packages feed

pure-noise 0.2.1.1 → 0.2.2.0

raw patch · 20 files changed

+3916/−201 lines, 20 filesdep +mtldep +splitmixdep −tasty-discoverdep ~tasty-benchdep ~vectorPVP ok

version bump matches the API change (PVP)

Dependencies added: mtl, splitmix

Dependencies removed: tasty-discover

Dependency ranges changed: tasty-bench, vector

API changes (from Hackage documentation)

+ Numeric.Noise: cellular3 :: (RealFrac a, Floating a) => CellularConfig a -> Noise3 a
+ Numeric.Noise: mkNoise1 :: (Seed -> p -> v) -> Noise1' p v
+ Numeric.Noise: mkNoise2 :: (Seed -> p -> p -> v) -> Noise2' p v
+ Numeric.Noise: mkNoise3 :: (Seed -> p -> p -> p -> v) -> Noise3' p v
+ Numeric.Noise: openSimplex3 :: RealFrac a => Noise3 a
+ Numeric.Noise: superSimplex3 :: RealFrac a => Noise3 a
+ Numeric.Noise.Cellular: noise3 :: (RealFrac a, Floating a) => CellularConfig a -> Noise3 a
+ Numeric.Noise.OpenSimplex: noise3 :: RealFrac a => Noise3 a
+ Numeric.Noise.OpenSimplex: noise3Base :: RealFrac a => Seed -> a -> a -> a -> a
+ Numeric.Noise.SuperSimplex: noise3 :: RealFrac a => Noise3 a
+ Numeric.Noise.SuperSimplex: noise3Base :: RealFrac a => Seed -> a -> a -> a -> a

Files

CHANGELOG.md view
@@ -8,6 +8,55 @@  ## Unreleased +## 0.2.2.0 2026-07-19++This is the final release of the 0.2.2.0 line.++### Added++- Exports for `mkNoise1`, `mkNoise2`, and `mkNoise3`, in case library users+  wish to hoist custom kernels.+- FastNoiseLite comparisons are now in-repository and reproducible. Comparison+  mechanism goes through tasty-bench.+- The sdist now includes `bench/fnl-compare/cbits/FastNoiseLite.h`, so the+  FNL comparison benchmark can build from a Hackage tarball.+- 3D counterparts for the simplex and cellular families: `openSimplex3`,+  `superSimplex3`, and `cellular3`.+- Flag `optimize`, with default `true`, which bumps the library's optimization level+  to `-O2` by default. Refer to the flag's documentation for rationale.+- Flag `mfma`, which will set `-mfma` for GHC, and may open pathways to+  fused-multiply-add optimizations+- Flag `mavx`, which will set `-mavx` for GHC, and may open pathways for AVX2+  SIMD optimization in the future.+- Flag `llvm-bench` (manual, default off) gating all LLVM-specific benchmark+  options (`-fllvm`, pinned `opt`/`llc`/`clang`, `-mavx`, `-mfma`,+  `-optlc-fp-contract=fast`).+- Benchmark runs now write a `.meta` provenance sidecar (toolchain versions,+  CPU governor/turbo state, AC power state, pinning) next to each CSV.++### Changed++- `pure-noise` now builds at `-O2` by default.+- Project now uses haskell.nix instead of Stack.+- Various updates to documentation, to improve quality+- Performance documentation and results tables re-collected on the 0.2.2.0+  toolchain (README.md and bench/README.md).+- `tested-with: GHC == 9.6.7, 9.8.4, 9.12.2`.++### Fixed++- Replace `round`-based rounding with a truncating `fastRound`. This replaces an+  `rintFloat` call in the core lowering with a `float2Int` call, which improves+  performance in hot loops.+  - This changes `cellular2` output at half-integer coordinates: cell selection+    now rounds half away from zero, matching FastNoiseLite's `FastRound`, rather+    than half to even. Numerically observable but visually imperceptible.+- Match FastNoiseLite's evaluation order in `openSimplex2`'s middle-corner+  contribution. Output changes by a few ULP.+- Fix an issue with the 3D benchmark harness where degenerate index math+  collapsed the sample set to ~4 distinct points, providing a poor signal on+  real-world performance.+ ## 0.2.1.1 2025-10-31  ### Changed
+ LICENSE_FastNoiseLite view
@@ -0,0 +1,22 @@+MIT License++Copyright(c) 2020 Jordan Peck (jordan.me2@gmail.com)+Copyright(c) 2020 Contributors++Permission is hereby granted, free of charge, to any person obtaining a copy+of this software and associated documentation files (the "Software"), to deal+in the Software without restriction, including without limitation the rights+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell+copies of the Software, and to permit persons to whom the Software is+furnished to do so, subject to the following conditions:++The above copyright notice and this permission notice shall be included in all+copies or substantial portions of the Software.++THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE+SOFTWARE.
README.md view
@@ -9,8 +9,9 @@   `Fractional`, `Monad`, etc). - **Complex effects** like domain warping and multi-octave fractals with clean,   type-safe composition.-- **84-95% of C++ FastNoiseLite performance** through aggressive optimization and-  LLVM compilation.+- **Competitive with C++**: roughly 70-100% of FastNoiseLite's single-threaded+  random-access throughput under the LLVM backend. And 2D cellular noise runs+  ahead of C++, measured against an optimized FNL build.  **For detailed FastNoiseLite comparison, methodology, and reproducibility instructions, see the [benchmark README](https://github.com/jtnuttall/pure-noise/blob/main/bench/README.md).**@@ -24,7 +25,7 @@ - This project grew from a port of the excellent   [FastNoiseLite](https://github.com/Auburn/FastNoiseLite) library. The library   structure has been tuned to perform well in Haskell and fit well with Haskell-  semantics, but the core noise implementations are the same.+  semantics, but the core noise share their origin with FNL. - All credit for the original design, algorithms, and implementation goes to its   creator **[Jordan Peck (@Auburn)](https://github.com/Auburn)**. I'm grateful for   their work and the opportunity to learn from it.@@ -32,6 +33,26 @@   were originally ported, is (C) 2020 Jordan Peck and is licensed under the MIT   license, a copy of which is included in this repository. +## FastNoiseLite compatibility++pure-noise shares its lineage with FNL, but it isn't intended as a 1:1 port —+kernels are restructured for GHC, and some families intentionally diverge.+Where outputs stand today:++| family                                   | output vs FNL                                 |+| ---------------------------------------- | --------------------------------------------- |+| `perlin2/3`, `cellular2/3`               | bit-exact                                     |+| `openSimplex2/3`, `superSimplex2/3`      | within a few ULP                              |+| `value2/3`, `valueCubic2/3`              | diverges (hash finalization)                  |+| `fractal2/3`, `ridged2/3`, `pingPong2/3` | diverges (octave normalization and weighting) |+| `billow2/3`                              | no FNL counterpart                            |++> [!IMPORTANT]+>+> The `value`, `valueCubic`, and fractal families will align with FNL in 0.3,+> which changes their output for a given seed. Pin `pure-noise < 0.3` if you+> depend on seed-stable output from them.+ ## Usage  The library provides composable noise functions. `Noise2` and `Noise3` are type@@ -136,47 +157,99 @@  ## Performance notes -- In single-threaded scenarios with LLVM enabled, this library achieves **84-95%-  of C++ FastNoiseLite performance**.-- This library benefits considerably from compilation with the LLVM backend-  (`-fllvm`). Benchmarks suggest a ~50-80% difference depending on the kind of noise.+- In single-threaded scenarios with LLVM enabled, this library reaches+  **roughly 70-100% of C++ FastNoiseLite throughput at random access, with 2D+  cellular noise ~7% faster than C++**. Grid-coherent workloads measure lower+  for the simplex family. These numbers are measured against an optimized+  FNL build (AVX2+FMA, FP contraction on).+- This library performs significantly better under the LLVM backend+  (`-fllvm`); the native code generator is not recommended where generation+  speed is a real concern.+- See the [benchmark README](https://github.com/jtnuttall/pure-noise/blob/main/bench/README.md)+  for per-algorithm, per-workload results and methodology. +### Recommended build flags++#### Native optimization flags for the library++For the library itself, copy `cabal.project.local.template` into your project+(`+optimize` is on by default; `+mfma +mavx` are safe on modern x86-64).++For bulk generation, prefer parallel evaluation via `massiv`.++#### Fused multiply add++For the fastest downstream executables, compile the modules that _call_ the+noise functions (the kernels inline into your code) with:++```+-fllvm -mavx -mfma -optlc-fp-contract=fast+```++> [!WARNING]+>+> 1. Results differ from a build without fma fusion by a few ULP.+> 2. Fusion decisions may vary across LLVM versions.+>    Skip this flag if you need bit-identical output across builds/toolchains.+> 3. You must use `-fllvm` to use this flag. It has no effect on the native+>    code generator.++`-optlc-fp-contract=fast` lets LLVM fuse multiply-add chains into FMA+instructions, an optimization modern C++ compilers apply by default. This+gives a ~5-15% improvement on these kernels in testing.+ ### Parallel noise generation  This library integrates well with [massiv](https://hackage.haskell.org/package/massiv)-for parallel computation. Parallel performance can reach 10-15x single-threaded-performance.+for parallel computation. Parallel evaluation reaches roughly 6-9x+single-threaded throughput on a 14-core machine in pinned-clock measurements. -**This is the recommended approach for generating large noise textures or datasets.**+> [!IMPORTANT]+>+> Massiv integration is the recommended approach for generating large noise+> textures or datasets.  ### Benchmarks  #### Results -Measured by values / second generated by the noise functions. These results come-from a benchmark with `-fllvm` enabled.+Measured by values / second generated by the noise functions, in the+`llvm-bench` configuration (LLVM backend + FP contraction), on an i7-1370P+**pinned at 1.9 GHz** for measurement stability — -There's inevitably some noise in the measurements because all of the results are-forced into an unboxed vector.+Absolute figures scale with CPU clock. The FastNoiseLite ratios are intended+to be clock-invariant. +There's inevitably some noise in the measurements because the results are forced+into an unboxed vector.++> [!NOTE]+>+> These numbers are lower than the initial release because they were re-run on a+> slower processor. Order-of-magnitude/comparative difference remains reasonably+> stable. See the benchmark README for details.+ ##### 2D  | name          | Float (values/sec) | Double (values/sec) | | ------------- | ------------------ | ------------------- |-| value2        | 173_511_654        | 189_119_731         |-| perlin2       | 154_674_464        | 161_114_532         |-| openSimplex2  | 74_747_031         | 74_332_345          |-| valueCubic2   | 61_415_544         | 62_481_313          |-| superSimplex2 | 51_295_369         | 50_383_577          |-| cellular2     | 34_996_382         | 32_652_899          |+| value2        | 64_407_126         | 68_260_653          |+| perlin2       | 61_301_663         | 65_143_707          |+| openSimplex2  | 25_982_291         | 27_045_472          |+| valueCubic2   | 22_743_642         | 23_403_842          |+| superSimplex2 | 17_167_069         | 17_762_669          |+| cellular2     | 16_025_950         | 16_007_044          |  ##### 3D -| name        | Float (values/sec) | Double (values/sec) |-| ----------- | ------------------ | ------------------- |-| value3      | 90_805_572         | 93_188_363          |-| perlin3     | 74_080_032         | 82_477_882          |-| valueCubic3 | 18_765_912         | 18_284_749          |+| name          | Float (values/sec) | Double (values/sec) |+| ------------- | ------------------ | ------------------- |+| value3        | 34_673_623         | 35_929_146          |+| perlin3       | 29_325_590         | 30_432_482          |+| openSimplex3  | 10_975_857         | 10_922_644          |+| superSimplex3 | 9_232_128          | 9_166_843           |+| valueCubic3   | 7_453_365          | 7_278_612           |+| cellular3     | 5_238_497          | 5_061_911           |  ## Examples 
bench/Bench.hs view
@@ -3,7 +3,7 @@ import Data.Typeable import Data.Vector.Unboxed qualified as U import Numeric.Noise-import System.Random.Stateful+import System.Random.Stateful (StatefulGen, UniformRange, newAtomicGenM, newStdGen, uniformRM)  main :: IO () main = do@@ -27,16 +27,17 @@         "3D"         ( baseline3 sz             <> benchPerlin3 octaves sz+            <> benchOpenSimplex3 octaves sz+            <> benchSuperSimplex3 octaves sz             <> benchValue3 octaves sz             <> benchValueCubic3 octaves sz+            <> benchCellular3 sz         )     , bgroup         "2D massiv"         ( benchMassivBase2 massivW massivH             <> benchMassivFractal2 octaves massivW massivH         )-    , bgroup "FNL compare 2D" benchFnlCompare2D-    , bgroup "FNL compare 3D" benchFnlCompare3D     ]  label :: (Typeable a) => String -> Int -> Proxy a -> String@@ -242,66 +243,21 @@       ]   ] --- | Create environment for FNL-style 2D benchmarks with integer-aligned coordinates--- Mimics FastNoiseLite's benchmark methodology exactly:--- for(y = 0; y < gridSize; y++)---   for(x = 0; x < gridSize; x++)---     noise(float(x), float(y))-createEnvFnl2 :: Int -> IO (Seed, U.Vector (Float, Float))-createEnvFnl2 gridSize = do-  g <- newAtomicGenM =<< newStdGen-  seed <- uniformRM (minBound, maxBound) g-  let v = U.generate (gridSize * gridSize) $ \i ->-        let x = fromIntegral (i `mod` gridSize) :: Float-            y = fromIntegral (i `div` gridSize) :: Float-         in (x, y)-  pure (seed, v)-{-# INLINE createEnvFnl2 #-}---- | Benchmark for FNL comparison (integer coordinates, Float only)--- Uses foldl' to avoid vector materialization overhead, matching FNL's DoNotOptimize approach--- Sums results to prevent DCE while keeping overhead minimal-benchFnlCompare2-  :: String-  -> Int-  -> Noise2 Float-  -> Benchmark-benchFnlCompare2 lbl gridSize f =-  env (createEnvFnl2 gridSize) $ \ ~(seed, v) ->-    bench (lbl <> ": Float (FNL grid) x" <> show (gridSize * gridSize)) $-      nf (\vec -> U.foldl' (\acc (x, y) -> acc + (noise2At f seed) x y) 0 vec) v-{-# INLINE benchFnlCompare2 #-}---- | FNL-style 2D benchmarks: 512x512 grid with integer coordinates-benchFnlCompare2D :: [Benchmark]-benchFnlCompare2D =-  let gridSize = 512-   in [ bgroup-          "FNL compare"-          [ benchFnlCompare2 "value2" gridSize value2-          , benchFnlCompare2 "perlin2" gridSize perlin2-          , benchFnlCompare2 "openSimplex2" gridSize openSimplex2-          , benchFnlCompare2 "superSimplex2" gridSize superSimplex2-          , benchFnlCompare2 "valueCubic2" gridSize valueCubic2-          , benchFnlCompare2-              "cellular2 (Distance)"-              gridSize-              (cellular2 defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = Distance})-          ]-      ]-+-- offsets are drawn per element (like createEnv2) and the strides are+-- independent per axis; the previous version drew one offset triple and+-- collapsed the index math to ~4 distinct points createEnv3 :: (U.Unbox a, UniformRange a, RealFrac a) => Int -> IO (Seed, U.Vector (a, a, a)) createEnv3 sz = do   g <- newAtomicGenM =<< newStdGen   seed <- uniformRM (minBound, maxBound) g-  offsetX <- uniformRM (0.00001, 0.99999) g-  offsetY <- uniformRM (0.00001, 0.99999) g-  offsetZ <- uniformRM (0.00001, 0.99999) g+  let d = ceiling (fromIntegral sz ** (1 / 3) :: Double)   !ixs <- U.generateM sz $ \i -> do-    let d = sz `div` 3-        !x = fromIntegral $ i `div` d `mod` d-        !y = fromIntegral $ i `div` (d * d)-        !z = fromIntegral $ i `div` d+    offsetX <- uniformRM (0.00001, 0.99999) g+    offsetY <- uniformRM (0.00001, 0.99999) g+    offsetZ <- uniformRM (0.00001, 0.99999) g+    let !x = fromIntegral (i `mod` d)+        !y = fromIntegral ((i `div` d) `mod` d)+        !z = fromIntegral (i `div` (d * d))     pure (x + offsetX, y + offsetY, z + offsetZ)   pure (seed, ixs) {-# INLINE createEnv3 #-}@@ -345,6 +301,51 @@       ]   ] +benchOpenSimplex3 :: Int -> Int -> [Benchmark]+benchOpenSimplex3 octaves sz =+  [ bgroup+      "openSimplex3"+      [ benchMany3 @Float "" sz openSimplex3+      , benchMany3 @Double "" sz openSimplex3+      , benchMany3 @Float "fractal" sz (fractal3 defaultFractalConfig{octaves} openSimplex3)+      , benchMany3 @Double "fractal" sz (fractal3 defaultFractalConfig{octaves} openSimplex3)+      ]+  ]++benchSuperSimplex3 :: Int -> Int -> [Benchmark]+benchSuperSimplex3 octaves sz =+  [ bgroup+      "superSimplex3"+      [ benchMany3 @Float "" sz superSimplex3+      , benchMany3 @Double "" sz superSimplex3+      , benchMany3 @Float "fractal" sz (fractal3 defaultFractalConfig{octaves} superSimplex3)+      , benchMany3 @Double "fractal" sz (fractal3 defaultFractalConfig{octaves} superSimplex3)+      ]+  ]++benchCellular3 :: Int -> [Benchmark]+benchCellular3 sz =+  [ bgroup+      "cellular3"+      [ benchMany3 @Float+          "DistEuclidean CellValue"+          sz+          (cellular3 defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = CellValue})+      , benchMany3 @Float+          "DistEuclidean Distance2Add"+          sz+          (cellular3 defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = Distance2Add})+      , benchMany3 @Double+          "DistEuclidean CellValue"+          sz+          (cellular3 defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = CellValue})+      , benchMany3 @Double+          "DistEuclidean Distance2Add"+          sz+          (cellular3 defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = Distance2Add})+      ]+  ]+ benchValue3 :: Int -> Int -> [Benchmark] benchValue3 octaves sz =   [ bgroup@@ -378,51 +379,6 @@       , benchMany3 @Double "pingPong" sz (pingPong3 defaultFractalConfig{octaves} defaultPingPongStrength valueCubic3)       ]   ]---- | Create environment for FNL-style 3D benchmarks with integer-aligned coordinates--- Mimics FastNoiseLite's benchmark methodology exactly:--- for(z = 0; z < gridSize; z++)---   for(y = 0; y < gridSize; y++)---     for(x = 0; x < gridSize; x++)---       noise(float(x), float(y), float(z))-createEnvFnl3 :: Int -> IO (Seed, U.Vector (Float, Float, Float))-createEnvFnl3 gridSize = do-  g <- newAtomicGenM =<< newStdGen-  seed <- uniformRM (minBound, maxBound) g-  let gridSq = gridSize * gridSize-      v = U.generate (gridSize * gridSize * gridSize) $ \i ->-        let x = fromIntegral (i `mod` gridSize) :: Float-            y = fromIntegral ((i `div` gridSize) `mod` gridSize) :: Float-            z = fromIntegral (i `div` gridSq) :: Float-         in (x, y, z)-  pure (seed, v)-{-# INLINE createEnvFnl3 #-}---- | Benchmark for FNL 3D comparison (integer coordinates, Float only)--- Uses foldl' to avoid vector materialization overhead, matching FNL's DoNotOptimize approach--- Sums results to prevent DCE while keeping overhead minimal-benchFnlCompare3-  :: String-  -> Int-  -> Noise3 Float-  -> Benchmark-benchFnlCompare3 lbl gridSize f =-  env (createEnvFnl3 gridSize) $ \ ~(seed, v) ->-    bench (lbl <> ": Float (FNL grid) x" <> show (gridSize * gridSize * gridSize)) $-      nf (\vec -> U.foldl' (\acc (x, y, z) -> acc + noise3At f seed x y z) 0 vec) v-{-# INLINE benchFnlCompare3 #-}---- | FNL-style 3D benchmarks: 64x64x64 grid with integer coordinates-benchFnlCompare3D :: [Benchmark]-benchFnlCompare3D =-  let gridSize = 64-   in [ bgroup-          "FNL compare"-          [ benchFnlCompare3 "value3" gridSize value3-          , benchFnlCompare3 "perlin3" gridSize perlin3-          , benchFnlCompare3 "valueCubic3" gridSize valueCubic3-          ]-      ]  benchMassiv2   :: forall a
+ bench/fnl-compare/FnlBench.hs view
@@ -0,0 +1,234 @@+-- | Performance comparison against the vendored FastNoiseLite+--+-- Ratios are printed by tasty-bench's 'bcompare': each @pure-noise@ line ends+-- with e.g. @1.05x@ = its mean time relative to the matching @fnl@ line+-- (lower is better; 1.05x means pure-noise is 5% slower).+--+-- Ratios only print when the matching fnl benchmark runs in the same invocation,+-- so filter by group (e.g. @-p \/perlin\/@), not by side.+module Main (main) where++import Control.Monad.State.Strict (evalState, state)+import Data.Vector.Storable qualified as S+import Foreign.C.Types (CFloat (..), CInt (..))+import Foreign.Ptr (Ptr)+import Numeric.Noise+import System.Random.SplitMix (SMGen, mkSMGen, nextFloat)+import Test.Tasty.Bench+import Test.Tasty.Patterns.Printer (printAwkExpr)++main :: IO ()+main =+  defaultMain (map (mapLeafBenchmarks addCompare) benchmarks)++-- | Attach a bcompare ratio to every pure-noise leaf, targeting the fnl leaf+-- in the same group (the pattern matches exactly one benchmark).+addCompare :: [String] -> Benchmark -> Benchmark+addCompare (leaf : path) b+  | leaf == pnLeaf = bcompare (printAwkExpr (locateBenchmark (fnlLeaf : path))) b+addCompare _ b = b++fnlLeaf, pnLeaf :: String+fnlLeaf = "fnl x" <> show samples+pnLeaf = "pure-noise x" <> show samples++benchmarks :: [Benchmark]+benchmarks =+  [ env mkEnv2 $ \ ~(grid, rand, kOne, kSmall) ->+      bgroup+        "2D"+        [ algo2 grid rand kOne kSmall "value" fnlValue value2+        , algo2 grid rand kOne kSmall "valueCubic" fnlValueCubic valueCubic2+        , algo2 grid rand kOne kSmall "perlin" fnlPerlin perlin2+        , algo2 grid rand kOne kSmall "openSimplex2" fnlOpenSimplex2 openSimplex2+        , algo2 grid rand kOne kSmall "superSimplex2" fnlOpenSimplex2S superSimplex2+        , algo2 grid rand kOne kSmall "cellular" fnlCellular (cellular2 benchCellularConfig)+        ]+  , env mkEnv3 $ \ ~(grid, rand, kOne, kSmall) ->+      bgroup+        "3D"+        [ algo3 grid rand kOne kSmall "value" fnlValue value3+        , algo3 grid rand kOne kSmall "valueCubic" fnlValueCubic valueCubic3+        , algo3 grid rand kOne kSmall "perlin" fnlPerlin perlin3+        , algo3 grid rand kOne kSmall "openSimplex2" fnlOpenSimplex2 openSimplex3+        , algo3 grid rand kOne kSmall "superSimplex2" fnlOpenSimplex2S superSimplex3+        , algo3 grid rand kOne kSmall "cellular" fnlCellular (cellular3 benchCellularConfig)+        ]+  ]++algo2+  :: (S.Vector Float, S.Vector Float)+  -> (S.Vector Float, S.Vector Float)+  -> Float+  -> Float+  -> String+  -> CInt+  -> Noise2 Float+  -> Benchmark+algo2 grid rand kOne kSmall name ty nz =+  bgroup+    name+    [ bgroup "grid" (pair grid kOne)+    , bgroup "freq001" (pair grid kSmall)+    , bgroup "random" (pair rand kOne)+    ]+ where+  pair ~(xs, ys) k =+    [ bench fnlLeaf $ whnfIO (fnlSum2 ty k xs ys)+    , bench pnLeaf $ whnf (\kk -> sumNoise2 nz kk xs ys) k+    ]+{-# INLINE algo2 #-}++algo3+  :: (S.Vector Float, S.Vector Float, S.Vector Float)+  -> (S.Vector Float, S.Vector Float, S.Vector Float)+  -> Float+  -> Float+  -> String+  -> CInt+  -> Noise3 Float+  -> Benchmark+algo3 grid rand kOne kSmall name ty nz =+  bgroup+    name+    [ bgroup "grid" (pair grid kOne)+    , bgroup "freq001" (pair grid kSmall)+    , bgroup "random" (pair rand kOne)+    ]+ where+  -- lazy pattern: see algo2+  pair ~(xs, ys, zs) k =+    [ bench fnlLeaf $ whnfIO (fnlSum3 ty k xs ys zs)+    , bench pnLeaf $ whnf (\kk -> sumNoise3 nz kk xs ys zs) k+    ]+{-# INLINE algo3 #-}++----------------------------------------------------------------------------------------------------+-- Matched sweeps+----------------------------------------------------------------------------------------------------++sumNoise2 :: Noise2 Float -> Float -> S.Vector Float -> S.Vector Float -> Float+sumNoise2 nz k xs ys = go 0 0+ where+  !n = S.length xs+  go !i !acc+    | i >= n = acc+    | otherwise =+        go (i + 1) (acc + noise2At nz benchSeed (k * S.unsafeIndex xs i) (k * S.unsafeIndex ys i))+{-# INLINE sumNoise2 #-}++sumNoise3 :: Noise3 Float -> Float -> S.Vector Float -> S.Vector Float -> S.Vector Float -> Float+sumNoise3 nz k xs ys zs = go 0 0+ where+  !n = S.length xs+  go !i !acc+    | i >= n = acc+    | otherwise =+        go+          (i + 1)+          (acc + noise3At nz benchSeed (k * S.unsafeIndex xs i) (k * S.unsafeIndex ys i) (k * S.unsafeIndex zs i))+{-# INLINE sumNoise3 #-}++----------------------------------------------------------------------------------------------------+-- FFI (fnl_bench_shim.cpp)+----------------------------------------------------------------------------------------------------++foreign import ccall unsafe "fnl_bench_sum2"+  c_fnlSum2 :: CInt -> CInt -> CFloat -> CInt -> Ptr Float -> Ptr Float -> IO CFloat++foreign import ccall unsafe "fnl_bench_sum3"+  c_fnlSum3 :: CInt -> CInt -> CFloat -> CInt -> Ptr Float -> Ptr Float -> Ptr Float -> IO CFloat++fnlSum2 :: CInt -> Float -> S.Vector Float -> S.Vector Float -> IO Float+fnlSum2 ty k xs ys =+  S.unsafeWith xs $ \px ->+    S.unsafeWith ys $ \py -> do+      CFloat r <- c_fnlSum2 ty fnlBenchSeed (CFloat k) (fromIntegral (S.length xs)) px py+      pure r++fnlSum3 :: CInt -> Float -> S.Vector Float -> S.Vector Float -> S.Vector Float -> IO Float+fnlSum3 ty k xs ys zs =+  S.unsafeWith xs $ \px ->+    S.unsafeWith ys $ \py ->+      S.unsafeWith zs $ \pz -> do+        CFloat r <- c_fnlSum3 ty fnlBenchSeed (CFloat k) (fromIntegral (S.length xs)) px py pz+        pure r++fnlOpenSimplex2, fnlOpenSimplex2S, fnlCellular, fnlPerlin, fnlValueCubic, fnlValue :: CInt+fnlOpenSimplex2 = 0+fnlOpenSimplex2S = 1+fnlCellular = 2+fnlPerlin = 3+fnlValueCubic = 4+fnlValue = 5++-- 1337 is the FNL bench convention.+benchSeed :: Seed+benchSeed = 1337++fnlBenchSeed :: CInt+fnlBenchSeed = 1337++benchCellularConfig :: CellularConfig Float+benchCellularConfig = defaultCellularConfig{cellularDistanceFn = DistEuclidean, cellularResult = Distance}++-- ---------------------------------------------------------------------------+-- Coordinate environments (built once per dimension)+-- ---------------------------------------------------------------------------++-- n = 262144 everywhere: 512*512 (2D grid), 64^3 (3D grid), and the random+-- streams, so every leaf carries the same "xN" count for vps.py.+gridSize2, gridSize3, samples :: Int+gridSize2 = 512+gridSize3 = 64+samples = gridSize2 * gridSize2++-- (grid xs/ys, random xs/ys, kOne = 1.0, kSmall = 0.01). The ks come out of+-- the IO action so the simplifier cannot constant-fold the per-point multiply+-- on the Haskell side only.+mkEnv2 :: IO ((S.Vector Float, S.Vector Float), (S.Vector Float, S.Vector Float), Float, Float)+mkEnv2 = pure (mkGrid2, mkRand2 samples, 1.0, 0.01)++mkEnv3+  :: IO+       ( (S.Vector Float, S.Vector Float, S.Vector Float)+       , (S.Vector Float, S.Vector Float, S.Vector Float)+       , Float+       , Float+       )+mkEnv3 = pure (mkGrid3, mkRand3 samples, 1.0, 0.01)++-- Integer lattice in FNL benchmark loop order (x fastest, then y, then z).+mkGrid2 :: (S.Vector Float, S.Vector Float)+mkGrid2 =+  ( S.generate samples (\i -> fromIntegral (i `mod` gridSize2))+  , S.generate samples (\i -> fromIntegral (i `div` gridSize2))+  )++mkGrid3 :: (S.Vector Float, S.Vector Float, S.Vector Float)+mkGrid3 =+  ( S.generate samples (\i -> fromIntegral (i `mod` gridSize3))+  , S.generate samples (\i -> fromIntegral ((i `div` gridSize3) `mod` gridSize3))+  , S.generate samples (\i -> fromIntegral (i `div` (gridSize3 * gridSize3)))+  )++smgen :: SMGen+smgen = mkSMGen 0x9E3779B97F4A7C15++draws :: Int -> S.Vector Float+draws n = S.map (* 512) . evalState (S.generateM n . const . state $ nextFloat) $ smgen++mkRand2 :: Int -> (S.Vector Float, S.Vector Float)+mkRand2 n =+  let v = draws (2 * n)+   in ( S.generate n (\i -> v S.! (2 * i))+      , S.generate n (\i -> v S.! (2 * i + 1))+      )++mkRand3 :: Int -> (S.Vector Float, S.Vector Float, S.Vector Float)+mkRand3 n =+  let v = draws (3 * n)+   in ( S.generate n (\i -> v S.! (3 * i))+      , S.generate n (\i -> v S.! (3 * i + 1))+      , S.generate n (\i -> v S.! (3 * i + 2))+      )
+ bench/fnl-compare/cbits/FastNoiseLite.h view
@@ -0,0 +1,2586 @@+// MIT License+//+// Copyright(c) 2023 Jordan Peck (jordan.me2@gmail.com)+// Copyright(c) 2023 Contributors+//+// Permission is hereby granted, free of charge, to any person obtaining a copy+// of this software and associated documentation files(the "Software"), to deal+// in the Software without restriction, including without limitation the rights+// to use, copy, modify, merge, publish, distribute, sublicense, and / or sell+// copies of the Software, and to permit persons to whom the Software is+// furnished to do so, subject to the following conditions :+//+// The above copyright notice and this permission notice shall be included in all+// copies or substantial portions of the Software.+//+// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR+// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,+// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.IN NO EVENT SHALL THE+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,+// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE+// SOFTWARE.+//+// .'',;:cldxkO00KKXXNNWWWNNXKOkxdollcc::::::;:::ccllloooolllllllllooollc:,'...        ...........',;cldxkO000Okxdlc::;;;,,;;;::cclllllll+// ..',;:ldxO0KXXNNNNNNNNXXK0kxdolcc::::::;;;,,,,,,;;;;;;;;;;:::cclllllc:;'....       ...........',;:ldxO0KXXXK0Okxdolc::;;;;::cllodddddo+// ...',:loxO0KXNNNNNXXKK0Okxdolc::;::::::::;;;,,'''''.....''',;:clllllc:;,'............''''''''',;:loxO0KXNNNNNXK0Okxdollccccllodxxxxxxd+// ....';:ldkO0KXXXKK00Okxdolcc:;;;;;::cclllcc:;;,''..... ....',;clooddolcc:;;;;,,;;;;;::::;;;;;;:cloxk0KXNWWWWWWNXKK0Okxddoooddxxkkkkkxx+// .....';:ldxkOOOOOkxxdolcc:;;;,,,;;:cllooooolcc:;'...      ..,:codxkkkxddooollloooooooollcc:::::clodkO0KXNWWWWWWNNXK00Okxxxxxxxxkkkkxxx+// . ....';:cloddddo___________,,,,;;:clooddddoolc:,...      ..,:ldx__00OOOkkk___kkkkkkxxdollc::::cclodkO0KXXNNNNNNXXK0OOkxxxxxxxxxxxxddd+// .......',;:cccc:|           |,,,;;:cclooddddoll:;'..     ..';cox|  \KKK000|   |KK00OOkxdocc___;::clldxxkO0KKKKK00Okkxdddddddddddddddoo+// .......'',,,,,''|   ________|',,;;::cclloooooolc:;'......___:ldk|   \KK000|   |XKKK0Okxolc|   |;;::cclodxxkkkkxxdoolllcclllooodddooooo+// ''......''''....|   |  ....'',,,,;;;::cclloooollc:;,''.'|   |oxk|    \OOO0|   |KKK00Oxdoll|___|;;;;;::ccllllllcc::;;,,;;;:cclloooooooo+// ;;,''.......... |   |_____',,;;;____:___cllo________.___|   |___|     \xkk|   |KK_______ool___:::;________;;;_______...'',;;:ccclllloo+// c:;,''......... |         |:::/     '   |lo/        |           |      \dx|   |0/       \d|   |cc/        |'/       \......',,;;:ccllo+// ol:;,'..........|    _____|ll/    __    |o/   ______|____    ___|   |   \o|   |/   ___   \|   |o/   ______|/   ___   \ .......'',;:clo+// dlc;,...........|   |::clooo|    /  |   |x\___   \KXKKK0|   |dol|   |\   \|   |   |   |   |   |d\___   \..|   |  /   /       ....',:cl+// xoc;'...  .....'|   |llodddd|    \__|   |_____\   \KKK0O|   |lc:|   |'\       |   |___|   |   |_____\   \.|   |_/___/...      ...',;:c+// dlc;'... ....',;|   |oddddddo\          |          |Okkx|   |::;|   |..\      |\         /|   |          | \         |...    ....',;:c+// ol:,'.......',:c|___|xxxddollc\_____,___|_________/ddoll|___|,,,|___|...\_____|:\ ______/l|___|_________/...\________|'........',;::cc+// c:;'.......';:codxxkkkkxxolc::;::clodxkOO0OOkkxdollc::;;,,''''',,,,''''''''''',,'''''',;:loxkkOOkxol:;,'''',,;:ccllcc:;,'''''',;::ccll+// ;,'.......',:codxkOO0OOkxdlc:;,,;;:cldxxkkxxdolc:;;,,''.....'',;;:::;;,,,'''''........,;cldkO0KK0Okdoc::;;::cloodddoolc:;;;;;::ccllooo+// .........',;:lodxOO0000Okdoc:,,',,;:clloddoolc:;,''.......'',;:clooollc:;;,,''.......',:ldkOKXNNXX0Oxdolllloddxxxxxxdolccccccllooodddd+// .    .....';:cldxkO0000Okxol:;,''',,;::cccc:;,,'.......'',;:cldxxkkxxdolc:;;,'.......';coxOKXNWWWNXKOkxddddxxkkkkkkxdoollllooddxxxxkkk+//       ....',;:codxkO000OOxdoc:;,''',,,;;;;,''.......',,;:clodkO00000Okxolc::;,,''..',;:ldxOKXNWWWNNK0OkkkkkkkkkkkxxddooooodxxkOOOOO000+//       ....',;;clodxkkOOOkkdolc:;,,,,,,,,'..........,;:clodxkO0KKXKK0Okxdolcc::;;,,,;;:codkO0XXNNNNXKK0OOOOOkkkkxxdoollloodxkO0KKKXXXXX+//+// VERSION: 1.1.1+// https://github.com/Auburn/FastNoiseLite++#ifndef FASTNOISELITE_H+#define FASTNOISELITE_H++#include <cmath>++class FastNoiseLite+{+public:+    enum NoiseType+    {+        NoiseType_OpenSimplex2,+        NoiseType_OpenSimplex2S,+        NoiseType_Cellular,+        NoiseType_Perlin,+        NoiseType_ValueCubic,+        NoiseType_Value+    };++    enum RotationType3D+    {+        RotationType3D_None,+        RotationType3D_ImproveXYPlanes,+        RotationType3D_ImproveXZPlanes+    };++    enum FractalType+    {+        FractalType_None,+        FractalType_FBm,+        FractalType_Ridged,+        FractalType_PingPong,+        FractalType_DomainWarpProgressive,+        FractalType_DomainWarpIndependent+    };++    enum CellularDistanceFunction+    {+        CellularDistanceFunction_Euclidean,+        CellularDistanceFunction_EuclideanSq,+        CellularDistanceFunction_Manhattan,+        CellularDistanceFunction_Hybrid+    };++    enum CellularReturnType+    {+        CellularReturnType_CellValue,+        CellularReturnType_Distance,+        CellularReturnType_Distance2,+        CellularReturnType_Distance2Add,+        CellularReturnType_Distance2Sub,+        CellularReturnType_Distance2Mul,+        CellularReturnType_Distance2Div+    };++    enum DomainWarpType+    {+        DomainWarpType_OpenSimplex2,+        DomainWarpType_OpenSimplex2Reduced,+        DomainWarpType_BasicGrid+    };++    /// <summary>+    /// Create new FastNoise object with optional seed+    /// </summary>+    FastNoiseLite(int seed = 1337)+    {+        mSeed = seed;+        mFrequency = 0.01f;+        mNoiseType = NoiseType_OpenSimplex2;+        mRotationType3D = RotationType3D_None;+        mTransformType3D = TransformType3D_DefaultOpenSimplex2;++        mFractalType = FractalType_None;+        mOctaves = 3;+        mLacunarity = 2.0f;+        mGain = 0.5f;+        mWeightedStrength = 0.0f;+        mPingPongStrength = 2.0f;++        mFractalBounding = 1 / 1.75f;++        mCellularDistanceFunction = CellularDistanceFunction_EuclideanSq;+        mCellularReturnType = CellularReturnType_Distance;+        mCellularJitterModifier = 1.0f;++        mDomainWarpType = DomainWarpType_OpenSimplex2;+        mWarpTransformType3D = TransformType3D_DefaultOpenSimplex2;+        mDomainWarpAmp = 1.0f;+    }++    /// <summary>+    /// Sets seed used for all noise types+    /// </summary>+    /// <remarks>+    /// Default: 1337+    /// </remarks>+    void SetSeed(int seed) { mSeed = seed; }++    /// <summary>+    /// Sets frequency for all noise types+    /// </summary>+    /// <remarks>+    /// Default: 0.01+    /// </remarks>+    void SetFrequency(float frequency) { mFrequency = frequency; }++    /// <summary>+    /// Sets noise algorithm used for GetNoise(...)+    /// </summary>+    /// <remarks>+    /// Default: OpenSimplex2+    /// </remarks>+    void SetNoiseType(NoiseType noiseType)+    {+        mNoiseType = noiseType;+        UpdateTransformType3D();+    }++    /// <summary>+    /// Sets domain rotation type for 3D Noise and 3D DomainWarp.+    /// Can aid in reducing directional artifacts when sampling a 2D plane in 3D+    /// </summary>+    /// <remarks>+    /// Default: None+    /// </remarks>+    void SetRotationType3D(RotationType3D rotationType3D)+    {+        mRotationType3D = rotationType3D;+        UpdateTransformType3D();+        UpdateWarpTransformType3D();+    }++    /// <summary>+    /// Sets method for combining octaves in all fractal noise types+    /// </summary>+    /// <remarks>+    /// Default: None+    /// Note: FractalType_DomainWarp... only affects DomainWarp(...)+    /// </remarks>+    void SetFractalType(FractalType fractalType) { mFractalType = fractalType; }++    /// <summary>+    /// Sets octave count for all fractal noise types +    /// </summary>+    /// <remarks>+    /// Default: 3+    /// </remarks>+    void SetFractalOctaves(int octaves)+    {+        mOctaves = octaves;+        CalculateFractalBounding();+    }++    /// <summary>+    /// Sets octave lacunarity for all fractal noise types+    /// </summary>+    /// <remarks>+    /// Default: 2.0+    /// </remarks>+    void SetFractalLacunarity(float lacunarity) { mLacunarity = lacunarity; }++    /// <summary>+    /// Sets octave gain for all fractal noise types+    /// </summary>+    /// <remarks>+    /// Default: 0.5+    /// </remarks>+    void SetFractalGain(float gain)+    {+        mGain = gain;+        CalculateFractalBounding();+    }++    /// <summary>+    /// Sets octave weighting for all none DomainWarp fratal types+    /// </summary>+    /// <remarks>+    /// Default: 0.0+    /// Note: Keep between 0...1 to maintain -1...1 output bounding+    /// </remarks>+    void SetFractalWeightedStrength(float weightedStrength) { mWeightedStrength = weightedStrength; }++    /// <summary>+    /// Sets strength of the fractal ping pong effect+    /// </summary>+    /// <remarks>+    /// Default: 2.0+    /// </remarks>+    void SetFractalPingPongStrength(float pingPongStrength) { mPingPongStrength = pingPongStrength; }+++    /// <summary>+    /// Sets distance function used in cellular noise calculations+    /// </summary>+    /// <remarks>+    /// Default: Distance+    /// </remarks>+    void SetCellularDistanceFunction(CellularDistanceFunction cellularDistanceFunction) { mCellularDistanceFunction = cellularDistanceFunction; }++    /// <summary>+    /// Sets return type from cellular noise calculations+    /// </summary>+    /// <remarks>+    /// Default: EuclideanSq+    /// </remarks>+    void SetCellularReturnType(CellularReturnType cellularReturnType) { mCellularReturnType = cellularReturnType; }++    /// <summary>+    /// Sets the maximum distance a cellular point can move from it's grid position+    /// </summary>+    /// <remarks>+    /// Default: 1.0+    /// Note: Setting this higher than 1 will cause artifacts+    /// </remarks> +    void SetCellularJitter(float cellularJitter) { mCellularJitterModifier = cellularJitter; }+++    /// <summary>+    /// Sets the warp algorithm when using DomainWarp(...)+    /// </summary>+    /// <remarks>+    /// Default: OpenSimplex2+    /// </remarks>+    void SetDomainWarpType(DomainWarpType domainWarpType)+    {+        mDomainWarpType = domainWarpType;+        UpdateWarpTransformType3D();+    }+++    /// <summary>+    /// Sets the maximum warp distance from original position when using DomainWarp(...)+    /// </summary>+    /// <remarks>+    /// Default: 1.0+    /// </remarks>+    void SetDomainWarpAmp(float domainWarpAmp) { mDomainWarpAmp = domainWarpAmp; }+++    /// <summary>+    /// 2D noise at given position using current settings+    /// </summary>+    /// <returns>+    /// Noise output bounded between -1...1+    /// </returns>+    template <typename FNfloat>+    float GetNoise(FNfloat x, FNfloat y) const+    {+        Arguments_must_be_floating_point_values<FNfloat>();++        TransformNoiseCoordinate(x, y);++        switch (mFractalType)+        {+        default:+            return GenNoiseSingle(mSeed, x, y);+        case FractalType_FBm:+            return GenFractalFBm(x, y);+        case FractalType_Ridged:+            return GenFractalRidged(x, y);+        case FractalType_PingPong:+            return GenFractalPingPong(x, y);+        }+    }++    /// <summary>+    /// 3D noise at given position using current settings+    /// </summary>+    /// <returns>+    /// Noise output bounded between -1...1+    /// </returns>+    template <typename FNfloat>+    float GetNoise(FNfloat x, FNfloat y, FNfloat z) const+    {+        Arguments_must_be_floating_point_values<FNfloat>();++        TransformNoiseCoordinate(x, y, z);++        switch (mFractalType)+        {+        default:+            return GenNoiseSingle(mSeed, x, y, z);+        case FractalType_FBm:+            return GenFractalFBm(x, y, z);+        case FractalType_Ridged:+            return GenFractalRidged(x, y, z);+        case FractalType_PingPong:+            return GenFractalPingPong(x, y, z);+        }+    }+++    /// <summary>+    /// 2D warps the input position using current domain warp settings+    /// </summary>+    /// <example>+    /// Example usage with GetNoise+    /// <code>DomainWarp(x, y)+    /// noise = GetNoise(x, y)</code>+    /// </example>+    template <typename FNfloat>+    void DomainWarp(FNfloat& x, FNfloat& y) const+    {+        Arguments_must_be_floating_point_values<FNfloat>();++        switch (mFractalType)+        {+        default:+            DomainWarpSingle(x, y);+            break;+        case FractalType_DomainWarpProgressive:+            DomainWarpFractalProgressive(x, y);+            break;+        case FractalType_DomainWarpIndependent:+            DomainWarpFractalIndependent(x, y);+            break;+        }+    }++    /// <summary>+    /// 3D warps the input position using current domain warp settings+    /// </summary>+    /// <example>+    /// Example usage with GetNoise+    /// <code>DomainWarp(x, y, z)+    /// noise = GetNoise(x, y, z)</code>+    /// </example>+    template <typename FNfloat>+    void DomainWarp(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        Arguments_must_be_floating_point_values<FNfloat>();++        switch (mFractalType)+        {+        default:+            DomainWarpSingle(x, y, z);+            break;+        case FractalType_DomainWarpProgressive:+            DomainWarpFractalProgressive(x, y, z);+            break;+        case FractalType_DomainWarpIndependent:+            DomainWarpFractalIndependent(x, y, z);+            break;+        }+    }++private:+    template <typename T>+    struct Arguments_must_be_floating_point_values;++    enum TransformType3D+    {+        TransformType3D_None,+        TransformType3D_ImproveXYPlanes,+        TransformType3D_ImproveXZPlanes,+        TransformType3D_DefaultOpenSimplex2+    };++    int mSeed;+    float mFrequency;+    NoiseType mNoiseType;+    RotationType3D mRotationType3D;+    TransformType3D mTransformType3D;++    FractalType mFractalType;+    int mOctaves;+    float mLacunarity;+    float mGain;+    float mWeightedStrength;+    float mPingPongStrength;++    float mFractalBounding;++    CellularDistanceFunction mCellularDistanceFunction;+    CellularReturnType mCellularReturnType;+    float mCellularJitterModifier;++    DomainWarpType mDomainWarpType;+    TransformType3D mWarpTransformType3D;+    float mDomainWarpAmp;+++    template <typename T>+    struct Lookup+    {+        static const T Gradients2D[];+        static const T Gradients3D[];+        static const T RandVecs2D[];+        static const T RandVecs3D[];+    };++    static float FastMin(float a, float b) { return a < b ? a : b; }++    static float FastMax(float a, float b) { return a > b ? a : b; }++    static float FastAbs(float f) { return f < 0 ? -f : f; }++    static float FastSqrt(float f) { return sqrtf(f); }++    template <typename FNfloat>+    static int FastFloor(FNfloat f) { return f >= 0 ? (int)f : (int)f - 1; }++    template <typename FNfloat>+    static int FastRound(FNfloat f) { return f >= 0 ? (int)(f + 0.5f) : (int)(f - 0.5f); }++    static float Lerp(float a, float b, float t) { return a + t * (b - a); }++    static float InterpHermite(float t) { return t * t * (3 - 2 * t); }++    static float InterpQuintic(float t) { return t * t * t * (t * (t * 6 - 15) + 10); }++    static float CubicLerp(float a, float b, float c, float d, float t)+    {+        float p = (d - c) - (a - b);+        return t * t * t * p + t * t * ((a - b) - p) + t * (c - a) + b;+    }++    static float PingPong(float t)+    {+        t -= (int)(t * 0.5f) * 2;+        return t < 1 ? t : 2 - t;+    }++    void CalculateFractalBounding()+    {+        float gain = FastAbs(mGain);+        float amp = gain;+        float ampFractal = 1.0f;+        for (int i = 1; i < mOctaves; i++)+        {+            ampFractal += amp;+            amp *= gain;+        }+        mFractalBounding = 1 / ampFractal;+    }++    // Hashing+    static const int PrimeX = 501125321;+    static const int PrimeY = 1136930381;+    static const int PrimeZ = 1720413743;++    static int Hash(int seed, int xPrimed, int yPrimed)+    {+        int hash = seed ^ xPrimed ^ yPrimed;++        hash *= 0x27d4eb2d;+        return hash;+    }+++    static int Hash(int seed, int xPrimed, int yPrimed, int zPrimed)+    {+        int hash = seed ^ xPrimed ^ yPrimed ^ zPrimed;++        hash *= 0x27d4eb2d;+        return hash;+    }+++    static float ValCoord(int seed, int xPrimed, int yPrimed)+    {+        int hash = Hash(seed, xPrimed, yPrimed);++        hash *= hash;+        hash ^= hash << 19;+        return hash * (1 / 2147483648.0f);+    }+++    static float ValCoord(int seed, int xPrimed, int yPrimed, int zPrimed)+    {+        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);++        hash *= hash;+        hash ^= hash << 19;+        return hash * (1 / 2147483648.0f);+    }+++    float GradCoord(int seed, int xPrimed, int yPrimed, float xd, float yd) const+    {+        int hash = Hash(seed, xPrimed, yPrimed);+        hash ^= hash >> 15;+        hash &= 127 << 1;++        float xg = Lookup<float>::Gradients2D[hash];+        float yg = Lookup<float>::Gradients2D[hash | 1];++        return xd * xg + yd * yg;+    }+++    float GradCoord(int seed, int xPrimed, int yPrimed, int zPrimed, float xd, float yd, float zd) const+    {+        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);+        hash ^= hash >> 15;+        hash &= 63 << 2;++        float xg = Lookup<float>::Gradients3D[hash];+        float yg = Lookup<float>::Gradients3D[hash | 1];+        float zg = Lookup<float>::Gradients3D[hash | 2];++        return xd * xg + yd * yg + zd * zg;+    }+++    void GradCoordOut(int seed, int xPrimed, int yPrimed, float& xo, float& yo) const+    {+        int hash = Hash(seed, xPrimed, yPrimed) & (255 << 1);++        xo = Lookup<float>::RandVecs2D[hash];+        yo = Lookup<float>::RandVecs2D[hash | 1];+    }+++    void GradCoordOut(int seed, int xPrimed, int yPrimed, int zPrimed, float& xo, float& yo, float& zo) const+    {+        int hash = Hash(seed, xPrimed, yPrimed, zPrimed) & (255 << 2);++        xo = Lookup<float>::RandVecs3D[hash];+        yo = Lookup<float>::RandVecs3D[hash | 1];+        zo = Lookup<float>::RandVecs3D[hash | 2];+    }+++    void GradCoordDual(int seed, int xPrimed, int yPrimed, float xd, float yd, float& xo, float& yo) const+    {+        int hash = Hash(seed, xPrimed, yPrimed);+        int index1 = hash & (127 << 1);+        int index2 = (hash >> 7) & (255 << 1);++        float xg = Lookup<float>::Gradients2D[index1];+        float yg = Lookup<float>::Gradients2D[index1 | 1];+        float value = xd * xg + yd * yg;++        float xgo = Lookup<float>::RandVecs2D[index2];+        float ygo = Lookup<float>::RandVecs2D[index2 | 1];++        xo = value * xgo;+        yo = value * ygo;+    }+++    void GradCoordDual(int seed, int xPrimed, int yPrimed, int zPrimed, float xd, float yd, float zd, float& xo, float& yo, float& zo) const+    {+        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);+        int index1 = hash & (63 << 2);+        int index2 = (hash >> 6) & (255 << 2);++        float xg = Lookup<float>::Gradients3D[index1];+        float yg = Lookup<float>::Gradients3D[index1 | 1];+        float zg = Lookup<float>::Gradients3D[index1 | 2];+        float value = xd * xg + yd * yg + zd * zg;++        float xgo = Lookup<float>::RandVecs3D[index2];+        float ygo = Lookup<float>::RandVecs3D[index2 | 1];+        float zgo = Lookup<float>::RandVecs3D[index2 | 2];++        xo = value * xgo;+        yo = value * ygo;+        zo = value * zgo;+    }+++    // Generic noise gen++    template <typename FNfloat>+    float GenNoiseSingle(int seed, FNfloat x, FNfloat y) const+    {+        switch (mNoiseType)+        {+        case NoiseType_OpenSimplex2:+            return SingleSimplex(seed, x, y);+        case NoiseType_OpenSimplex2S:+            return SingleOpenSimplex2S(seed, x, y);+        case NoiseType_Cellular:+            return SingleCellular(seed, x, y);+        case NoiseType_Perlin:+            return SinglePerlin(seed, x, y);+        case NoiseType_ValueCubic:+            return SingleValueCubic(seed, x, y);+        case NoiseType_Value:+            return SingleValue(seed, x, y);+        default:+            return 0;+        }+    }++    template <typename FNfloat>+    float GenNoiseSingle(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        switch (mNoiseType)+        {+        case NoiseType_OpenSimplex2:+            return SingleOpenSimplex2(seed, x, y, z);+        case NoiseType_OpenSimplex2S:+            return SingleOpenSimplex2S(seed, x, y, z);+        case NoiseType_Cellular:+            return SingleCellular(seed, x, y, z);+        case NoiseType_Perlin:+            return SinglePerlin(seed, x, y, z);+        case NoiseType_ValueCubic:+            return SingleValueCubic(seed, x, y, z);+        case NoiseType_Value:+            return SingleValue(seed, x, y, z);+        default:+            return 0;+        }+    }+++    // Noise Coordinate Transforms (frequency, and possible skew or rotation)++    template <typename FNfloat>+    void TransformNoiseCoordinate(FNfloat& x, FNfloat& y) const+    {+        x *= mFrequency;+        y *= mFrequency;++        switch (mNoiseType)+        {+        case NoiseType_OpenSimplex2:+        case NoiseType_OpenSimplex2S:+            {+                const FNfloat SQRT3 = (FNfloat)1.7320508075688772935274463415059;+                const FNfloat F2 = 0.5f * (SQRT3 - 1);+                FNfloat t = (x + y) * F2;+                x += t;+                y += t;+            }+            break;+        default:+            break;+        }+    }++    template <typename FNfloat>+    void TransformNoiseCoordinate(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        x *= mFrequency;+        y *= mFrequency;+        z *= mFrequency;++        switch (mTransformType3D)+        {+        case TransformType3D_ImproveXYPlanes:+            {+                FNfloat xy = x + y;+                FNfloat s2 = xy * -(FNfloat)0.211324865405187;+                z *= (FNfloat)0.577350269189626;+                x += s2 - z;+                y = y + s2 - z;+                z += xy * (FNfloat)0.577350269189626;+            }+            break;+        case TransformType3D_ImproveXZPlanes:+            {+                FNfloat xz = x + z;+                FNfloat s2 = xz * -(FNfloat)0.211324865405187;+                y *= (FNfloat)0.577350269189626;+                x += s2 - y;+                z += s2 - y;+                y += xz * (FNfloat)0.577350269189626;+            }+            break;+        case TransformType3D_DefaultOpenSimplex2:+            {+                const FNfloat R3 = (FNfloat)(2.0 / 3.0);+                FNfloat r = (x + y + z) * R3; // Rotation, not skew+                x = r - x;+                y = r - y;+                z = r - z;+            }+            break;+        default:+            break;+        }+    }++    void UpdateTransformType3D()+    {+        switch (mRotationType3D)+        {+        case RotationType3D_ImproveXYPlanes:+            mTransformType3D = TransformType3D_ImproveXYPlanes;+            break;+        case RotationType3D_ImproveXZPlanes:+            mTransformType3D = TransformType3D_ImproveXZPlanes;+            break;+        default:+            switch (mNoiseType)+            {+            case NoiseType_OpenSimplex2:+            case NoiseType_OpenSimplex2S:+                mTransformType3D = TransformType3D_DefaultOpenSimplex2;+                break;+            default:+                mTransformType3D = TransformType3D_None;+                break;+            }+            break;+        }+    }+++    // Domain Warp Coordinate Transforms++    template <typename FNfloat>+    void TransformDomainWarpCoordinate(FNfloat& x, FNfloat& y) const+    {+        switch (mDomainWarpType)+        {+        case DomainWarpType_OpenSimplex2:+        case DomainWarpType_OpenSimplex2Reduced:+            {+                const FNfloat SQRT3 = (FNfloat)1.7320508075688772935274463415059;+                const FNfloat F2 = 0.5f * (SQRT3 - 1);+                FNfloat t = (x + y) * F2;+                x += t;+                y += t;+            }+            break;+        default:+            break;+        }+    }++    template <typename FNfloat>+    void TransformDomainWarpCoordinate(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        switch (mWarpTransformType3D)+        {+        case TransformType3D_ImproveXYPlanes:+            {+                FNfloat xy = x + y;+                FNfloat s2 = xy * -(FNfloat)0.211324865405187;+                z *= (FNfloat)0.577350269189626;+                x += s2 - z;+                y = y + s2 - z;+                z += xy * (FNfloat)0.577350269189626;+            }+            break;+        case TransformType3D_ImproveXZPlanes:+            {+                FNfloat xz = x + z;+                FNfloat s2 = xz * -(FNfloat)0.211324865405187;+                y *= (FNfloat)0.577350269189626;+                x += s2 - y;+                z += s2 - y;+                y += xz * (FNfloat)0.577350269189626;+            }+            break;+        case TransformType3D_DefaultOpenSimplex2:+            {+                const FNfloat R3 = (FNfloat)(2.0 / 3.0);+                FNfloat r = (x + y + z) * R3; // Rotation, not skew+                x = r - x;+                y = r - y;+                z = r - z;+            }+            break;+        default:+            break;+        }+    }++    void UpdateWarpTransformType3D()+    {+        switch (mRotationType3D)+        {+        case RotationType3D_ImproveXYPlanes:+            mWarpTransformType3D = TransformType3D_ImproveXYPlanes;+            break;+        case RotationType3D_ImproveXZPlanes:+            mWarpTransformType3D = TransformType3D_ImproveXZPlanes;+            break;+        default:+            switch (mDomainWarpType)+            {+            case DomainWarpType_OpenSimplex2:+            case DomainWarpType_OpenSimplex2Reduced:+                mWarpTransformType3D = TransformType3D_DefaultOpenSimplex2;+                break;+            default:+                mWarpTransformType3D = TransformType3D_None;+                break;+            }+            break;+        }+    }+++    // Fractal FBm++    template <typename FNfloat>+    float GenFractalFBm(FNfloat x, FNfloat y) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = GenNoiseSingle(seed++, x, y);+            sum += noise * amp;+            amp *= Lerp(1.0f, FastMin(noise + 1, 2) * 0.5f, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }++    template <typename FNfloat>+    float GenFractalFBm(FNfloat x, FNfloat y, FNfloat z) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = GenNoiseSingle(seed++, x, y, z);+            sum += noise * amp;+            amp *= Lerp(1.0f, (noise + 1) * 0.5f, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            z *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }+++    // Fractal Ridged++    template <typename FNfloat>+    float GenFractalRidged(FNfloat x, FNfloat y) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = FastAbs(GenNoiseSingle(seed++, x, y));+            sum += (noise * -2 + 1) * amp;+            amp *= Lerp(1.0f, 1 - noise, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }++    template <typename FNfloat>+    float GenFractalRidged(FNfloat x, FNfloat y, FNfloat z) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = FastAbs(GenNoiseSingle(seed++, x, y, z));+            sum += (noise * -2 + 1) * amp;+            amp *= Lerp(1.0f, 1 - noise, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            z *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }+++    // Fractal PingPong ++    template <typename FNfloat>+    float GenFractalPingPong(FNfloat x, FNfloat y) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = PingPong((GenNoiseSingle(seed++, x, y) + 1) * mPingPongStrength);+            sum += (noise - 0.5f) * 2 * amp;+            amp *= Lerp(1.0f, noise, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }++    template <typename FNfloat>+    float GenFractalPingPong(FNfloat x, FNfloat y, FNfloat z) const+    {+        int seed = mSeed;+        float sum = 0;+        float amp = mFractalBounding;++        for (int i = 0; i < mOctaves; i++)+        {+            float noise = PingPong((GenNoiseSingle(seed++, x, y, z) + 1) * mPingPongStrength);+            sum += (noise - 0.5f) * 2 * amp;+            amp *= Lerp(1.0f, noise, mWeightedStrength);++            x *= mLacunarity;+            y *= mLacunarity;+            z *= mLacunarity;+            amp *= mGain;+        }++        return sum;+    }+++    // Simplex/OpenSimplex2 Noise++    template <typename FNfloat>+    float SingleSimplex(int seed, FNfloat x, FNfloat y) const+    {+        // 2D OpenSimplex2 case uses the same algorithm as ordinary Simplex.++        const float SQRT3 = 1.7320508075688772935274463415059f;+        const float G2 = (3 - SQRT3) / 6;++        /*+         * --- Skew moved to TransformNoiseCoordinate method ---+         * const FNfloat F2 = 0.5f * (SQRT3 - 1);+         * FNfloat s = (x + y) * F2;+         * x += s; y += s;+        */++        int i = FastFloor(x);+        int j = FastFloor(y);+        float xi = (float)(x - i);+        float yi = (float)(y - j);++        float t = (xi + yi) * G2;+        float x0 = (float)(xi - t);+        float y0 = (float)(yi - t);++        i *= PrimeX;+        j *= PrimeY;++        float n0, n1, n2;++        float a = 0.5f - x0 * x0 - y0 * y0;+        if (a <= 0) n0 = 0;+        else+        {+            n0 = (a * a) * (a * a) * GradCoord(seed, i, j, x0, y0);+        }++        float c = (float)(2 * (1 - 2 * G2) * (1 / G2 - 2)) * t + ((float)(-2 * (1 - 2 * G2) * (1 - 2 * G2)) + a);+        if (c <= 0) n2 = 0;+        else+        {+            float x2 = x0 + (2 * (float)G2 - 1);+            float y2 = y0 + (2 * (float)G2 - 1);+            n2 = (c * c) * (c * c) * GradCoord(seed, i + PrimeX, j + PrimeY, x2, y2);+        }++        if (y0 > x0)+        {+            float x1 = x0 + (float)G2;+            float y1 = y0 + ((float)G2 - 1);+            float b = 0.5f - x1 * x1 - y1 * y1;+            if (b <= 0) n1 = 0;+            else+            {+                n1 = (b * b) * (b * b) * GradCoord(seed, i, j + PrimeY, x1, y1);+            }+        }+        else+        {+            float x1 = x0 + ((float)G2 - 1);+            float y1 = y0 + (float)G2;+            float b = 0.5f - x1 * x1 - y1 * y1;+            if (b <= 0) n1 = 0;+            else+            {+                n1 = (b * b) * (b * b) * GradCoord(seed, i + PrimeX, j, x1, y1);+            }+        }++        return (n0 + n1 + n2) * 99.83685446303647f;+    }++    template <typename FNfloat>+    float SingleOpenSimplex2(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        // 3D OpenSimplex2 case uses two offset rotated cube grids.++        /*+         * --- Rotation moved to TransformNoiseCoordinate method ---+         * const FNfloat R3 = (FNfloat)(2.0 / 3.0);+         * FNfloat r = (x + y + z) * R3; // Rotation, not skew+         * x = r - x; y = r - y; z = r - z;+        */++        int i = FastRound(x);+        int j = FastRound(y);+        int k = FastRound(z);+        float x0 = (float)(x - i);+        float y0 = (float)(y - j);+        float z0 = (float)(z - k);++        int xNSign = (int)(-1.0f - x0) | 1;+        int yNSign = (int)(-1.0f - y0) | 1;+        int zNSign = (int)(-1.0f - z0) | 1;++        float ax0 = xNSign * -x0;+        float ay0 = yNSign * -y0;+        float az0 = zNSign * -z0;++        i *= PrimeX;+        j *= PrimeY;+        k *= PrimeZ;++        float value = 0;+        float a = (0.6f - x0 * x0) - (y0 * y0 + z0 * z0);++        for (int l = 0; ; l++)+        {+            if (a > 0)+            {+                value += (a * a) * (a * a) * GradCoord(seed, i, j, k, x0, y0, z0);+            }++            float b = a + 1;+            int i1 = i;+            int j1 = j;+            int k1 = k;+            float x1 = x0;+            float y1 = y0;+            float z1 = z0;++            if (ax0 >= ay0 && ax0 >= az0)+            {+                x1 += xNSign;+                b -= xNSign * 2 * x1;+                i1 -= xNSign * PrimeX;+            }+            else if (ay0 > ax0 && ay0 >= az0)+            {+                y1 += yNSign;+                b -= yNSign * 2 * y1;+                j1 -= yNSign * PrimeY;+            }+            else+            {+                z1 += zNSign;+                b -= zNSign * 2 * z1;+                k1 -= zNSign * PrimeZ;+            }++            if (b > 0)+            {+                value += (b * b) * (b * b) * GradCoord(seed, i1, j1, k1, x1, y1, z1);+            }++            if (l == 1) break;++            ax0 = 0.5f - ax0;+            ay0 = 0.5f - ay0;+            az0 = 0.5f - az0;++            x0 = xNSign * ax0;+            y0 = yNSign * ay0;+            z0 = zNSign * az0;++            a += (0.75f - ax0) - (ay0 + az0);++            i += (xNSign >> 1) & PrimeX;+            j += (yNSign >> 1) & PrimeY;+            k += (zNSign >> 1) & PrimeZ;++            xNSign = -xNSign;+            yNSign = -yNSign;+            zNSign = -zNSign;++            seed = ~seed;+        }++        return value * 32.69428253173828125f;+    }+++    // OpenSimplex2S Noise++    template <typename FNfloat>+    float SingleOpenSimplex2S(int seed, FNfloat x, FNfloat y) const+    {+        // 2D OpenSimplex2S case is a modified 2D simplex noise.++        const FNfloat SQRT3 = (FNfloat)1.7320508075688772935274463415059;+        const FNfloat G2 = (3 - SQRT3) / 6;++        /*+         * --- Skew moved to TransformNoiseCoordinate method ---+         * const FNfloat F2 = 0.5f * (SQRT3 - 1);+         * FNfloat s = (x + y) * F2;+         * x += s; y += s;+        */++        int i = FastFloor(x);+        int j = FastFloor(y);+        float xi = (float)(x - i);+        float yi = (float)(y - j);++        i *= PrimeX;+        j *= PrimeY;+        int i1 = i + PrimeX;+        int j1 = j + PrimeY;++        float t = (xi + yi) * (float)G2;+        float x0 = xi - t;+        float y0 = yi - t;++        float a0 = (2.0f / 3.0f) - x0 * x0 - y0 * y0;+        float value = (a0 * a0) * (a0 * a0) * GradCoord(seed, i, j, x0, y0);++        float a1 = (float)(2 * (1 - 2 * G2) * (1 / G2 - 2)) * t + ((float)(-2 * (1 - 2 * G2) * (1 - 2 * G2)) + a0);+        float x1 = x0 - (float)(1 - 2 * G2);+        float y1 = y0 - (float)(1 - 2 * G2);+        value += (a1 * a1) * (a1 * a1) * GradCoord(seed, i1, j1, x1, y1);++        // Nested conditionals were faster than compact bit logic/arithmetic.+        float xmyi = xi - yi;+        if (t > G2)+        {+            if (xi + xmyi > 1)+            {+                float x2 = x0 + (float)(3 * G2 - 2);+                float y2 = y0 + (float)(3 * G2 - 1);+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i + (PrimeX << 1), j + PrimeY, x2, y2);+                }+            }+            else+            {+                float x2 = x0 + (float)G2;+                float y2 = y0 + (float)(G2 - 1);+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i, j + PrimeY, x2, y2);+                }+            }++            if (yi - xmyi > 1)+            {+                float x3 = x0 + (float)(3 * G2 - 1);+                float y3 = y0 + (float)(3 * G2 - 2);+                float a3 = (2.0f / 3.0f) - x3 * x3 - y3 * y3;+                if (a3 > 0)+                {+                    value += (a3 * a3) * (a3 * a3) * GradCoord(seed, i + PrimeX, j + (PrimeY << 1), x3, y3);+                }+            }+            else+            {+                float x3 = x0 + (float)(G2 - 1);+                float y3 = y0 + (float)G2;+                float a3 = (2.0f / 3.0f) - x3 * x3 - y3 * y3;+                if (a3 > 0)+                {+                    value += (a3 * a3) * (a3 * a3) * GradCoord(seed, i + PrimeX, j, x3, y3);+                }+            }+        }+        else+        {+            if (xi + xmyi < 0)+            {+                float x2 = x0 + (float)(1 - G2);+                float y2 = y0 - (float)G2;+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i - PrimeX, j, x2, y2);+                }+            }+            else+            {+                float x2 = x0 + (float)(G2 - 1);+                float y2 = y0 + (float)G2;+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i + PrimeX, j, x2, y2);+                }+            }++            if (yi < xmyi)+            {+                float x2 = x0 - (float)G2;+                float y2 = y0 - (float)(G2 - 1);+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i, j - PrimeY, x2, y2);+                }+            }+            else+            {+                float x2 = x0 + (float)G2;+                float y2 = y0 + (float)(G2 - 1);+                float a2 = (2.0f / 3.0f) - x2 * x2 - y2 * y2;+                if (a2 > 0)+                {+                    value += (a2 * a2) * (a2 * a2) * GradCoord(seed, i, j + PrimeY, x2, y2);+                }+            }+        }++        return value * 18.24196194486065f;+    }++    template <typename FNfloat>+    float SingleOpenSimplex2S(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        // 3D OpenSimplex2S case uses two offset rotated cube grids.++        /*+         * --- Rotation moved to TransformNoiseCoordinate method ---+         * const FNfloat R3 = (FNfloat)(2.0 / 3.0);+         * FNfloat r = (x + y + z) * R3; // Rotation, not skew+         * x = r - x; y = r - y; z = r - z;+        */++        int i = FastFloor(x);+        int j = FastFloor(y);+        int k = FastFloor(z);+        float xi = (float)(x - i);+        float yi = (float)(y - j);+        float zi = (float)(z - k);++        i *= PrimeX;+        j *= PrimeY;+        k *= PrimeZ;+        int seed2 = seed + 1293373;++        int xNMask = (int)(-0.5f - xi);+        int yNMask = (int)(-0.5f - yi);+        int zNMask = (int)(-0.5f - zi);++        float x0 = xi + xNMask;+        float y0 = yi + yNMask;+        float z0 = zi + zNMask;+        float a0 = 0.75f - x0 * x0 - y0 * y0 - z0 * z0;+        float value = (a0 * a0) * (a0 * a0) * GradCoord(seed,+                                                        i + (xNMask & PrimeX), j + (yNMask & PrimeY), k + (zNMask & PrimeZ), x0, y0, z0);++        float x1 = xi - 0.5f;+        float y1 = yi - 0.5f;+        float z1 = zi - 0.5f;+        float a1 = 0.75f - x1 * x1 - y1 * y1 - z1 * z1;+        value += (a1 * a1) * (a1 * a1) * GradCoord(seed2,+                                                   i + PrimeX, j + PrimeY, k + PrimeZ, x1, y1, z1);++        float xAFlipMask0 = ((xNMask | 1) << 1) * x1;+        float yAFlipMask0 = ((yNMask | 1) << 1) * y1;+        float zAFlipMask0 = ((zNMask | 1) << 1) * z1;+        float xAFlipMask1 = (-2 - (xNMask << 2)) * x1 - 1.0f;+        float yAFlipMask1 = (-2 - (yNMask << 2)) * y1 - 1.0f;+        float zAFlipMask1 = (-2 - (zNMask << 2)) * z1 - 1.0f;++        bool skip5 = false;+        float a2 = xAFlipMask0 + a0;+        if (a2 > 0)+        {+            float x2 = x0 - (xNMask | 1);+            float y2 = y0;+            float z2 = z0;+            value += (a2 * a2) * (a2 * a2) * GradCoord(seed,+                                                       i + (~xNMask & PrimeX), j + (yNMask & PrimeY), k + (zNMask & PrimeZ), x2, y2, z2);+        }+        else+        {+            float a3 = yAFlipMask0 + zAFlipMask0 + a0;+            if (a3 > 0)+            {+                float x3 = x0;+                float y3 = y0 - (yNMask | 1);+                float z3 = z0 - (zNMask | 1);+                value += (a3 * a3) * (a3 * a3) * GradCoord(seed,+                                                           i + (xNMask & PrimeX), j + (~yNMask & PrimeY), k + (~zNMask & PrimeZ), x3, y3, z3);+            }++            float a4 = xAFlipMask1 + a1;+            if (a4 > 0)+            {+                float x4 = (xNMask | 1) + x1;+                float y4 = y1;+                float z4 = z1;+                value += (a4 * a4) * (a4 * a4) * GradCoord(seed2,+                                                           i + (xNMask & (PrimeX * 2)), j + PrimeY, k + PrimeZ, x4, y4, z4);+                skip5 = true;+            }+        }++        bool skip9 = false;+        float a6 = yAFlipMask0 + a0;+        if (a6 > 0)+        {+            float x6 = x0;+            float y6 = y0 - (yNMask | 1);+            float z6 = z0;+            value += (a6 * a6) * (a6 * a6) * GradCoord(seed,+                                                       i + (xNMask & PrimeX), j + (~yNMask & PrimeY), k + (zNMask & PrimeZ), x6, y6, z6);+        }+        else+        {+            float a7 = xAFlipMask0 + zAFlipMask0 + a0;+            if (a7 > 0)+            {+                float x7 = x0 - (xNMask | 1);+                float y7 = y0;+                float z7 = z0 - (zNMask | 1);+                value += (a7 * a7) * (a7 * a7) * GradCoord(seed,+                                                           i + (~xNMask & PrimeX), j + (yNMask & PrimeY), k + (~zNMask & PrimeZ), x7, y7, z7);+            }++            float a8 = yAFlipMask1 + a1;+            if (a8 > 0)+            {+                float x8 = x1;+                float y8 = (yNMask | 1) + y1;+                float z8 = z1;+                value += (a8 * a8) * (a8 * a8) * GradCoord(seed2,+                                                           i + PrimeX, j + (yNMask & (PrimeY << 1)), k + PrimeZ, x8, y8, z8);+                skip9 = true;+            }+        }++        bool skipD = false;+        float aA = zAFlipMask0 + a0;+        if (aA > 0)+        {+            float xA = x0;+            float yA = y0;+            float zA = z0 - (zNMask | 1);+            value += (aA * aA) * (aA * aA) * GradCoord(seed,+                                                       i + (xNMask & PrimeX), j + (yNMask & PrimeY), k + (~zNMask & PrimeZ), xA, yA, zA);+        }+        else+        {+            float aB = xAFlipMask0 + yAFlipMask0 + a0;+            if (aB > 0)+            {+                float xB = x0 - (xNMask | 1);+                float yB = y0 - (yNMask | 1);+                float zB = z0;+                value += (aB * aB) * (aB * aB) * GradCoord(seed,+                                                           i + (~xNMask & PrimeX), j + (~yNMask & PrimeY), k + (zNMask & PrimeZ), xB, yB, zB);+            }++            float aC = zAFlipMask1 + a1;+            if (aC > 0)+            {+                float xC = x1;+                float yC = y1;+                float zC = (zNMask | 1) + z1;+                value += (aC * aC) * (aC * aC) * GradCoord(seed2,+                                                           i + PrimeX, j + PrimeY, k + (zNMask & (PrimeZ << 1)), xC, yC, zC);+                skipD = true;+            }+        }++        if (!skip5)+        {+            float a5 = yAFlipMask1 + zAFlipMask1 + a1;+            if (a5 > 0)+            {+                float x5 = x1;+                float y5 = (yNMask | 1) + y1;+                float z5 = (zNMask | 1) + z1;+                value += (a5 * a5) * (a5 * a5) * GradCoord(seed2,+                                                           i + PrimeX, j + (yNMask & (PrimeY << 1)), k + (zNMask & (PrimeZ << 1)), x5, y5, z5);+            }+        }++        if (!skip9)+        {+            float a9 = xAFlipMask1 + zAFlipMask1 + a1;+            if (a9 > 0)+            {+                float x9 = (xNMask | 1) + x1;+                float y9 = y1;+                float z9 = (zNMask | 1) + z1;+                value += (a9 * a9) * (a9 * a9) * GradCoord(seed2,+                                                           i + (xNMask & (PrimeX * 2)), j + PrimeY, k + (zNMask & (PrimeZ << 1)), x9, y9, z9);+            }+        }++        if (!skipD)+        {+            float aD = xAFlipMask1 + yAFlipMask1 + a1;+            if (aD > 0)+            {+                float xD = (xNMask | 1) + x1;+                float yD = (yNMask | 1) + y1;+                float zD = z1;+                value += (aD * aD) * (aD * aD) * GradCoord(seed2,+                                                           i + (xNMask & (PrimeX << 1)), j + (yNMask & (PrimeY << 1)), k + PrimeZ, xD, yD, zD);+            }+        }++        return value * 9.046026385208288f;+    }+++    // Cellular Noise++    template <typename FNfloat>+    float SingleCellular(int seed, FNfloat x, FNfloat y) const+    {+        int xr = FastRound(x);+        int yr = FastRound(y);++        float distance0 = 1e10f;+        float distance1 = 1e10f;+        int closestHash = 0;++        float cellularJitter = 0.43701595f * mCellularJitterModifier;++        int xPrimed = (xr - 1) * PrimeX;+        int yPrimedBase = (yr - 1) * PrimeY;++        switch (mCellularDistanceFunction)+        {+        default:+        case CellularDistanceFunction_Euclidean:+        case CellularDistanceFunction_EuclideanSq:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int hash = Hash(seed, xPrimed, yPrimed);+                    int idx = hash & (255 << 1);++                    float vecX = (float)(xi - x) + Lookup<float>::RandVecs2D[idx] * cellularJitter;+                    float vecY = (float)(yi - y) + Lookup<float>::RandVecs2D[idx | 1] * cellularJitter;++                    float newDistance = vecX * vecX + vecY * vecY;++                    distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                    if (newDistance < distance0)+                    {+                        distance0 = newDistance;+                        closestHash = hash;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        case CellularDistanceFunction_Manhattan:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int hash = Hash(seed, xPrimed, yPrimed);+                    int idx = hash & (255 << 1);++                    float vecX = (float)(xi - x) + Lookup<float>::RandVecs2D[idx] * cellularJitter;+                    float vecY = (float)(yi - y) + Lookup<float>::RandVecs2D[idx | 1] * cellularJitter;++                    float newDistance = FastAbs(vecX) + FastAbs(vecY);++                    distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                    if (newDistance < distance0)+                    {+                        distance0 = newDistance;+                        closestHash = hash;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        case CellularDistanceFunction_Hybrid:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int hash = Hash(seed, xPrimed, yPrimed);+                    int idx = hash & (255 << 1);++                    float vecX = (float)(xi - x) + Lookup<float>::RandVecs2D[idx] * cellularJitter;+                    float vecY = (float)(yi - y) + Lookup<float>::RandVecs2D[idx | 1] * cellularJitter;++                    float newDistance = (FastAbs(vecX) + FastAbs(vecY)) + (vecX * vecX + vecY * vecY);++                    distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                    if (newDistance < distance0)+                    {+                        distance0 = newDistance;+                        closestHash = hash;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        }++        if (mCellularDistanceFunction == CellularDistanceFunction_Euclidean && mCellularReturnType >= CellularReturnType_Distance)+        {+            distance0 = FastSqrt(distance0);++            if (mCellularReturnType >= CellularReturnType_Distance2)+            {+                distance1 = FastSqrt(distance1);+            }+        }++        switch (mCellularReturnType)+        {+        case CellularReturnType_CellValue:+            return closestHash * (1 / 2147483648.0f);+        case CellularReturnType_Distance:+            return distance0 - 1;+        case CellularReturnType_Distance2:+            return distance1 - 1;+        case CellularReturnType_Distance2Add:+            return (distance1 + distance0) * 0.5f - 1;+        case CellularReturnType_Distance2Sub:+            return distance1 - distance0 - 1;+        case CellularReturnType_Distance2Mul:+            return distance1 * distance0 * 0.5f - 1;+        case CellularReturnType_Distance2Div:+            return distance0 / distance1 - 1;+        default:+            return 0;+        }+    }++    template <typename FNfloat>+    float SingleCellular(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        int xr = FastRound(x);+        int yr = FastRound(y);+        int zr = FastRound(z);++        float distance0 = 1e10f;+        float distance1 = 1e10f;+        int closestHash = 0;++        float cellularJitter = 0.39614353f * mCellularJitterModifier;++        int xPrimed = (xr - 1) * PrimeX;+        int yPrimedBase = (yr - 1) * PrimeY;+        int zPrimedBase = (zr - 1) * PrimeZ;++        switch (mCellularDistanceFunction)+        {+        case CellularDistanceFunction_Euclidean:+        case CellularDistanceFunction_EuclideanSq:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int zPrimed = zPrimedBase;++                    for (int zi = zr - 1; zi <= zr + 1; zi++)+                    {+                        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);+                        int idx = hash & (255 << 2);++                        float vecX = (float)(xi - x) + Lookup<float>::RandVecs3D[idx] * cellularJitter;+                        float vecY = (float)(yi - y) + Lookup<float>::RandVecs3D[idx | 1] * cellularJitter;+                        float vecZ = (float)(zi - z) + Lookup<float>::RandVecs3D[idx | 2] * cellularJitter;++                        float newDistance = vecX * vecX + vecY * vecY + vecZ * vecZ;++                        distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                        if (newDistance < distance0)+                        {+                            distance0 = newDistance;+                            closestHash = hash;+                        }+                        zPrimed += PrimeZ;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        case CellularDistanceFunction_Manhattan:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int zPrimed = zPrimedBase;++                    for (int zi = zr - 1; zi <= zr + 1; zi++)+                    {+                        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);+                        int idx = hash & (255 << 2);++                        float vecX = (float)(xi - x) + Lookup<float>::RandVecs3D[idx] * cellularJitter;+                        float vecY = (float)(yi - y) + Lookup<float>::RandVecs3D[idx | 1] * cellularJitter;+                        float vecZ = (float)(zi - z) + Lookup<float>::RandVecs3D[idx | 2] * cellularJitter;++                        float newDistance = FastAbs(vecX) + FastAbs(vecY) + FastAbs(vecZ);++                        distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                        if (newDistance < distance0)+                        {+                            distance0 = newDistance;+                            closestHash = hash;+                        }+                        zPrimed += PrimeZ;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        case CellularDistanceFunction_Hybrid:+            for (int xi = xr - 1; xi <= xr + 1; xi++)+            {+                int yPrimed = yPrimedBase;++                for (int yi = yr - 1; yi <= yr + 1; yi++)+                {+                    int zPrimed = zPrimedBase;++                    for (int zi = zr - 1; zi <= zr + 1; zi++)+                    {+                        int hash = Hash(seed, xPrimed, yPrimed, zPrimed);+                        int idx = hash & (255 << 2);++                        float vecX = (float)(xi - x) + Lookup<float>::RandVecs3D[idx] * cellularJitter;+                        float vecY = (float)(yi - y) + Lookup<float>::RandVecs3D[idx | 1] * cellularJitter;+                        float vecZ = (float)(zi - z) + Lookup<float>::RandVecs3D[idx | 2] * cellularJitter;++                        float newDistance = (FastAbs(vecX) + FastAbs(vecY) + FastAbs(vecZ)) + (vecX * vecX + vecY * vecY + vecZ * vecZ);++                        distance1 = FastMax(FastMin(distance1, newDistance), distance0);+                        if (newDistance < distance0)+                        {+                            distance0 = newDistance;+                            closestHash = hash;+                        }+                        zPrimed += PrimeZ;+                    }+                    yPrimed += PrimeY;+                }+                xPrimed += PrimeX;+            }+            break;+        default:+            break;+        }++        if (mCellularDistanceFunction == CellularDistanceFunction_Euclidean && mCellularReturnType >= CellularReturnType_Distance)+        {+            distance0 = FastSqrt(distance0);++            if (mCellularReturnType >= CellularReturnType_Distance2)+            {+                distance1 = FastSqrt(distance1);+            }+        }++        switch (mCellularReturnType)+        {+        case CellularReturnType_CellValue:+            return closestHash * (1 / 2147483648.0f);+        case CellularReturnType_Distance:+            return distance0 - 1;+        case CellularReturnType_Distance2:+            return distance1 - 1;+        case CellularReturnType_Distance2Add:+            return (distance1 + distance0) * 0.5f - 1;+        case CellularReturnType_Distance2Sub:+            return distance1 - distance0 - 1;+        case CellularReturnType_Distance2Mul:+            return distance1 * distance0 * 0.5f - 1;+        case CellularReturnType_Distance2Div:+            return distance0 / distance1 - 1;+        default:+            return 0;+        }+    }+++    // Perlin Noise++    template <typename FNfloat>+    float SinglePerlin(int seed, FNfloat x, FNfloat y) const+    {+        int x0 = FastFloor(x);+        int y0 = FastFloor(y);++        float xd0 = (float)(x - x0);+        float yd0 = (float)(y - y0);+        float xd1 = xd0 - 1;+        float yd1 = yd0 - 1;++        float xs = InterpQuintic(xd0);+        float ys = InterpQuintic(yd0);++        x0 *= PrimeX;+        y0 *= PrimeY;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;++        float xf0 = Lerp(GradCoord(seed, x0, y0, xd0, yd0), GradCoord(seed, x1, y0, xd1, yd0), xs);+        float xf1 = Lerp(GradCoord(seed, x0, y1, xd0, yd1), GradCoord(seed, x1, y1, xd1, yd1), xs);++        return Lerp(xf0, xf1, ys) * 1.4247691104677813f;+    }++    template <typename FNfloat>+    float SinglePerlin(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        int x0 = FastFloor(x);+        int y0 = FastFloor(y);+        int z0 = FastFloor(z);++        float xd0 = (float)(x - x0);+        float yd0 = (float)(y - y0);+        float zd0 = (float)(z - z0);+        float xd1 = xd0 - 1;+        float yd1 = yd0 - 1;+        float zd1 = zd0 - 1;++        float xs = InterpQuintic(xd0);+        float ys = InterpQuintic(yd0);+        float zs = InterpQuintic(zd0);++        x0 *= PrimeX;+        y0 *= PrimeY;+        z0 *= PrimeZ;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;+        int z1 = z0 + PrimeZ;++        float xf00 = Lerp(GradCoord(seed, x0, y0, z0, xd0, yd0, zd0), GradCoord(seed, x1, y0, z0, xd1, yd0, zd0), xs);+        float xf10 = Lerp(GradCoord(seed, x0, y1, z0, xd0, yd1, zd0), GradCoord(seed, x1, y1, z0, xd1, yd1, zd0), xs);+        float xf01 = Lerp(GradCoord(seed, x0, y0, z1, xd0, yd0, zd1), GradCoord(seed, x1, y0, z1, xd1, yd0, zd1), xs);+        float xf11 = Lerp(GradCoord(seed, x0, y1, z1, xd0, yd1, zd1), GradCoord(seed, x1, y1, z1, xd1, yd1, zd1), xs);++        float yf0 = Lerp(xf00, xf10, ys);+        float yf1 = Lerp(xf01, xf11, ys);++        return Lerp(yf0, yf1, zs) * 0.964921414852142333984375f;+    }+++    // Value Cubic Noise++    template <typename FNfloat>+    float SingleValueCubic(int seed, FNfloat x, FNfloat y) const+    {+        int x1 = FastFloor(x);+        int y1 = FastFloor(y);++        float xs = (float)(x - x1);+        float ys = (float)(y - y1);++        x1 *= PrimeX;+        y1 *= PrimeY;+        int x0 = x1 - PrimeX;+        int y0 = y1 - PrimeY;+        int x2 = x1 + PrimeX;+        int y2 = y1 + PrimeY;+        int x3 = x1 + (int)((long)PrimeX << 1);+        int y3 = y1 + (int)((long)PrimeY << 1);++        return CubicLerp(+            CubicLerp(ValCoord(seed, x0, y0), ValCoord(seed, x1, y0), ValCoord(seed, x2, y0), ValCoord(seed, x3, y0),+                      xs),+            CubicLerp(ValCoord(seed, x0, y1), ValCoord(seed, x1, y1), ValCoord(seed, x2, y1), ValCoord(seed, x3, y1),+                      xs),+            CubicLerp(ValCoord(seed, x0, y2), ValCoord(seed, x1, y2), ValCoord(seed, x2, y2), ValCoord(seed, x3, y2),+                      xs),+            CubicLerp(ValCoord(seed, x0, y3), ValCoord(seed, x1, y3), ValCoord(seed, x2, y3), ValCoord(seed, x3, y3),+                      xs),+            ys) * (1 / (1.5f * 1.5f));+    }++    template <typename FNfloat>+    float SingleValueCubic(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        int x1 = FastFloor(x);+        int y1 = FastFloor(y);+        int z1 = FastFloor(z);++        float xs = (float)(x - x1);+        float ys = (float)(y - y1);+        float zs = (float)(z - z1);++        x1 *= PrimeX;+        y1 *= PrimeY;+        z1 *= PrimeZ;++        int x0 = x1 - PrimeX;+        int y0 = y1 - PrimeY;+        int z0 = z1 - PrimeZ;+        int x2 = x1 + PrimeX;+        int y2 = y1 + PrimeY;+        int z2 = z1 + PrimeZ;+        int x3 = x1 + (int)((long)PrimeX << 1);+        int y3 = y1 + (int)((long)PrimeY << 1);+        int z3 = z1 + (int)((long)PrimeZ << 1);+++        return CubicLerp(+            CubicLerp(+                CubicLerp(ValCoord(seed, x0, y0, z0), ValCoord(seed, x1, y0, z0), ValCoord(seed, x2, y0, z0), ValCoord(seed, x3, y0, z0), xs),+                CubicLerp(ValCoord(seed, x0, y1, z0), ValCoord(seed, x1, y1, z0), ValCoord(seed, x2, y1, z0), ValCoord(seed, x3, y1, z0), xs),+                CubicLerp(ValCoord(seed, x0, y2, z0), ValCoord(seed, x1, y2, z0), ValCoord(seed, x2, y2, z0), ValCoord(seed, x3, y2, z0), xs),+                CubicLerp(ValCoord(seed, x0, y3, z0), ValCoord(seed, x1, y3, z0), ValCoord(seed, x2, y3, z0), ValCoord(seed, x3, y3, z0), xs),+                ys),+            CubicLerp(+                CubicLerp(ValCoord(seed, x0, y0, z1), ValCoord(seed, x1, y0, z1), ValCoord(seed, x2, y0, z1), ValCoord(seed, x3, y0, z1), xs),+                CubicLerp(ValCoord(seed, x0, y1, z1), ValCoord(seed, x1, y1, z1), ValCoord(seed, x2, y1, z1), ValCoord(seed, x3, y1, z1), xs),+                CubicLerp(ValCoord(seed, x0, y2, z1), ValCoord(seed, x1, y2, z1), ValCoord(seed, x2, y2, z1), ValCoord(seed, x3, y2, z1), xs),+                CubicLerp(ValCoord(seed, x0, y3, z1), ValCoord(seed, x1, y3, z1), ValCoord(seed, x2, y3, z1), ValCoord(seed, x3, y3, z1), xs),+                ys),+            CubicLerp(+                CubicLerp(ValCoord(seed, x0, y0, z2), ValCoord(seed, x1, y0, z2), ValCoord(seed, x2, y0, z2), ValCoord(seed, x3, y0, z2), xs),+                CubicLerp(ValCoord(seed, x0, y1, z2), ValCoord(seed, x1, y1, z2), ValCoord(seed, x2, y1, z2), ValCoord(seed, x3, y1, z2), xs),+                CubicLerp(ValCoord(seed, x0, y2, z2), ValCoord(seed, x1, y2, z2), ValCoord(seed, x2, y2, z2), ValCoord(seed, x3, y2, z2), xs),+                CubicLerp(ValCoord(seed, x0, y3, z2), ValCoord(seed, x1, y3, z2), ValCoord(seed, x2, y3, z2), ValCoord(seed, x3, y3, z2), xs),+                ys),+            CubicLerp(+                CubicLerp(ValCoord(seed, x0, y0, z3), ValCoord(seed, x1, y0, z3), ValCoord(seed, x2, y0, z3), ValCoord(seed, x3, y0, z3), xs),+                CubicLerp(ValCoord(seed, x0, y1, z3), ValCoord(seed, x1, y1, z3), ValCoord(seed, x2, y1, z3), ValCoord(seed, x3, y1, z3), xs),+                CubicLerp(ValCoord(seed, x0, y2, z3), ValCoord(seed, x1, y2, z3), ValCoord(seed, x2, y2, z3), ValCoord(seed, x3, y2, z3), xs),+                CubicLerp(ValCoord(seed, x0, y3, z3), ValCoord(seed, x1, y3, z3), ValCoord(seed, x2, y3, z3), ValCoord(seed, x3, y3, z3), xs),+                ys),+            zs) * (1 / (1.5f * 1.5f * 1.5f));+    }+++    // Value Noise++    template <typename FNfloat>+    float SingleValue(int seed, FNfloat x, FNfloat y) const+    {+        int x0 = FastFloor(x);+        int y0 = FastFloor(y);++        float xs = InterpHermite((float)(x - x0));+        float ys = InterpHermite((float)(y - y0));++        x0 *= PrimeX;+        y0 *= PrimeY;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;++        float xf0 = Lerp(ValCoord(seed, x0, y0), ValCoord(seed, x1, y0), xs);+        float xf1 = Lerp(ValCoord(seed, x0, y1), ValCoord(seed, x1, y1), xs);++        return Lerp(xf0, xf1, ys);+    }++    template <typename FNfloat>+    float SingleValue(int seed, FNfloat x, FNfloat y, FNfloat z) const+    {+        int x0 = FastFloor(x);+        int y0 = FastFloor(y);+        int z0 = FastFloor(z);++        float xs = InterpHermite((float)(x - x0));+        float ys = InterpHermite((float)(y - y0));+        float zs = InterpHermite((float)(z - z0));++        x0 *= PrimeX;+        y0 *= PrimeY;+        z0 *= PrimeZ;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;+        int z1 = z0 + PrimeZ;++        float xf00 = Lerp(ValCoord(seed, x0, y0, z0), ValCoord(seed, x1, y0, z0), xs);+        float xf10 = Lerp(ValCoord(seed, x0, y1, z0), ValCoord(seed, x1, y1, z0), xs);+        float xf01 = Lerp(ValCoord(seed, x0, y0, z1), ValCoord(seed, x1, y0, z1), xs);+        float xf11 = Lerp(ValCoord(seed, x0, y1, z1), ValCoord(seed, x1, y1, z1), xs);++        float yf0 = Lerp(xf00, xf10, ys);+        float yf1 = Lerp(xf01, xf11, ys);++        return Lerp(yf0, yf1, zs);+    }+++    // Domain Warp++    template <typename FNfloat>+    void DoSingleDomainWarp(int seed, float amp, float freq, FNfloat x, FNfloat y, FNfloat& xr, FNfloat& yr) const+    {+        switch (mDomainWarpType)+        {+        case DomainWarpType_OpenSimplex2:+            SingleDomainWarpSimplexGradient(seed, amp * 38.283687591552734375f, freq, x, y, xr, yr, false);+            break;+        case DomainWarpType_OpenSimplex2Reduced:+            SingleDomainWarpSimplexGradient(seed, amp * 16.0f, freq, x, y, xr, yr, true);+            break;+        case DomainWarpType_BasicGrid:+            SingleDomainWarpBasicGrid(seed, amp, freq, x, y, xr, yr);+            break;+        }+    }++    template <typename FNfloat>+    void DoSingleDomainWarp(int seed, float amp, float freq, FNfloat x, FNfloat y, FNfloat z, FNfloat& xr, FNfloat& yr, FNfloat& zr) const+    {+        switch (mDomainWarpType)+        {+        case DomainWarpType_OpenSimplex2:+            SingleDomainWarpOpenSimplex2Gradient(seed, amp * 32.69428253173828125f, freq, x, y, z, xr, yr, zr, false);+            break;+        case DomainWarpType_OpenSimplex2Reduced:+            SingleDomainWarpOpenSimplex2Gradient(seed, amp * 7.71604938271605f, freq, x, y, z, xr, yr, zr, true);+            break;+        case DomainWarpType_BasicGrid:+            SingleDomainWarpBasicGrid(seed, amp, freq, x, y, z, xr, yr, zr);+            break;+        }+    }+++    // Domain Warp Single Wrapper++    template <typename FNfloat>+    void DomainWarpSingle(FNfloat& x, FNfloat& y) const+    {+        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        FNfloat xs = x;+        FNfloat ys = y;+        TransformDomainWarpCoordinate(xs, ys);++        DoSingleDomainWarp(seed, amp, freq, xs, ys, x, y);+    }++    template <typename FNfloat>+    void DomainWarpSingle(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        FNfloat xs = x;+        FNfloat ys = y;+        FNfloat zs = z;+        TransformDomainWarpCoordinate(xs, ys, zs);++        DoSingleDomainWarp(seed, amp, freq, xs, ys, zs, x, y, z);+    }+++    // Domain Warp Fractal Progressive++    template <typename FNfloat>+    void DomainWarpFractalProgressive(FNfloat& x, FNfloat& y) const+    {+        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        for (int i = 0; i < mOctaves; i++)+        {+            FNfloat xs = x;+            FNfloat ys = y;+            TransformDomainWarpCoordinate(xs, ys);++            DoSingleDomainWarp(seed, amp, freq, xs, ys, x, y);++            seed++;+            amp *= mGain;+            freq *= mLacunarity;+        }+    }++    template <typename FNfloat>+    void DomainWarpFractalProgressive(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        for (int i = 0; i < mOctaves; i++)+        {+            FNfloat xs = x;+            FNfloat ys = y;+            FNfloat zs = z;+            TransformDomainWarpCoordinate(xs, ys, zs);++            DoSingleDomainWarp(seed, amp, freq, xs, ys, zs, x, y, z);++            seed++;+            amp *= mGain;+            freq *= mLacunarity;+        }+    }+++    // Domain Warp Fractal Independant++    template <typename FNfloat>+    void DomainWarpFractalIndependent(FNfloat& x, FNfloat& y) const+    {+        FNfloat xs = x;+        FNfloat ys = y;+        TransformDomainWarpCoordinate(xs, ys);++        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        for (int i = 0; i < mOctaves; i++)+        {+            DoSingleDomainWarp(seed, amp, freq, xs, ys, x, y);++            seed++;+            amp *= mGain;+            freq *= mLacunarity;+        }+    }++    template <typename FNfloat>+    void DomainWarpFractalIndependent(FNfloat& x, FNfloat& y, FNfloat& z) const+    {+        FNfloat xs = x;+        FNfloat ys = y;+        FNfloat zs = z;+        TransformDomainWarpCoordinate(xs, ys, zs);++        int seed = mSeed;+        float amp = mDomainWarpAmp * mFractalBounding;+        float freq = mFrequency;++        for (int i = 0; i < mOctaves; i++)+        {+            DoSingleDomainWarp(seed, amp, freq, xs, ys, zs, x, y, z);++            seed++;+            amp *= mGain;+            freq *= mLacunarity;+        }+    }+++    // Domain Warp Basic Grid++    template <typename FNfloat>+    void SingleDomainWarpBasicGrid(int seed, float warpAmp, float frequency, FNfloat x, FNfloat y, FNfloat& xr, FNfloat& yr) const+    {+        FNfloat xf = x * frequency;+        FNfloat yf = y * frequency;++        int x0 = FastFloor(xf);+        int y0 = FastFloor(yf);++        float xs = InterpHermite((float)(xf - x0));+        float ys = InterpHermite((float)(yf - y0));++        x0 *= PrimeX;+        y0 *= PrimeY;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;++        int hash0 = Hash(seed, x0, y0) & (255 << 1);+        int hash1 = Hash(seed, x1, y0) & (255 << 1);++        float lx0x = Lerp(Lookup<float>::RandVecs2D[hash0], Lookup<float>::RandVecs2D[hash1], xs);+        float ly0x = Lerp(Lookup<float>::RandVecs2D[hash0 | 1], Lookup<float>::RandVecs2D[hash1 | 1], xs);++        hash0 = Hash(seed, x0, y1) & (255 << 1);+        hash1 = Hash(seed, x1, y1) & (255 << 1);++        float lx1x = Lerp(Lookup<float>::RandVecs2D[hash0], Lookup<float>::RandVecs2D[hash1], xs);+        float ly1x = Lerp(Lookup<float>::RandVecs2D[hash0 | 1], Lookup<float>::RandVecs2D[hash1 | 1], xs);++        xr += Lerp(lx0x, lx1x, ys) * warpAmp;+        yr += Lerp(ly0x, ly1x, ys) * warpAmp;+    }++    template <typename FNfloat>+    void SingleDomainWarpBasicGrid(int seed, float warpAmp, float frequency, FNfloat x, FNfloat y, FNfloat z, FNfloat& xr, FNfloat& yr, FNfloat& zr) const+    {+        FNfloat xf = x * frequency;+        FNfloat yf = y * frequency;+        FNfloat zf = z * frequency;++        int x0 = FastFloor(xf);+        int y0 = FastFloor(yf);+        int z0 = FastFloor(zf);++        float xs = InterpHermite((float)(xf - x0));+        float ys = InterpHermite((float)(yf - y0));+        float zs = InterpHermite((float)(zf - z0));++        x0 *= PrimeX;+        y0 *= PrimeY;+        z0 *= PrimeZ;+        int x1 = x0 + PrimeX;+        int y1 = y0 + PrimeY;+        int z1 = z0 + PrimeZ;++        int hash0 = Hash(seed, x0, y0, z0) & (255 << 2);+        int hash1 = Hash(seed, x1, y0, z0) & (255 << 2);++        float lx0x = Lerp(Lookup<float>::RandVecs3D[hash0], Lookup<float>::RandVecs3D[hash1], xs);+        float ly0x = Lerp(Lookup<float>::RandVecs3D[hash0 | 1], Lookup<float>::RandVecs3D[hash1 | 1], xs);+        float lz0x = Lerp(Lookup<float>::RandVecs3D[hash0 | 2], Lookup<float>::RandVecs3D[hash1 | 2], xs);++        hash0 = Hash(seed, x0, y1, z0) & (255 << 2);+        hash1 = Hash(seed, x1, y1, z0) & (255 << 2);++        float lx1x = Lerp(Lookup<float>::RandVecs3D[hash0], Lookup<float>::RandVecs3D[hash1], xs);+        float ly1x = Lerp(Lookup<float>::RandVecs3D[hash0 | 1], Lookup<float>::RandVecs3D[hash1 | 1], xs);+        float lz1x = Lerp(Lookup<float>::RandVecs3D[hash0 | 2], Lookup<float>::RandVecs3D[hash1 | 2], xs);++        float lx0y = Lerp(lx0x, lx1x, ys);+        float ly0y = Lerp(ly0x, ly1x, ys);+        float lz0y = Lerp(lz0x, lz1x, ys);++        hash0 = Hash(seed, x0, y0, z1) & (255 << 2);+        hash1 = Hash(seed, x1, y0, z1) & (255 << 2);++        lx0x = Lerp(Lookup<float>::RandVecs3D[hash0], Lookup<float>::RandVecs3D[hash1], xs);+        ly0x = Lerp(Lookup<float>::RandVecs3D[hash0 | 1], Lookup<float>::RandVecs3D[hash1 | 1], xs);+        lz0x = Lerp(Lookup<float>::RandVecs3D[hash0 | 2], Lookup<float>::RandVecs3D[hash1 | 2], xs);++        hash0 = Hash(seed, x0, y1, z1) & (255 << 2);+        hash1 = Hash(seed, x1, y1, z1) & (255 << 2);++        lx1x = Lerp(Lookup<float>::RandVecs3D[hash0], Lookup<float>::RandVecs3D[hash1], xs);+        ly1x = Lerp(Lookup<float>::RandVecs3D[hash0 | 1], Lookup<float>::RandVecs3D[hash1 | 1], xs);+        lz1x = Lerp(Lookup<float>::RandVecs3D[hash0 | 2], Lookup<float>::RandVecs3D[hash1 | 2], xs);++        xr += Lerp(lx0y, Lerp(lx0x, lx1x, ys), zs) * warpAmp;+        yr += Lerp(ly0y, Lerp(ly0x, ly1x, ys), zs) * warpAmp;+        zr += Lerp(lz0y, Lerp(lz0x, lz1x, ys), zs) * warpAmp;+    }+++    // Domain Warp Simplex/OpenSimplex2++    template <typename FNfloat>+    void SingleDomainWarpSimplexGradient(int seed, float warpAmp, float frequency, FNfloat x, FNfloat y, FNfloat& xr, FNfloat& yr, bool outGradOnly) const+    {+        const float SQRT3 = 1.7320508075688772935274463415059f;+        const float G2 = (3 - SQRT3) / 6;++        x *= frequency;+        y *= frequency;++        /*+         * --- Skew moved to TransformNoiseCoordinate method ---+         * const FNfloat F2 = 0.5f * (SQRT3 - 1);+         * FNfloat s = (x + y) * F2;+         * x += s; y += s;+        */++        int i = FastFloor(x);+        int j = FastFloor(y);+        float xi = (float)(x - i);+        float yi = (float)(y - j);++        float t = (xi + yi) * G2;+        float x0 = (float)(xi - t);+        float y0 = (float)(yi - t);++        i *= PrimeX;+        j *= PrimeY;++        float vx, vy;+        vx = vy = 0;++        float a = 0.5f - x0 * x0 - y0 * y0;+        if (a > 0)+        {+            float aaaa = (a * a) * (a * a);+            float xo, yo;+            if (outGradOnly)+                GradCoordOut(seed, i, j, xo, yo);+            else+                GradCoordDual(seed, i, j, x0, y0, xo, yo);+            vx += aaaa * xo;+            vy += aaaa * yo;+        }++        float c = (float)(2 * (1 - 2 * G2) * (1 / G2 - 2)) * t + ((float)(-2 * (1 - 2 * G2) * (1 - 2 * G2)) + a);+        if (c > 0)+        {+            float x2 = x0 + (2 * (float)G2 - 1);+            float y2 = y0 + (2 * (float)G2 - 1);+            float cccc = (c * c) * (c * c);+            float xo, yo;+            if (outGradOnly)+                GradCoordOut(seed, i + PrimeX, j + PrimeY, xo, yo);+            else+                GradCoordDual(seed, i + PrimeX, j + PrimeY, x2, y2, xo, yo);+            vx += cccc * xo;+            vy += cccc * yo;+        }++        if (y0 > x0)+        {+            float x1 = x0 + (float)G2;+            float y1 = y0 + ((float)G2 - 1);+            float b = 0.5f - x1 * x1 - y1 * y1;+            if (b > 0)+            {+                float bbbb = (b * b) * (b * b);+                float xo, yo;+                if (outGradOnly)+                    GradCoordOut(seed, i, j + PrimeY, xo, yo);+                else+                    GradCoordDual(seed, i, j + PrimeY, x1, y1, xo, yo);+                vx += bbbb * xo;+                vy += bbbb * yo;+            }+        }+        else+        {+            float x1 = x0 + ((float)G2 - 1);+            float y1 = y0 + (float)G2;+            float b = 0.5f - x1 * x1 - y1 * y1;+            if (b > 0)+            {+                float bbbb = (b * b) * (b * b);+                float xo, yo;+                if (outGradOnly)+                    GradCoordOut(seed, i + PrimeX, j, xo, yo);+                else+                    GradCoordDual(seed, i + PrimeX, j, x1, y1, xo, yo);+                vx += bbbb * xo;+                vy += bbbb * yo;+            }+        }++        xr += vx * warpAmp;+        yr += vy * warpAmp;+    }++    template <typename FNfloat>+    void SingleDomainWarpOpenSimplex2Gradient(int seed, float warpAmp, float frequency, FNfloat x, FNfloat y, FNfloat z, FNfloat& xr, FNfloat& yr, FNfloat& zr, bool outGradOnly) const+    {+        x *= frequency;+        y *= frequency;+        z *= frequency;++        /*+         * --- Rotation moved to TransformDomainWarpCoordinate method ---+         * const FNfloat R3 = (FNfloat)(2.0 / 3.0);+         * FNfloat r = (x + y + z) * R3; // Rotation, not skew+         * x = r - x; y = r - y; z = r - z;+        */++        int i = FastRound(x);+        int j = FastRound(y);+        int k = FastRound(z);+        float x0 = (float)x - i;+        float y0 = (float)y - j;+        float z0 = (float)z - k;++        int xNSign = (int)(-x0 - 1.0f) | 1;+        int yNSign = (int)(-y0 - 1.0f) | 1;+        int zNSign = (int)(-z0 - 1.0f) | 1;++        float ax0 = xNSign * -x0;+        float ay0 = yNSign * -y0;+        float az0 = zNSign * -z0;++        i *= PrimeX;+        j *= PrimeY;+        k *= PrimeZ;++        float vx, vy, vz;+        vx = vy = vz = 0;++        float a = (0.6f - x0 * x0) - (y0 * y0 + z0 * z0);+        for (int l = 0; l < 2; l++)+        {+            if (a > 0)+            {+                float aaaa = (a * a) * (a * a);+                float xo, yo, zo;+                if (outGradOnly)+                    GradCoordOut(seed, i, j, k, xo, yo, zo);+                else+                    GradCoordDual(seed, i, j, k, x0, y0, z0, xo, yo, zo);+                vx += aaaa * xo;+                vy += aaaa * yo;+                vz += aaaa * zo;+            }++            float b = a + 1;+            int i1 = i;+            int j1 = j;+            int k1 = k;+            float x1 = x0;+            float y1 = y0;+            float z1 = z0;++            if (ax0 >= ay0 && ax0 >= az0)+            {+                x1 += xNSign;+                b -= xNSign * 2 * x1;+                i1 -= xNSign * PrimeX;+            }+            else if (ay0 > ax0 && ay0 >= az0)+            {+                y1 += yNSign;+                b -= yNSign * 2 * y1;+                j1 -= yNSign * PrimeY;+            }+            else+            {+                z1 += zNSign;+                b -= zNSign * 2 * z1;+                k1 -= zNSign * PrimeZ;+            }++            if (b > 0)+            {+                float bbbb = (b * b) * (b * b);+                float xo, yo, zo;+                if (outGradOnly)+                    GradCoordOut(seed, i1, j1, k1, xo, yo, zo);+                else+                    GradCoordDual(seed, i1, j1, k1, x1, y1, z1, xo, yo, zo);+                vx += bbbb * xo;+                vy += bbbb * yo;+                vz += bbbb * zo;+            }++            if (l == 1) break;++            ax0 = 0.5f - ax0;+            ay0 = 0.5f - ay0;+            az0 = 0.5f - az0;++            x0 = xNSign * ax0;+            y0 = yNSign * ay0;+            z0 = zNSign * az0;++            a += (0.75f - ax0) - (ay0 + az0);++            i += (xNSign >> 1) & PrimeX;+            j += (yNSign >> 1) & PrimeY;+            k += (zNSign >> 1) & PrimeZ;++            xNSign = -xNSign;+            yNSign = -yNSign;+            zNSign = -zNSign;++            seed += 1293373;+        }++        xr += vx * warpAmp;+        yr += vy * warpAmp;+        zr += vz * warpAmp;+    }+};++template <>+struct FastNoiseLite::Arguments_must_be_floating_point_values<float> {};+template <>+struct FastNoiseLite::Arguments_must_be_floating_point_values<double> {};+template <>+struct FastNoiseLite::Arguments_must_be_floating_point_values<long double> {};++template <typename T>+const T FastNoiseLite::Lookup<T>::Gradients2D[] =+{+    0.130526192220052f, 0.99144486137381f, 0.38268343236509f, 0.923879532511287f, 0.608761429008721f, 0.793353340291235f, 0.793353340291235f, 0.608761429008721f,+    0.923879532511287f, 0.38268343236509f, 0.99144486137381f, 0.130526192220051f, 0.99144486137381f, -0.130526192220051f, 0.923879532511287f, -0.38268343236509f,+    0.793353340291235f, -0.60876142900872f, 0.608761429008721f, -0.793353340291235f, 0.38268343236509f, -0.923879532511287f, 0.130526192220052f, -0.99144486137381f,+    -0.130526192220052f, -0.99144486137381f, -0.38268343236509f, -0.923879532511287f, -0.608761429008721f, -0.793353340291235f, -0.793353340291235f, -0.608761429008721f,+    -0.923879532511287f, -0.38268343236509f, -0.99144486137381f, -0.130526192220052f, -0.99144486137381f, 0.130526192220051f, -0.923879532511287f, 0.38268343236509f,+    -0.793353340291235f, 0.608761429008721f, -0.608761429008721f, 0.793353340291235f, -0.38268343236509f, 0.923879532511287f, -0.130526192220052f, 0.99144486137381f,+    0.130526192220052f, 0.99144486137381f, 0.38268343236509f, 0.923879532511287f, 0.608761429008721f, 0.793353340291235f, 0.793353340291235f, 0.608761429008721f,+    0.923879532511287f, 0.38268343236509f, 0.99144486137381f, 0.130526192220051f, 0.99144486137381f, -0.130526192220051f, 0.923879532511287f, -0.38268343236509f,+    0.793353340291235f, -0.60876142900872f, 0.608761429008721f, -0.793353340291235f, 0.38268343236509f, -0.923879532511287f, 0.130526192220052f, -0.99144486137381f,+    -0.130526192220052f, -0.99144486137381f, -0.38268343236509f, -0.923879532511287f, -0.608761429008721f, -0.793353340291235f, -0.793353340291235f, -0.608761429008721f,+    -0.923879532511287f, -0.38268343236509f, -0.99144486137381f, -0.130526192220052f, -0.99144486137381f, 0.130526192220051f, -0.923879532511287f, 0.38268343236509f,+    -0.793353340291235f, 0.608761429008721f, -0.608761429008721f, 0.793353340291235f, -0.38268343236509f, 0.923879532511287f, -0.130526192220052f, 0.99144486137381f,+    0.130526192220052f, 0.99144486137381f, 0.38268343236509f, 0.923879532511287f, 0.608761429008721f, 0.793353340291235f, 0.793353340291235f, 0.608761429008721f,+    0.923879532511287f, 0.38268343236509f, 0.99144486137381f, 0.130526192220051f, 0.99144486137381f, -0.130526192220051f, 0.923879532511287f, -0.38268343236509f,+    0.793353340291235f, -0.60876142900872f, 0.608761429008721f, -0.793353340291235f, 0.38268343236509f, -0.923879532511287f, 0.130526192220052f, -0.99144486137381f,+    -0.130526192220052f, -0.99144486137381f, -0.38268343236509f, -0.923879532511287f, -0.608761429008721f, -0.793353340291235f, -0.793353340291235f, -0.608761429008721f,+    -0.923879532511287f, -0.38268343236509f, -0.99144486137381f, -0.130526192220052f, -0.99144486137381f, 0.130526192220051f, -0.923879532511287f, 0.38268343236509f,+    -0.793353340291235f, 0.608761429008721f, -0.608761429008721f, 0.793353340291235f, -0.38268343236509f, 0.923879532511287f, -0.130526192220052f, 0.99144486137381f,+    0.130526192220052f, 0.99144486137381f, 0.38268343236509f, 0.923879532511287f, 0.608761429008721f, 0.793353340291235f, 0.793353340291235f, 0.608761429008721f,+    0.923879532511287f, 0.38268343236509f, 0.99144486137381f, 0.130526192220051f, 0.99144486137381f, -0.130526192220051f, 0.923879532511287f, -0.38268343236509f,+    0.793353340291235f, -0.60876142900872f, 0.608761429008721f, -0.793353340291235f, 0.38268343236509f, -0.923879532511287f, 0.130526192220052f, -0.99144486137381f,+    -0.130526192220052f, -0.99144486137381f, -0.38268343236509f, -0.923879532511287f, -0.608761429008721f, -0.793353340291235f, -0.793353340291235f, -0.608761429008721f,+    -0.923879532511287f, -0.38268343236509f, -0.99144486137381f, -0.130526192220052f, -0.99144486137381f, 0.130526192220051f, -0.923879532511287f, 0.38268343236509f,+    -0.793353340291235f, 0.608761429008721f, -0.608761429008721f, 0.793353340291235f, -0.38268343236509f, 0.923879532511287f, -0.130526192220052f, 0.99144486137381f,+    0.130526192220052f, 0.99144486137381f, 0.38268343236509f, 0.923879532511287f, 0.608761429008721f, 0.793353340291235f, 0.793353340291235f, 0.608761429008721f,+    0.923879532511287f, 0.38268343236509f, 0.99144486137381f, 0.130526192220051f, 0.99144486137381f, -0.130526192220051f, 0.923879532511287f, -0.38268343236509f,+    0.793353340291235f, -0.60876142900872f, 0.608761429008721f, -0.793353340291235f, 0.38268343236509f, -0.923879532511287f, 0.130526192220052f, -0.99144486137381f,+    -0.130526192220052f, -0.99144486137381f, -0.38268343236509f, -0.923879532511287f, -0.608761429008721f, -0.793353340291235f, -0.793353340291235f, -0.608761429008721f,+    -0.923879532511287f, -0.38268343236509f, -0.99144486137381f, -0.130526192220052f, -0.99144486137381f, 0.130526192220051f, -0.923879532511287f, 0.38268343236509f,+    -0.793353340291235f, 0.608761429008721f, -0.608761429008721f, 0.793353340291235f, -0.38268343236509f, 0.923879532511287f, -0.130526192220052f, 0.99144486137381f,+    0.38268343236509f, 0.923879532511287f, 0.923879532511287f, 0.38268343236509f, 0.923879532511287f, -0.38268343236509f, 0.38268343236509f, -0.923879532511287f,+    -0.38268343236509f, -0.923879532511287f, -0.923879532511287f, -0.38268343236509f, -0.923879532511287f, 0.38268343236509f, -0.38268343236509f, 0.923879532511287f,+};++template <typename T>+const T FastNoiseLite::Lookup<T>::RandVecs2D[] =+{+    -0.2700222198f, -0.9628540911f, 0.3863092627f, -0.9223693152f, 0.04444859006f, -0.999011673f, -0.5992523158f, -0.8005602176f, -0.7819280288f, 0.6233687174f, 0.9464672271f, 0.3227999196f, -0.6514146797f, -0.7587218957f, 0.9378472289f, 0.347048376f,+    -0.8497875957f, -0.5271252623f, -0.879042592f, 0.4767432447f, -0.892300288f, -0.4514423508f, -0.379844434f, -0.9250503802f, -0.9951650832f, 0.0982163789f, 0.7724397808f, -0.6350880136f, 0.7573283322f, -0.6530343002f, -0.9928004525f, -0.119780055f,+    -0.0532665713f, 0.9985803285f, 0.9754253726f, -0.2203300762f, -0.7665018163f, 0.6422421394f, 0.991636706f, 0.1290606184f, -0.994696838f, 0.1028503788f, -0.5379205513f, -0.84299554f, 0.5022815471f, -0.8647041387f, 0.4559821461f, -0.8899889226f,+    -0.8659131224f, -0.5001944266f, 0.0879458407f, -0.9961252577f, -0.5051684983f, 0.8630207346f, 0.7753185226f, -0.6315704146f, -0.6921944612f, 0.7217110418f, -0.5191659449f, -0.8546734591f, 0.8978622882f, -0.4402764035f, -0.1706774107f, 0.9853269617f,+    -0.9353430106f, -0.3537420705f, -0.9992404798f, 0.03896746794f, -0.2882064021f, -0.9575683108f, -0.9663811329f, 0.2571137995f, -0.8759714238f, -0.4823630009f, -0.8303123018f, -0.5572983775f, 0.05110133755f, -0.9986934731f, -0.8558373281f, -0.5172450752f,+    0.09887025282f, 0.9951003332f, 0.9189016087f, 0.3944867976f, -0.2439375892f, -0.9697909324f, -0.8121409387f, -0.5834613061f, -0.9910431363f, 0.1335421355f, 0.8492423985f, -0.5280031709f, -0.9717838994f, -0.2358729591f, 0.9949457207f, 0.1004142068f,+    0.6241065508f, -0.7813392434f, 0.662910307f, 0.7486988212f, -0.7197418176f, 0.6942418282f, -0.8143370775f, -0.5803922158f, 0.104521054f, -0.9945226741f, -0.1065926113f, -0.9943027784f, 0.445799684f, -0.8951327509f, 0.105547406f, 0.9944142724f,+    -0.992790267f, 0.1198644477f, -0.8334366408f, 0.552615025f, 0.9115561563f, -0.4111755999f, 0.8285544909f, -0.5599084351f, 0.7217097654f, -0.6921957921f, 0.4940492677f, -0.8694339084f, -0.3652321272f, -0.9309164803f, -0.9696606758f, 0.2444548501f,+    0.08925509731f, -0.996008799f, 0.5354071276f, -0.8445941083f, -0.1053576186f, 0.9944343981f, -0.9890284586f, 0.1477251101f, 0.004856104961f, 0.9999882091f, 0.9885598478f, 0.1508291331f, 0.9286129562f, -0.3710498316f, -0.5832393863f, -0.8123003252f,+    0.3015207509f, 0.9534596146f, -0.9575110528f, 0.2883965738f, 0.9715802154f, -0.2367105511f, 0.229981792f, 0.9731949318f, 0.955763816f, -0.2941352207f, 0.740956116f, 0.6715534485f, -0.9971513787f, -0.07542630764f, 0.6905710663f, -0.7232645452f,+    -0.290713703f, -0.9568100872f, 0.5912777791f, -0.8064679708f, -0.9454592212f, -0.325740481f, 0.6664455681f, 0.74555369f, 0.6236134912f, 0.7817328275f, 0.9126993851f, -0.4086316587f, -0.8191762011f, 0.5735419353f, -0.8812745759f, -0.4726046147f,+    0.9953313627f, 0.09651672651f, 0.9855650846f, -0.1692969699f, -0.8495980887f, 0.5274306472f, 0.6174853946f, -0.7865823463f, 0.8508156371f, 0.52546432f, 0.9985032451f, -0.05469249926f, 0.1971371563f, -0.9803759185f, 0.6607855748f, -0.7505747292f,+    -0.03097494063f, 0.9995201614f, -0.6731660801f, 0.739491331f, -0.7195018362f, -0.6944905383f, 0.9727511689f, 0.2318515979f, 0.9997059088f, -0.0242506907f, 0.4421787429f, -0.8969269532f, 0.9981350961f, -0.061043673f, -0.9173660799f, -0.3980445648f,+    -0.8150056635f, -0.5794529907f, -0.8789331304f, 0.4769450202f, 0.0158605829f, 0.999874213f, -0.8095464474f, 0.5870558317f, -0.9165898907f, -0.3998286786f, -0.8023542565f, 0.5968480938f, -0.5176737917f, 0.8555780767f, -0.8154407307f, -0.5788405779f,+    0.4022010347f, -0.9155513791f, -0.9052556868f, -0.4248672045f, 0.7317445619f, 0.6815789728f, -0.5647632201f, -0.8252529947f, -0.8403276335f, -0.5420788397f, -0.9314281527f, 0.363925262f, 0.5238198472f, 0.8518290719f, 0.7432803869f, -0.6689800195f,+    -0.985371561f, -0.1704197369f, 0.4601468731f, 0.88784281f, 0.825855404f, 0.5638819483f, 0.6182366099f, 0.7859920446f, 0.8331502863f, -0.553046653f, 0.1500307506f, 0.9886813308f, -0.662330369f, -0.7492119075f, -0.668598664f, 0.743623444f,+    0.7025606278f, 0.7116238924f, -0.5419389763f, -0.8404178401f, -0.3388616456f, 0.9408362159f, 0.8331530315f, 0.5530425174f, -0.2989720662f, -0.9542618632f, 0.2638522993f, 0.9645630949f, 0.124108739f, -0.9922686234f, -0.7282649308f, -0.6852956957f,+    0.6962500149f, 0.7177993569f, -0.9183535368f, 0.3957610156f, -0.6326102274f, -0.7744703352f, -0.9331891859f, -0.359385508f, -0.1153779357f, -0.9933216659f, 0.9514974788f, -0.3076565421f, -0.08987977445f, -0.9959526224f, 0.6678496916f, 0.7442961705f,+    0.7952400393f, -0.6062947138f, -0.6462007402f, -0.7631674805f, -0.2733598753f, 0.9619118351f, 0.9669590226f, -0.254931851f, -0.9792894595f, 0.2024651934f, -0.5369502995f, -0.8436138784f, -0.270036471f, -0.9628500944f, -0.6400277131f, 0.7683518247f,+    -0.7854537493f, -0.6189203566f, 0.06005905383f, -0.9981948257f, -0.02455770378f, 0.9996984141f, -0.65983623f, 0.751409442f, -0.6253894466f, -0.7803127835f, -0.6210408851f, -0.7837781695f, 0.8348888491f, 0.5504185768f, -0.1592275245f, 0.9872419133f,+    0.8367622488f, 0.5475663786f, -0.8675753916f, -0.4973056806f, -0.2022662628f, -0.9793305667f, 0.9399189937f, 0.3413975472f, 0.9877404807f, -0.1561049093f, -0.9034455656f, 0.4287028224f, 0.1269804218f, -0.9919052235f, -0.3819600854f, 0.924178821f,+    0.9754625894f, 0.2201652486f, -0.3204015856f, -0.9472818081f, -0.9874760884f, 0.1577687387f, 0.02535348474f, -0.9996785487f, 0.4835130794f, -0.8753371362f, -0.2850799925f, -0.9585037287f, -0.06805516006f, -0.99768156f, -0.7885244045f, -0.6150034663f,+    0.3185392127f, -0.9479096845f, 0.8880043089f, 0.4598351306f, 0.6476921488f, -0.7619021462f, 0.9820241299f, 0.1887554194f, 0.9357275128f, -0.3527237187f, -0.8894895414f, 0.4569555293f, 0.7922791302f, 0.6101588153f, 0.7483818261f, 0.6632681526f,+    -0.7288929755f, -0.6846276581f, 0.8729032783f, -0.4878932944f, 0.8288345784f, 0.5594937369f, 0.08074567077f, 0.9967347374f, 0.9799148216f, -0.1994165048f, -0.580730673f, -0.8140957471f, -0.4700049791f, -0.8826637636f, 0.2409492979f, 0.9705377045f,+    0.9437816757f, -0.3305694308f, -0.8927998638f, -0.4504535528f, -0.8069622304f, 0.5906030467f, 0.06258973166f, 0.9980393407f, -0.9312597469f, 0.3643559849f, 0.5777449785f, 0.8162173362f, -0.3360095855f, -0.941858566f, 0.697932075f, -0.7161639607f,+    -0.002008157227f, -0.9999979837f, -0.1827294312f, -0.9831632392f, -0.6523911722f, 0.7578824173f, -0.4302626911f, -0.9027037258f, -0.9985126289f, -0.05452091251f, -0.01028102172f, -0.9999471489f, -0.4946071129f, 0.8691166802f, -0.2999350194f, 0.9539596344f,+    0.8165471961f, 0.5772786819f, 0.2697460475f, 0.962931498f, -0.7306287391f, -0.6827749597f, -0.7590952064f, -0.6509796216f, -0.907053853f, 0.4210146171f, -0.5104861064f, -0.8598860013f, 0.8613350597f, 0.5080373165f, 0.5007881595f, -0.8655698812f,+    -0.654158152f, 0.7563577938f, -0.8382755311f, -0.545246856f, 0.6940070834f, 0.7199681717f, 0.06950936031f, 0.9975812994f, 0.1702942185f, -0.9853932612f, 0.2695973274f, 0.9629731466f, 0.5519612192f, -0.8338697815f, 0.225657487f, -0.9742067022f,+    0.4215262855f, -0.9068161835f, 0.4881873305f, -0.8727388672f, -0.3683854996f, -0.9296731273f, -0.9825390578f, 0.1860564427f, 0.81256471f, 0.5828709909f, 0.3196460933f, -0.9475370046f, 0.9570913859f, 0.2897862643f, -0.6876655497f, -0.7260276109f,+    -0.9988770922f, -0.047376731f, -0.1250179027f, 0.992154486f, -0.8280133617f, 0.560708367f, 0.9324863769f, -0.3612051451f, 0.6394653183f, 0.7688199442f, -0.01623847064f, -0.9998681473f, -0.9955014666f, -0.09474613458f, -0.81453315f, 0.580117012f,+    0.4037327978f, -0.9148769469f, 0.9944263371f, 0.1054336766f, -0.1624711654f, 0.9867132919f, -0.9949487814f, -0.100383875f, -0.6995302564f, 0.7146029809f, 0.5263414922f, -0.85027327f, -0.5395221479f, 0.841971408f, 0.6579370318f, 0.7530729462f,+    0.01426758847f, -0.9998982128f, -0.6734383991f, 0.7392433447f, 0.639412098f, -0.7688642071f, 0.9211571421f, 0.3891908523f, -0.146637214f, -0.9891903394f, -0.782318098f, 0.6228791163f, -0.5039610839f, -0.8637263605f, -0.7743120191f, -0.6328039957f,+};++template <typename T>+const T FastNoiseLite::Lookup<T>::Gradients3D[] =+{+    0, 1, 1, 0,  0,-1, 1, 0,  0, 1,-1, 0,  0,-1,-1, 0,+    1, 0, 1, 0, -1, 0, 1, 0,  1, 0,-1, 0, -1, 0,-1, 0,+    1, 1, 0, 0, -1, 1, 0, 0,  1,-1, 0, 0, -1,-1, 0, 0,+    0, 1, 1, 0,  0,-1, 1, 0,  0, 1,-1, 0,  0,-1,-1, 0,+    1, 0, 1, 0, -1, 0, 1, 0,  1, 0,-1, 0, -1, 0,-1, 0,+    1, 1, 0, 0, -1, 1, 0, 0,  1,-1, 0, 0, -1,-1, 0, 0,+    0, 1, 1, 0,  0,-1, 1, 0,  0, 1,-1, 0,  0,-1,-1, 0,+    1, 0, 1, 0, -1, 0, 1, 0,  1, 0,-1, 0, -1, 0,-1, 0,+    1, 1, 0, 0, -1, 1, 0, 0,  1,-1, 0, 0, -1,-1, 0, 0,+    0, 1, 1, 0,  0,-1, 1, 0,  0, 1,-1, 0,  0,-1,-1, 0,+    1, 0, 1, 0, -1, 0, 1, 0,  1, 0,-1, 0, -1, 0,-1, 0,+    1, 1, 0, 0, -1, 1, 0, 0,  1,-1, 0, 0, -1,-1, 0, 0,+    0, 1, 1, 0,  0,-1, 1, 0,  0, 1,-1, 0,  0,-1,-1, 0,+    1, 0, 1, 0, -1, 0, 1, 0,  1, 0,-1, 0, -1, 0,-1, 0,+    1, 1, 0, 0, -1, 1, 0, 0,  1,-1, 0, 0, -1,-1, 0, 0,+    1, 1, 0, 0,  0,-1, 1, 0, -1, 1, 0, 0,  0,-1,-1, 0+};++template <typename T>+const T FastNoiseLite::Lookup<T>::RandVecs3D[] =+{+    -0.7292736885f, -0.6618439697f, 0.1735581948f, 0, 0.790292081f, -0.5480887466f, -0.2739291014f, 0, 0.7217578935f, 0.6226212466f, -0.3023380997f, 0, 0.565683137f, -0.8208298145f, -0.0790000257f, 0, 0.760049034f, -0.5555979497f, -0.3370999617f, 0, 0.3713945616f, 0.5011264475f, 0.7816254623f, 0, -0.1277062463f, -0.4254438999f, -0.8959289049f, 0, -0.2881560924f, -0.5815838982f, 0.7607405838f, 0,+    0.5849561111f, -0.662820239f, -0.4674352136f, 0, 0.3307171178f, 0.0391653737f, 0.94291689f, 0, 0.8712121778f, -0.4113374369f, -0.2679381538f, 0, 0.580981015f, 0.7021915846f, 0.4115677815f, 0, 0.503756873f, 0.6330056931f, -0.5878203852f, 0, 0.4493712205f, 0.601390195f, 0.6606022552f, 0, -0.6878403724f, 0.09018890807f, -0.7202371714f, 0, -0.5958956522f, -0.6469350577f, 0.475797649f, 0,+    -0.5127052122f, 0.1946921978f, -0.8361987284f, 0, -0.9911507142f, -0.05410276466f, -0.1212153153f, 0, -0.2149721042f, 0.9720882117f, -0.09397607749f, 0, -0.7518650936f, -0.5428057603f, 0.3742469607f, 0, 0.5237068895f, 0.8516377189f, -0.02107817834f, 0, 0.6333504779f, 0.1926167129f, -0.7495104896f, 0, -0.06788241606f, 0.3998305789f, 0.9140719259f, 0, -0.5538628599f, -0.4729896695f, -0.6852128902f, 0,+    -0.7261455366f, -0.5911990757f, 0.3509933228f, 0, -0.9229274737f, -0.1782808786f, 0.3412049336f, 0, -0.6968815002f, 0.6511274338f, 0.3006480328f, 0, 0.9608044783f, -0.2098363234f, -0.1811724921f, 0, 0.06817146062f, -0.9743405129f, 0.2145069156f, 0, -0.3577285196f, -0.6697087264f, -0.6507845481f, 0, -0.1868621131f, 0.7648617052f, -0.6164974636f, 0, -0.6541697588f, 0.3967914832f, 0.6439087246f, 0,+    0.6993340405f, -0.6164538506f, 0.3618239211f, 0, -0.1546665739f, 0.6291283928f, 0.7617583057f, 0, -0.6841612949f, -0.2580482182f, -0.6821542638f, 0, 0.5383980957f, 0.4258654885f, 0.7271630328f, 0, -0.5026987823f, -0.7939832935f, -0.3418836993f, 0, 0.3202971715f, 0.2834415347f, 0.9039195862f, 0, 0.8683227101f, -0.0003762656404f, -0.4959995258f, 0, 0.791120031f, -0.08511045745f, 0.6057105799f, 0,+    -0.04011016052f, -0.4397248749f, 0.8972364289f, 0, 0.9145119872f, 0.3579346169f, -0.1885487608f, 0, -0.9612039066f, -0.2756484276f, 0.01024666929f, 0, 0.6510361721f, -0.2877799159f, -0.7023778346f, 0, -0.2041786351f, 0.7365237271f, 0.644859585f, 0, -0.7718263711f, 0.3790626912f, 0.5104855816f, 0, -0.3060082741f, -0.7692987727f, 0.5608371729f, 0, 0.454007341f, -0.5024843065f, 0.7357899537f, 0,+    0.4816795475f, 0.6021208291f, -0.6367380315f, 0, 0.6961980369f, -0.3222197429f, 0.641469197f, 0, -0.6532160499f, -0.6781148932f, 0.3368515753f, 0, 0.5089301236f, -0.6154662304f, -0.6018234363f, 0, -0.1635919754f, -0.9133604627f, -0.372840892f, 0, 0.52408019f, -0.8437664109f, 0.1157505864f, 0, 0.5902587356f, 0.4983817807f, -0.6349883666f, 0, 0.5863227872f, 0.494764745f, 0.6414307729f, 0,+    0.6779335087f, 0.2341345225f, 0.6968408593f, 0, 0.7177054546f, -0.6858979348f, 0.120178631f, 0, -0.5328819713f, -0.5205125012f, 0.6671608058f, 0, -0.8654874251f, -0.0700727088f, -0.4960053754f, 0, -0.2861810166f, 0.7952089234f, 0.5345495242f, 0, -0.04849529634f, 0.9810836427f, -0.1874115585f, 0, -0.6358521667f, 0.6058348682f, 0.4781800233f, 0, 0.6254794696f, -0.2861619734f, 0.7258696564f, 0,+    -0.2585259868f, 0.5061949264f, -0.8227581726f, 0, 0.02136306781f, 0.5064016808f, -0.8620330371f, 0, 0.200111773f, 0.8599263484f, 0.4695550591f, 0, 0.4743561372f, 0.6014985084f, -0.6427953014f, 0, 0.6622993731f, -0.5202474575f, -0.5391679918f, 0, 0.08084972818f, -0.6532720452f, 0.7527940996f, 0, -0.6893687501f, 0.0592860349f, 0.7219805347f, 0, -0.1121887082f, -0.9673185067f, 0.2273952515f, 0,+    0.7344116094f, 0.5979668656f, -0.3210532909f, 0, 0.5789393465f, -0.2488849713f, 0.7764570201f, 0, 0.6988182827f, 0.3557169806f, -0.6205791146f, 0, -0.8636845529f, -0.2748771249f, -0.4224826141f, 0, -0.4247027957f, -0.4640880967f, 0.777335046f, 0, 0.5257722489f, -0.8427017621f, 0.1158329937f, 0, 0.9343830603f, 0.316302472f, -0.1639543925f, 0, -0.1016836419f, -0.8057303073f, -0.5834887393f, 0,+    -0.6529238969f, 0.50602126f, -0.5635892736f, 0, -0.2465286165f, -0.9668205684f, -0.06694497494f, 0, -0.9776897119f, -0.2099250524f, -0.007368825344f, 0, 0.7736893337f, 0.5734244712f, 0.2694238123f, 0, -0.6095087895f, 0.4995678998f, 0.6155736747f, 0, 0.5794535482f, 0.7434546771f, 0.3339292269f, 0, -0.8226211154f, 0.08142581855f, 0.5627293636f, 0, -0.510385483f, 0.4703667658f, 0.7199039967f, 0,+    -0.5764971849f, -0.07231656274f, -0.8138926898f, 0, 0.7250628871f, 0.3949971505f, -0.5641463116f, 0, -0.1525424005f, 0.4860840828f, -0.8604958341f, 0, -0.5550976208f, -0.4957820792f, 0.667882296f, 0, -0.1883614327f, 0.9145869398f, 0.357841725f, 0, 0.7625556724f, -0.5414408243f, -0.3540489801f, 0, -0.5870231946f, -0.3226498013f, -0.7424963803f, 0, 0.3051124198f, 0.2262544068f, -0.9250488391f, 0,+    0.6379576059f, 0.577242424f, -0.5097070502f, 0, -0.5966775796f, 0.1454852398f, -0.7891830656f, 0, -0.658330573f, 0.6555487542f, -0.3699414651f, 0, 0.7434892426f, 0.2351084581f, 0.6260573129f, 0, 0.5562114096f, 0.8264360377f, -0.0873632843f, 0, -0.3028940016f, -0.8251527185f, 0.4768419182f, 0, 0.1129343818f, -0.985888439f, -0.1235710781f, 0, 0.5937652891f, -0.5896813806f, 0.5474656618f, 0,+    0.6757964092f, -0.5835758614f, -0.4502648413f, 0, 0.7242302609f, -0.1152719764f, 0.6798550586f, 0, -0.9511914166f, 0.0753623979f, -0.2992580792f, 0, 0.2539470961f, -0.1886339355f, 0.9486454084f, 0, 0.571433621f, -0.1679450851f, -0.8032795685f, 0, -0.06778234979f, 0.3978269256f, 0.9149531629f, 0, 0.6074972649f, 0.733060024f, -0.3058922593f, 0, -0.5435478392f, 0.1675822484f, 0.8224791405f, 0,+    -0.5876678086f, -0.3380045064f, -0.7351186982f, 0, -0.7967562402f, 0.04097822706f, -0.6029098428f, 0, -0.1996350917f, 0.8706294745f, 0.4496111079f, 0, -0.02787660336f, -0.9106232682f, -0.4122962022f, 0, -0.7797625996f, -0.6257634692f, 0.01975775581f, 0, -0.5211232846f, 0.7401644346f, -0.4249554471f, 0, 0.8575424857f, 0.4053272873f, -0.3167501783f, 0, 0.1045223322f, 0.8390195772f, -0.5339674439f, 0,+    0.3501822831f, 0.9242524096f, -0.1520850155f, 0, 0.1987849858f, 0.07647613266f, 0.9770547224f, 0, 0.7845996363f, 0.6066256811f, -0.1280964233f, 0, 0.09006737436f, -0.9750989929f, -0.2026569073f, 0, -0.8274343547f, -0.542299559f, 0.1458203587f, 0, -0.3485797732f, -0.415802277f, 0.840000362f, 0, -0.2471778936f, -0.7304819962f, -0.6366310879f, 0, -0.3700154943f, 0.8577948156f, 0.3567584454f, 0,+    0.5913394901f, -0.548311967f, -0.5913303597f, 0, 0.1204873514f, -0.7626472379f, -0.6354935001f, 0, 0.616959265f, 0.03079647928f, 0.7863922953f, 0, 0.1258156836f, -0.6640829889f, -0.7369967419f, 0, -0.6477565124f, -0.1740147258f, -0.7417077429f, 0, 0.6217889313f, -0.7804430448f, -0.06547655076f, 0, 0.6589943422f, -0.6096987708f, 0.4404473475f, 0, -0.2689837504f, -0.6732403169f, -0.6887635427f, 0,+    -0.3849775103f, 0.5676542638f, 0.7277093879f, 0, 0.5754444408f, 0.8110471154f, -0.1051963504f, 0, 0.9141593684f, 0.3832947817f, 0.131900567f, 0, -0.107925319f, 0.9245493968f, 0.3654593525f, 0, 0.377977089f, 0.3043148782f, 0.8743716458f, 0, -0.2142885215f, -0.8259286236f, 0.5214617324f, 0, 0.5802544474f, 0.4148098596f, -0.7008834116f, 0, -0.1982660881f, 0.8567161266f, -0.4761596756f, 0,+    -0.03381553704f, 0.3773180787f, -0.9254661404f, 0, -0.6867922841f, -0.6656597827f, 0.2919133642f, 0, 0.7731742607f, -0.2875793547f, -0.5652430251f, 0, -0.09655941928f, 0.9193708367f, -0.3813575004f, 0, 0.2715702457f, -0.9577909544f, -0.09426605581f, 0, 0.2451015704f, -0.6917998565f, -0.6792188003f, 0, 0.977700782f, -0.1753855374f, 0.1155036542f, 0, -0.5224739938f, 0.8521606816f, 0.02903615945f, 0,+    -0.7734880599f, -0.5261292347f, 0.3534179531f, 0, -0.7134492443f, -0.269547243f, 0.6467878011f, 0, 0.1644037271f, 0.5105846203f, -0.8439637196f, 0, 0.6494635788f, 0.05585611296f, 0.7583384168f, 0, -0.4711970882f, 0.5017280509f, -0.7254255765f, 0, -0.6335764307f, -0.2381686273f, -0.7361091029f, 0, -0.9021533097f, -0.270947803f, -0.3357181763f, 0, -0.3793711033f, 0.872258117f, 0.3086152025f, 0,+    -0.6855598966f, -0.3250143309f, 0.6514394162f, 0, 0.2900942212f, -0.7799057743f, -0.5546100667f, 0, -0.2098319339f, 0.85037073f, 0.4825351604f, 0, -0.4592603758f, 0.6598504336f, -0.5947077538f, 0, 0.8715945488f, 0.09616365406f, -0.4807031248f, 0, -0.6776666319f, 0.7118504878f, -0.1844907016f, 0, 0.7044377633f, 0.312427597f, 0.637304036f, 0, -0.7052318886f, -0.2401093292f, -0.6670798253f, 0,+    0.081921007f, -0.7207336136f, -0.6883545647f, 0, -0.6993680906f, -0.5875763221f, -0.4069869034f, 0, -0.1281454481f, 0.6419895885f, 0.7559286424f, 0, -0.6337388239f, -0.6785471501f, -0.3714146849f, 0, 0.5565051903f, -0.2168887573f, -0.8020356851f, 0, -0.5791554484f, 0.7244372011f, -0.3738578718f, 0, 0.1175779076f, -0.7096451073f, 0.6946792478f, 0, -0.6134619607f, 0.1323631078f, 0.7785527795f, 0,+    0.6984635305f, -0.02980516237f, -0.715024719f, 0, 0.8318082963f, -0.3930171956f, 0.3919597455f, 0, 0.1469576422f, 0.05541651717f, -0.9875892167f, 0, 0.708868575f, -0.2690503865f, 0.6520101478f, 0, 0.2726053183f, 0.67369766f, -0.68688995f, 0, -0.6591295371f, 0.3035458599f, -0.6880466294f, 0, 0.4815131379f, -0.7528270071f, 0.4487723203f, 0, 0.9430009463f, 0.1675647412f, -0.2875261255f, 0,+    0.434802957f, 0.7695304522f, -0.4677277752f, 0, 0.3931996188f, 0.594473625f, 0.7014236729f, 0, 0.7254336655f, -0.603925654f, 0.3301814672f, 0, 0.7590235227f, -0.6506083235f, 0.02433313207f, 0, -0.8552768592f, -0.3430042733f, 0.3883935666f, 0, -0.6139746835f, 0.6981725247f, 0.3682257648f, 0, -0.7465905486f, -0.5752009504f, 0.3342849376f, 0, 0.5730065677f, 0.810555537f, -0.1210916791f, 0,+    -0.9225877367f, -0.3475211012f, -0.167514036f, 0, -0.7105816789f, -0.4719692027f, -0.5218416899f, 0, -0.08564609717f, 0.3583001386f, 0.929669703f, 0, -0.8279697606f, -0.2043157126f, 0.5222271202f, 0, 0.427944023f, 0.278165994f, 0.8599346446f, 0, 0.5399079671f, -0.7857120652f, -0.3019204161f, 0, 0.5678404253f, -0.5495413974f, -0.6128307303f, 0, -0.9896071041f, 0.1365639107f, -0.04503418428f, 0,+    -0.6154342638f, -0.6440875597f, 0.4543037336f, 0, 0.1074204368f, -0.7946340692f, 0.5975094525f, 0, -0.3595449969f, -0.8885529948f, 0.28495784f, 0, -0.2180405296f, 0.1529888965f, 0.9638738118f, 0, -0.7277432317f, -0.6164050508f, -0.3007234646f, 0, 0.7249729114f, -0.00669719484f, 0.6887448187f, 0, -0.5553659455f, -0.5336586252f, 0.6377908264f, 0, 0.5137558015f, 0.7976208196f, -0.3160000073f, 0,+    -0.3794024848f, 0.9245608561f, -0.03522751494f, 0, 0.8229248658f, 0.2745365933f, -0.4974176556f, 0, -0.5404114394f, 0.6091141441f, 0.5804613989f, 0, 0.8036581901f, -0.2703029469f, 0.5301601931f, 0, 0.6044318879f, 0.6832968393f, 0.4095943388f, 0, 0.06389988817f, 0.9658208605f, -0.2512108074f, 0, 0.1087113286f, 0.7402471173f, -0.6634877936f, 0, -0.713427712f, -0.6926784018f, 0.1059128479f, 0,+    0.6458897819f, -0.5724548511f, -0.5050958653f, 0, -0.6553931414f, 0.7381471625f, 0.159995615f, 0, 0.3910961323f, 0.9188871375f, -0.05186755998f, 0, -0.4879022471f, -0.5904376907f, 0.6429111375f, 0, 0.6014790094f, 0.7707441366f, -0.2101820095f, 0, -0.5677173047f, 0.7511360995f, 0.3368851762f, 0, 0.7858573506f, 0.226674665f, 0.5753666838f, 0, -0.4520345543f, -0.604222686f, -0.6561857263f, 0,+    0.002272116345f, 0.4132844051f, -0.9105991643f, 0, -0.5815751419f, -0.5162925989f, 0.6286591339f, 0, -0.03703704785f, 0.8273785755f, 0.5604221175f, 0, -0.5119692504f, 0.7953543429f, -0.3244980058f, 0, -0.2682417366f, -0.9572290247f, -0.1084387619f, 0, -0.2322482736f, -0.9679131102f, -0.09594243324f, 0, 0.3554328906f, -0.8881505545f, 0.2913006227f, 0, 0.7346520519f, -0.4371373164f, 0.5188422971f, 0,+    0.9985120116f, 0.04659011161f, -0.02833944577f, 0, -0.3727687496f, -0.9082481361f, 0.1900757285f, 0, 0.91737377f, -0.3483642108f, 0.1925298489f, 0, 0.2714911074f, 0.4147529736f, -0.8684886582f, 0, 0.5131763485f, -0.7116334161f, 0.4798207128f, 0, -0.8737353606f, 0.18886992f, -0.4482350644f, 0, 0.8460043821f, -0.3725217914f, 0.3814499973f, 0, 0.8978727456f, -0.1780209141f, -0.4026575304f, 0,+    0.2178065647f, -0.9698322841f, -0.1094789531f, 0, -0.1518031304f, -0.7788918132f, -0.6085091231f, 0, -0.2600384876f, -0.4755398075f, -0.8403819825f, 0, 0.572313509f, -0.7474340931f, -0.3373418503f, 0, -0.7174141009f, 0.1699017182f, -0.6756111411f, 0, -0.684180784f, 0.02145707593f, -0.7289967412f, 0, -0.2007447902f, 0.06555605789f, -0.9774476623f, 0, -0.1148803697f, -0.8044887315f, 0.5827524187f, 0,+    -0.7870349638f, 0.03447489231f, 0.6159443543f, 0, -0.2015596421f, 0.6859872284f, 0.6991389226f, 0, -0.08581082512f, -0.10920836f, -0.9903080513f, 0, 0.5532693395f, 0.7325250401f, -0.396610771f, 0, -0.1842489331f, -0.9777375055f, -0.1004076743f, 0, 0.0775473789f, -0.9111505856f, 0.4047110257f, 0, 0.1399838409f, 0.7601631212f, -0.6344734459f, 0, 0.4484419361f, -0.845289248f, 0.2904925424f, 0+};++#endif
+ bench/fnl-compare/fnl_bench_shim.cpp view
@@ -0,0 +1,54 @@+// Bulk-loop shim over the vendored FastNoiseLite.h (./cbits/FastNoiseLite.h)+// to provide a fair comparison to FNL.+//+// Sticking the loop in C++ should prevent FFI overhead from skewing results. +// Unsafe calls should have double-digit nanosecond overhead, so the signal from +// the bench itself should dominate.+//+// tasty-bench handles timing and statistics; simpler and less numerically fraught +// than trying to cross-compare Google benchmark to tasty/criterion.+//+// This should be compiled the way a user would build FNL: -O3 -march=native. We+// want to give C++ every edge it can have for a meaningful comparison.+//+// NOTE: noiseType convention is:+//   0=OpenSimplex2 +//   1=OpenSimplex2S +//   2=Cellular +//   3=Perlin +//   4=ValueCubic +//   5=Value+#include "FastNoiseLite.h"++static FastNoiseLite mk(int noiseType, int seed, float freq) {+    FastNoiseLite fnl(seed);+    fnl.SetNoiseType(static_cast<FastNoiseLite::NoiseType>(noiseType));+    // pure-noise's loop multiplies every coordinate by a runtime k; FNL pays+    // the same multiply here via mFrequency.+    fnl.SetFrequency(freq);+    if (noiseType == 2) { // match audit-bench: Euclidean distance, Distance result+        fnl.SetCellularDistanceFunction(FastNoiseLite::CellularDistanceFunction_Euclidean);+        fnl.SetCellularReturnType(FastNoiseLite::CellularReturnType_Distance);+    }+    return fnl;+}++extern "C" {++float fnl_bench_sum2(int noiseType, int seed, float freq, int n,+                     const float* __restrict xs, const float* __restrict ys) {+    FastNoiseLite fnl = mk(noiseType, seed, freq);+    float acc = 0.0f;+    for (int i = 0; i < n; i++) acc += fnl.GetNoise(xs[i], ys[i]);+    return acc;+}++float fnl_bench_sum3(int noiseType, int seed, float freq, int n,+                     const float* __restrict xs, const float* __restrict ys, const float* __restrict zs) {+    FastNoiseLite fnl = mk(noiseType, seed, freq);+    float acc = 0.0f;+    for (int i = 0; i < n; i++) acc += fnl.GetNoise(xs[i], ys[i], zs[i]);+    return acc;+}++} // extern "C"
pure-noise.cabal view
@@ -1,32 +1,32 @@ cabal-version: 2.2 --- This file has been generated from package.yaml by hpack version 0.38.1.+-- This file has been generated from package.yaml by hpack version 0.39.6. -- -- see: https://github.com/sol/hpack  name:           pure-noise-version:        0.2.1.1-synopsis:       High-performance composable noise generation (Perlin, Simplex, Cellular)-description:    A high-performance noise generation library ported from FastNoiseLite.-                Provides N-dimensional noise functions (Perlin, OpenSimplex, SuperSimplex,-                Value, Cellular) that can be composed using Num or Fractional methods with-                minimal performance overhead. Noise values are generally clamped to [-1, 1].-                Benefits significantly from LLVM backend compilation (~50-80% performance improvement).+version:        0.2.2.0+synopsis:       Performant, modern noise generation (Perlin, OpenSimplex2, Cellular)+description:    Fast, modern noise generation In pure Haskell with an algebraic interface.+                Provides Perlin, OpenSimplex2, OpenSimplex2S, Value, and Cellular noise variants. category:       Math, Numeric, Noise homepage:       https://github.com/jtnuttall/pure-noise#readme bug-reports:    https://github.com/jtnuttall/pure-noise/issues author:         Jeremy Nuttall maintainer:     jeremy@jeremy-nuttall.com-copyright:      2024 Jeremy Nuttall+copyright:      2026 Jeremy Nuttall license:        BSD-3-Clause license-file:   LICENSE build-type:     Simple tested-with:     GHC == 9.6.7   , GHC == 9.8.4-  , GHC == 9.10.2+  , GHC == 9.12.2 extra-source-files:     README.md+    LICENSE+    LICENSE_FastNoiseLite+    bench/fnl-compare/cbits/FastNoiseLite.h extra-doc-files:     CHANGELOG.md @@ -34,6 +34,39 @@   type: git   location: https://github.com/jtnuttall/pure-noise +flag llvm-bench+  description: Build the benchmark suites via GHC's LLVM backend with a pinned toolchain+               (-fllvm, -pgmlo opt, -pgmlc llc, -pgmlas clang) plus -mavx -mfma and fast+               FP contraction in llc (-optlc-fp-contract=fast). Requires opt/llc/clang+               on PATH (LLVM <= 19 for GHC 9.12) and GHC >= 9.10 (for -pgmlas). Off by+               default so `cabal build --enable-benchmarks` works on machines and CI+               runners without LLVM. Published benchmark numbers are collected with this+               flag on; see bench/README.md.+  manual: True+  default: False++flag mavx+  description: Compile with AVX instructions+  manual: True+  default: False++flag mfma+  description: Compile with native fused-multiply-add instructions+  manual: True+  default: False++flag optimize+  description: Turns on -O2 for pure-noise. Since the library is pretty small, this shouldn't be+               too much trouble, but you can disable this flag if it's slowing your builds too+               much.+               .+               Rationale:+               - -O1 leaves ~3x on the table for tight noise loops in downstream code.+               - -O2 seems to produce higher-quality unfoldings, which in turn causes pure-noise's+               kernels to inline and optimize more reliably at call sites.+  manual: True+  default: True+ library   exposed-modules:       Numeric.Noise@@ -57,6 +90,14 @@       base >=4.16 && <5     , primitive >=0.8 && <0.10   default-language: GHC2021+  if flag(optimize)+    ghc-options: -O2+  if flag(mfma)+    ghc-options: -mfma+    cpp-options: -DHAS_FMA+  if flag(mavx)+    ghc-options: -mavx+    cpp-options: -DHAS_AVX  test-suite pure-noise-test   type: exitcode-stdio-1.0@@ -70,6 +111,7 @@       OpenSimplexSpec       PerlinSpec       SuperSimplexSpec+      TotalitySpec       ValueCubicSpec       ValueSpec       Paths_pure_noise@@ -78,6 +120,8 @@   hs-source-dirs:       test   ghc-options: -Wall -Wcompat -Widentities -Wincomplete-record-updates -Wincomplete-uni-patterns -Wmissing-export-lists -Wmissing-home-modules -Wpartial-fields -Wredundant-constraints -Wno-missing-export-lists -threaded -rtsopts -with-rtsopts=-N+  build-tool-depends:+      tasty-discover:tasty-discover   build-depends:       JuicyPixels ==3.3.*     , aeson >=2.0 && <2.3@@ -90,7 +134,6 @@     , primitive >=0.8 && <0.10     , pure-noise     , tasty-    , tasty-discover     , tasty-golden     , tasty-hunit     , tasty-quickcheck@@ -108,7 +151,7 @@       Paths_pure_noise   hs-source-dirs:       bench-  ghc-options: -Wall -Wcompat -Widentities -Wincomplete-record-updates -Wincomplete-uni-patterns -Wmissing-export-lists -Wmissing-home-modules -Wpartial-fields -Wredundant-constraints -threaded -rtsopts -with-rtsopts=-N +RTS -A32m --nonmoving-gc -T -RTS -O2 -optc-O3 -fsimpl-tick-factor=1000+  ghc-options: -Wall -Wcompat -Widentities -Wincomplete-record-updates -Wincomplete-uni-patterns -Wmissing-export-lists -Wmissing-home-modules -Wpartial-fields -Wredundant-constraints -threaded -rtsopts "-with-rtsopts=-N -A32m -T" -O2 -fsimpl-tick-factor=1000   build-depends:       base >=4.16 && <5     , deepseq@@ -118,5 +161,43 @@     , random     , tasty     , tasty-bench-    , vector <=0.14+    , vector >=0.12 && <0.14   default-language: GHC2021+  if flag(llvm-bench)+    ghc-options: -fllvm -pgmlo opt -pgmlc llc -pgmlas clang -mavx -mfma -optlc-fp-contract=fast++benchmark pure-noise-fnl-bench+  type: exitcode-stdio-1.0+  main-is: FnlBench.hs+  other-modules:+      Paths_pure_noise+  autogen-modules:+      Paths_pure_noise+  hs-source-dirs:+      bench/fnl-compare+  ghc-options: -Wall -Wcompat -Widentities -Wincomplete-record-updates -Wincomplete-uni-patterns -Wmissing-export-lists -Wmissing-home-modules -Wpartial-fields -Wredundant-constraints -threaded -rtsopts "-with-rtsopts=-A64m -T" -O2 -fsimpl-tick-factor=1000+  cxx-options: -O3 -std=c++14 -ffp-contract=fast -fstrict-overflow+  include-dirs:+      bench/fnl-compare/cbits+  cxx-sources:+      bench/fnl-compare/fnl_bench_shim.cpp+  build-depends:+      base >=4.16 && <5+    , mtl >=2.1 && <2.4+    , primitive >=0.8 && <0.10+    , pure-noise+    , splitmix ==0.1.*+    , tasty+    , tasty-bench >=0.4 && <0.6+    , vector >=0.12 && <0.14+  default-language: GHC2021+  if flag(llvm-bench)+    ghc-options: -fllvm -pgmlo opt -pgmlc llc -pgmlas clang -mavx -mfma -optlc-fp-contract=fast+  if arch(x86_64)+    cxx-options: -march=x86-64-v3+  if os(darwin)+    extra-libraries:+        c+++  else+    extra-libraries:+        stdc++
src/Numeric/Noise.hs view
@@ -12,8 +12,8 @@ -- type aliases for 2D and 3D noise. These can be composed algebraically -- with minimal performance overhead. ----- Noise values are generally clamped to @[-1, 1]@, though some functions may--- occasionally produce values slightly outside this range.+-- Noise values are generally clamped to @[-1, 1]@, although some noise+-- functions may occasionally produce values slightly outside this range. -- -- == Basic Usage --@@ -61,15 +61,28 @@ -- scaledAndLayered :: Noise.Noise2 Float -- scaledAndLayered = --  Noise.warp (\\(x, y) -> (x * 2, y * 2)) Noise.perlin2---    + fmap (logBase 2) Noise.perlin2+--    + fmap (* 0.5) Noise.perlin2 -- @ -- -- Layer independent noise with 'reseed' or 'next2': -- -- @ -- layered :: Noise.Noise2 Float--- layered = Noise.perlin2 + Noise.next2 Noise.perlin2 \/ 2+-- layered = (Noise.perlin2 + Noise.next2 Noise.perlin2) \/ 2 -- @+--+-- == Coordinate domain+--+-- Coordinates are supported on the Int32 lattice range (@|x| < 2^31@; in+-- practice 'Float' precision runs out well before that).+--+-- The OpenSimplex2\/2S family first rotates coordinates into its lattice domain,+-- which shrinks its usable range by the rotation factor — up to ~1.73x, so roughly+-- @|x| < 1.2e9@.+--+-- Outside those domains, or for non-finite inputs, results are unspecified.+--+-- This behavior mirrors FastNoiseLite, but may change in a future major version. module Numeric.Noise (   -- * Noise @@ -102,12 +115,15 @@    -- ** OpenSimplex   openSimplex2,+  openSimplex3,    -- ** OpenSimplex2S   superSimplex2,+  superSimplex3,    -- ** Cellular   cellular2,+  cellular3,    -- *** Configuration   CellularConfig (..),@@ -127,9 +143,10 @@    -- * Noise alteration -  --  ** Altering values+  -- ** Altering values   remap,-  --  ** Altering parameters++  -- ** Altering parameters   warp,   reseed,   next2,@@ -149,8 +166,8 @@   -- | Fractal noise combines multiple octaves at different frequencies and   -- amplitudes to create natural-looking, multi-scale patterns.   ---  -- For custom fractal implementations using modifier functions, see-  -- "Numeric.Noise.Fractal".+  -- For custom fractal implementations using per-octave modifier functions,+  -- see "Numeric.Noise.Fractal".    -- ** Fractal Brownian Motion (FBM)   fractal2,@@ -170,6 +187,18 @@   PingPongStrength (..),   defaultPingPongStrength, +  -- * Custom kernels++  --++  -- | Lift a plain @seed -> coordinates -> value@ function into a composable+  -- 'Noise' value — the inverses of the accessors above.+  --+  -- You may use these to construct custom kernels.+  mkNoise1,+  mkNoise2,+  mkNoise3,+   -- * Math utilities   clamp,   clamp2,@@ -199,17 +228,34 @@ cellular2 = Cellular.noise2 {-# INLINE cellular2 #-} +-- | 3D Cellular (Worley) noise. See 'cellular2'.+cellular3 :: (RealFrac a, Floating a) => CellularConfig a -> Noise3 a+cellular3 = Cellular.noise3+{-# INLINE cellular3 #-}+ -- | 2D OpenSimplex noise. Smooth gradient noise similar to Perlin but without -- directional artifacts. openSimplex2 :: (RealFrac a) => Noise2 a openSimplex2 = OpenSimplex.noise2 {-# INLINE openSimplex2 #-} +-- | 3D OpenSimplex noise (FastNoiseLite's OpenSimplex2, two offset rotated+-- cube grids), including its default coordinate rotation.+openSimplex3 :: (RealFrac a) => Noise3 a+openSimplex3 = OpenSimplex.noise3+{-# INLINE openSimplex3 #-}+ -- | 2D SuperSimplex noise. Improved OpenSimplex variant with better visual -- characteristics. superSimplex2 :: (RealFrac a) => Noise2 a superSimplex2 = SuperSimplex.noise2 {-# INLINE superSimplex2 #-}++-- | 3D SuperSimplex noise (FastNoiseLite's OpenSimplex2S, two offset rotated+-- cube grids), including its default coordinate rotation.+superSimplex3 :: (RealFrac a) => Noise3 a+superSimplex3 = SuperSimplex.noise3+{-# INLINE superSimplex3 #-}  -- | 2D Perlin noise. Classic gradient noise algorithm. perlin2 :: (RealFrac a) => Noise2 a
src/Numeric/Noise/Cellular.hs view
@@ -15,6 +15,9 @@    -- * 2D Noise   noise2,++  -- * 3D Noise+  noise3, ) where  import Data.Bits@@ -69,6 +72,9 @@ -- -- These options allow for different visual effects by returning different -- properties of the cell structure.+--+-- Distance-based results are not confined to @[-1, 1]@ under every metric —+-- 'DistManhattan' and 'DistHybrid' can exceed 1, matching FastNoiseLite. data CellularResult   = -- | Return the hash value of the nearest cell point.     -- Creates discrete regions with constant values.@@ -107,30 +113,22 @@   _ -> id {-# INLINE normDist #-} -noise2 :: (RealFrac a, Floating a) => CellularConfig a -> Noise2 a-noise2 CellularConfig{..} = mkNoise2 $ \ !seed !x !y ->-  let !jitter = cellularJitter * 0.43701595-      !rx = round x-      !ry = round y+distance3 :: (RealFrac a) => CellularDistanceFn -> a -> a -> a -> a+distance3 = \case+  DistEuclidean -> \ !x !y !z -> x * x + y * y + z * z+  DistEuclideanSq -> \ !x !y !z -> x * x + y * y + z * z+  DistManhattan -> \ !x !y !z -> abs x + abs y + abs z+  DistHybrid -> \ !x !y !z -> abs x + abs y + abs z + (x * x + y * y + z * z)+{-# INLINE distance3 #-} -      dist = distance cellularDistanceFn-      norm = normDist cellularDistanceFn+-- | Fold the candidate cell points with the selector 'cellularResult' needs.+-- Shared by 'noise2' and 'noise3'; INLINE so the folds fuse with the callers'+-- point comprehensions (see the note on @points@ in 'noise2').+selectResult :: (RealFrac a, Floating a) => CellularResult -> CellularDistanceFn -> [(Hash, a)] -> a+selectResult res distFn points =+  let norm = normDist distFn       coeff = 1 / (maxHash + 1) -      {-# INLINE pointDist #-}-      pointDist !xi !yi =-        let !px = fromIntegral xi - x-            !py = fromIntegral yi - y-            !h = hash2 seed (primeX * xi) (primeY * yi)-            !i = h .&. 0x1FE-            !rvx = lookupRandVec2d i-            !rvy = lookupRandVec2d (i .|. 1)-            !d = dist (px + rvx * jitter) (py + rvy * jitter)-         in (h, d)--      {-# INLINE points #-}-      points = [pointDist (rx + xi) (ry + yi) | !xi <- [-1 .. 1], !yi <- [-1 .. 1]]-       {-# INLINE selectMinHash #-}       selectMinHash =         let minHash (!hMin, !dMin) (!h, !d)@@ -152,7 +150,7 @@               | d < d1 = (c, d0, d)               | otherwise = (c, d0, d1)          in foldl' smallestTwo (0, infinity, infinity) points-   in case cellularResult of+   in case res of         CellValue ->           let (!hash, !_) = selectMinHash            in fromIntegral hash * coeff@@ -174,8 +172,69 @@         Distance2Div ->           let (!_, !d0, !d1) = selectSmallestTwo            in norm d0 / norm d1 - 1+{-# INLINE selectResult #-}++noise2 :: (RealFrac a, Floating a) => CellularConfig a -> Noise2 a+noise2 CellularConfig{..} = mkNoise2 $ \ !seed !x !y ->+  let !jitter = cellularJitter * 0.43701595+      !rx = fastRound x+      !ry = fastRound y++      dist = distance cellularDistanceFn++      {-# INLINE pointDist #-}+      pointDist !xi !yi =+        let !px = fromIntegral xi - x+            !py = fromIntegral yi - y+            !h = hash2 seed (primeX * xi) (primeY * yi)+            !i = h .&. 0x1FE+            !rvx = lookupRandVec2d i+            !rvy = lookupRandVec2d (i .|. 1)+            !d = dist (px + rvx * jitter) (py + rvy * jitter)+         in (h, d)++      -- The [-1 .. 1] comprehension is load-bearing: enumFromTo takes part in+      -- foldr/build fusion with the folds below, so no list is ever built.+      -- OverloadedLists literals here would desugar to fromListN (no fusion),+      -- and hand-unrolled cons-chains defeat the fold's fusion entirely+      -- (measured ~5x slower). Same applies to the 3D comprehension.+      {-# INLINE points #-}+      points = [pointDist (rx + xi) (ry + yi) | !xi <- [-1 .. 1], !yi <- [-1 .. 1]]+   in selectResult cellularResult cellularDistanceFn points {-# INLINE [2] noise2 #-} +noise3 :: (RealFrac a, Floating a) => CellularConfig a -> Noise3 a+noise3 CellularConfig{..} = mkNoise3 $ \ !seed !x !y !z ->+  let !jitter = cellularJitter * 0.39614353+      !rx = fastRound x+      !ry = fastRound y+      !rz = fastRound z++      dist = distance3 cellularDistanceFn++      {-# INLINE pointDist #-}+      pointDist !xi !yi !zi =+        let !px = fromIntegral xi - x+            !py = fromIntegral yi - y+            !pz = fromIntegral zi - z+            !h = hash3 seed (primeX * xi) (primeY * yi) (primeZ * zi)+            !i = h .&. 0x3FC+            !rvx = lookupRandVec3d i+            !rvy = lookupRandVec3d (i .|. 1)+            !rvz = lookupRandVec3d (i .|. 2)+            !d = dist (px + rvx * jitter) (py + rvy * jitter) (pz + rvz * jitter)+         in (h, d)++      {-# INLINE points #-}+      points =+        [ pointDist (rx + xi) (ry + yi) (rz + zi)+        | !xi <- [-1 .. 1]+        , !yi <- [-1 .. 1]+        , !zi <- [-1 .. 1]+        ]+   in selectResult cellularResult cellularDistanceFn points+{-# INLINE [2] noise3 #-}+ lookupRandVec2d :: (RealFrac a) => Hash -> a lookupRandVec2d = realToFrac . indexPrimArray randVecs2dd . fromIntegral {-# NOINLINE [1] lookupRandVec2d #-}@@ -192,7 +251,7 @@ randVecs2df :: PrimArray Float randVecs2df = mapPrimArray realToFrac randVecs2dd --- >>> sizeofPrimArray randVecs2d == 512+-- >>> sizeofPrimArray randVecs2dd == 512 -- True {- ORMOLU_DISABLE -} randVecs2dd :: PrimArray Double@@ -262,3 +321,155 @@   ,0.01426758847,-0.9998982128,-0.6734383991,0.7392433447,0.639412098,-0.7688642071,0.9211571421,0.3891908523   ,-0.146637214,-0.9891903394,-0.782318098,0.6228791163,-0.5039610839,-0.8637263605,-0.7743120191,-0.6328039957   ]++lookupRandVec3d :: (RealFrac a) => Hash -> a+lookupRandVec3d = realToFrac . indexPrimArray randVecs3dd . fromIntegral+{-# NOINLINE [1] lookupRandVec3d #-}++{-# RULES+"lookupRandVec3d/Float" forall h.+  lookupRandVec3d h =+    indexPrimArray randVecs3df (fromIntegral h)+"lookupRandVec3d/Double" forall h.+  lookupRandVec3d h =+    indexPrimArray randVecs3dd (fromIntegral h)+  #-}++randVecs3df :: PrimArray Float+randVecs3df = mapPrimArray realToFrac randVecs3dd++-- >>> sizeofPrimArray randVecs3dd == 1024+-- True+{- ORMOLU_DISABLE -}+randVecs3dd :: PrimArray Double+randVecs3dd =+  [-0.7292736885,-0.6618439697,0.1735581948,0,0.790292081,-0.5480887466,-0.2739291014,0+  ,0.7217578935,0.6226212466,-0.3023380997,0,0.565683137,-0.8208298145,-0.0790000257,0+  ,0.760049034,-0.5555979497,-0.3370999617,0,0.3713945616,0.5011264475,0.7816254623,0+  ,-0.1277062463,-0.4254438999,-0.8959289049,0,-0.2881560924,-0.5815838982,0.7607405838,0+  ,0.5849561111,-0.662820239,-0.4674352136,0,0.3307171178,0.0391653737,0.94291689,0+  ,0.8712121778,-0.4113374369,-0.2679381538,0,0.580981015,0.7021915846,0.4115677815,0+  ,0.503756873,0.6330056931,-0.5878203852,0,0.4493712205,0.601390195,0.6606022552,0+  ,-0.6878403724,0.09018890807,-0.7202371714,0,-0.5958956522,-0.6469350577,0.475797649,0+  ,-0.5127052122,0.1946921978,-0.8361987284,0,-0.9911507142,-0.05410276466,-0.1212153153,0+  ,-0.2149721042,0.9720882117,-0.09397607749,0,-0.7518650936,-0.5428057603,0.3742469607,0+  ,0.5237068895,0.8516377189,-0.02107817834,0,0.6333504779,0.1926167129,-0.7495104896,0+  ,-0.06788241606,0.3998305789,0.9140719259,0,-0.5538628599,-0.4729896695,-0.6852128902,0+  ,-0.7261455366,-0.5911990757,0.3509933228,0,-0.9229274737,-0.1782808786,0.3412049336,0+  ,-0.6968815002,0.6511274338,0.3006480328,0,0.9608044783,-0.2098363234,-0.1811724921,0+  ,0.06817146062,-0.9743405129,0.2145069156,0,-0.3577285196,-0.6697087264,-0.6507845481,0+  ,-0.1868621131,0.7648617052,-0.6164974636,0,-0.6541697588,0.3967914832,0.6439087246,0+  ,0.6993340405,-0.6164538506,0.3618239211,0,-0.1546665739,0.6291283928,0.7617583057,0+  ,-0.6841612949,-0.2580482182,-0.6821542638,0,0.5383980957,0.4258654885,0.7271630328,0+  ,-0.5026987823,-0.7939832935,-0.3418836993,0,0.3202971715,0.2834415347,0.9039195862,0+  ,0.8683227101,-0.0003762656404,-0.4959995258,0,0.791120031,-0.08511045745,0.6057105799,0+  ,-0.04011016052,-0.4397248749,0.8972364289,0,0.9145119872,0.3579346169,-0.1885487608,0+  ,-0.9612039066,-0.2756484276,0.01024666929,0,0.6510361721,-0.2877799159,-0.7023778346,0+  ,-0.2041786351,0.7365237271,0.644859585,0,-0.7718263711,0.3790626912,0.5104855816,0+  ,-0.3060082741,-0.7692987727,0.5608371729,0,0.454007341,-0.5024843065,0.7357899537,0+  ,0.4816795475,0.6021208291,-0.6367380315,0,0.6961980369,-0.3222197429,0.641469197,0+  ,-0.6532160499,-0.6781148932,0.3368515753,0,0.5089301236,-0.6154662304,-0.6018234363,0+  ,-0.1635919754,-0.9133604627,-0.372840892,0,0.52408019,-0.8437664109,0.1157505864,0+  ,0.5902587356,0.4983817807,-0.6349883666,0,0.5863227872,0.494764745,0.6414307729,0+  ,0.6779335087,0.2341345225,0.6968408593,0,0.7177054546,-0.6858979348,0.120178631,0+  ,-0.5328819713,-0.5205125012,0.6671608058,0,-0.8654874251,-0.0700727088,-0.4960053754,0+  ,-0.2861810166,0.7952089234,0.5345495242,0,-0.04849529634,0.9810836427,-0.1874115585,0+  ,-0.6358521667,0.6058348682,0.4781800233,0,0.6254794696,-0.2861619734,0.7258696564,0+  ,-0.2585259868,0.5061949264,-0.8227581726,0,0.02136306781,0.5064016808,-0.8620330371,0+  ,0.200111773,0.8599263484,0.4695550591,0,0.4743561372,0.6014985084,-0.6427953014,0+  ,0.6622993731,-0.5202474575,-0.5391679918,0,0.08084972818,-0.6532720452,0.7527940996,0+  ,-0.6893687501,0.0592860349,0.7219805347,0,-0.1121887082,-0.9673185067,0.2273952515,0+  ,0.7344116094,0.5979668656,-0.3210532909,0,0.5789393465,-0.2488849713,0.7764570201,0+  ,0.6988182827,0.3557169806,-0.6205791146,0,-0.8636845529,-0.2748771249,-0.4224826141,0+  ,-0.4247027957,-0.4640880967,0.777335046,0,0.5257722489,-0.8427017621,0.1158329937,0+  ,0.9343830603,0.316302472,-0.1639543925,0,-0.1016836419,-0.8057303073,-0.5834887393,0+  ,-0.6529238969,0.50602126,-0.5635892736,0,-0.2465286165,-0.9668205684,-0.06694497494,0+  ,-0.9776897119,-0.2099250524,-0.007368825344,0,0.7736893337,0.5734244712,0.2694238123,0+  ,-0.6095087895,0.4995678998,0.6155736747,0,0.5794535482,0.7434546771,0.3339292269,0+  ,-0.8226211154,0.08142581855,0.5627293636,0,-0.510385483,0.4703667658,0.7199039967,0+  ,-0.5764971849,-0.07231656274,-0.8138926898,0,0.7250628871,0.3949971505,-0.5641463116,0+  ,-0.1525424005,0.4860840828,-0.8604958341,0,-0.5550976208,-0.4957820792,0.667882296,0+  ,-0.1883614327,0.9145869398,0.357841725,0,0.7625556724,-0.5414408243,-0.3540489801,0+  ,-0.5870231946,-0.3226498013,-0.7424963803,0,0.3051124198,0.2262544068,-0.9250488391,0+  ,0.6379576059,0.577242424,-0.5097070502,0,-0.5966775796,0.1454852398,-0.7891830656,0+  ,-0.658330573,0.6555487542,-0.3699414651,0,0.7434892426,0.2351084581,0.6260573129,0+  ,0.5562114096,0.8264360377,-0.0873632843,0,-0.3028940016,-0.8251527185,0.4768419182,0+  ,0.1129343818,-0.985888439,-0.1235710781,0,0.5937652891,-0.5896813806,0.5474656618,0+  ,0.6757964092,-0.5835758614,-0.4502648413,0,0.7242302609,-0.1152719764,0.6798550586,0+  ,-0.9511914166,0.0753623979,-0.2992580792,0,0.2539470961,-0.1886339355,0.9486454084,0+  ,0.571433621,-0.1679450851,-0.8032795685,0,-0.06778234979,0.3978269256,0.9149531629,0+  ,0.6074972649,0.733060024,-0.3058922593,0,-0.5435478392,0.1675822484,0.8224791405,0+  ,-0.5876678086,-0.3380045064,-0.7351186982,0,-0.7967562402,0.04097822706,-0.6029098428,0+  ,-0.1996350917,0.8706294745,0.4496111079,0,-0.02787660336,-0.9106232682,-0.4122962022,0+  ,-0.7797625996,-0.6257634692,0.01975775581,0,-0.5211232846,0.7401644346,-0.4249554471,0+  ,0.8575424857,0.4053272873,-0.3167501783,0,0.1045223322,0.8390195772,-0.5339674439,0+  ,0.3501822831,0.9242524096,-0.1520850155,0,0.1987849858,0.07647613266,0.9770547224,0+  ,0.7845996363,0.6066256811,-0.1280964233,0,0.09006737436,-0.9750989929,-0.2026569073,0+  ,-0.8274343547,-0.542299559,0.1458203587,0,-0.3485797732,-0.415802277,0.840000362,0+  ,-0.2471778936,-0.7304819962,-0.6366310879,0,-0.3700154943,0.8577948156,0.3567584454,0+  ,0.5913394901,-0.548311967,-0.5913303597,0,0.1204873514,-0.7626472379,-0.6354935001,0+  ,0.616959265,0.03079647928,0.7863922953,0,0.1258156836,-0.6640829889,-0.7369967419,0+  ,-0.6477565124,-0.1740147258,-0.7417077429,0,0.6217889313,-0.7804430448,-0.06547655076,0+  ,0.6589943422,-0.6096987708,0.4404473475,0,-0.2689837504,-0.6732403169,-0.6887635427,0+  ,-0.3849775103,0.5676542638,0.7277093879,0,0.5754444408,0.8110471154,-0.1051963504,0+  ,0.9141593684,0.3832947817,0.131900567,0,-0.107925319,0.9245493968,0.3654593525,0+  ,0.377977089,0.3043148782,0.8743716458,0,-0.2142885215,-0.8259286236,0.5214617324,0+  ,0.5802544474,0.4148098596,-0.7008834116,0,-0.1982660881,0.8567161266,-0.4761596756,0+  ,-0.03381553704,0.3773180787,-0.9254661404,0,-0.6867922841,-0.6656597827,0.2919133642,0+  ,0.7731742607,-0.2875793547,-0.5652430251,0,-0.09655941928,0.9193708367,-0.3813575004,0+  ,0.2715702457,-0.9577909544,-0.09426605581,0,0.2451015704,-0.6917998565,-0.6792188003,0+  ,0.977700782,-0.1753855374,0.1155036542,0,-0.5224739938,0.8521606816,0.02903615945,0+  ,-0.7734880599,-0.5261292347,0.3534179531,0,-0.7134492443,-0.269547243,0.6467878011,0+  ,0.1644037271,0.5105846203,-0.8439637196,0,0.6494635788,0.05585611296,0.7583384168,0+  ,-0.4711970882,0.5017280509,-0.7254255765,0,-0.6335764307,-0.2381686273,-0.7361091029,0+  ,-0.9021533097,-0.270947803,-0.3357181763,0,-0.3793711033,0.872258117,0.3086152025,0+  ,-0.6855598966,-0.3250143309,0.6514394162,0,0.2900942212,-0.7799057743,-0.5546100667,0+  ,-0.2098319339,0.85037073,0.4825351604,0,-0.4592603758,0.6598504336,-0.5947077538,0+  ,0.8715945488,0.09616365406,-0.4807031248,0,-0.6776666319,0.7118504878,-0.1844907016,0+  ,0.7044377633,0.312427597,0.637304036,0,-0.7052318886,-0.2401093292,-0.6670798253,0+  ,0.081921007,-0.7207336136,-0.6883545647,0,-0.6993680906,-0.5875763221,-0.4069869034,0+  ,-0.1281454481,0.6419895885,0.7559286424,0,-0.6337388239,-0.6785471501,-0.3714146849,0+  ,0.5565051903,-0.2168887573,-0.8020356851,0,-0.5791554484,0.7244372011,-0.3738578718,0+  ,0.1175779076,-0.7096451073,0.6946792478,0,-0.6134619607,0.1323631078,0.7785527795,0+  ,0.6984635305,-0.02980516237,-0.715024719,0,0.8318082963,-0.3930171956,0.3919597455,0+  ,0.1469576422,0.05541651717,-0.9875892167,0,0.708868575,-0.2690503865,0.6520101478,0+  ,0.2726053183,0.67369766,-0.68688995,0,-0.6591295371,0.3035458599,-0.6880466294,0+  ,0.4815131379,-0.7528270071,0.4487723203,0,0.9430009463,0.1675647412,-0.2875261255,0+  ,0.434802957,0.7695304522,-0.4677277752,0,0.3931996188,0.594473625,0.7014236729,0+  ,0.7254336655,-0.603925654,0.3301814672,0,0.7590235227,-0.6506083235,0.02433313207,0+  ,-0.8552768592,-0.3430042733,0.3883935666,0,-0.6139746835,0.6981725247,0.3682257648,0+  ,-0.7465905486,-0.5752009504,0.3342849376,0,0.5730065677,0.810555537,-0.1210916791,0+  ,-0.9225877367,-0.3475211012,-0.167514036,0,-0.7105816789,-0.4719692027,-0.5218416899,0+  ,-0.08564609717,0.3583001386,0.929669703,0,-0.8279697606,-0.2043157126,0.5222271202,0+  ,0.427944023,0.278165994,0.8599346446,0,0.5399079671,-0.7857120652,-0.3019204161,0+  ,0.5678404253,-0.5495413974,-0.6128307303,0,-0.9896071041,0.1365639107,-0.04503418428,0+  ,-0.6154342638,-0.6440875597,0.4543037336,0,0.1074204368,-0.7946340692,0.5975094525,0+  ,-0.3595449969,-0.8885529948,0.28495784,0,-0.2180405296,0.1529888965,0.9638738118,0+  ,-0.7277432317,-0.6164050508,-0.3007234646,0,0.7249729114,-0.00669719484,0.6887448187,0+  ,-0.5553659455,-0.5336586252,0.6377908264,0,0.5137558015,0.7976208196,-0.3160000073,0+  ,-0.3794024848,0.9245608561,-0.03522751494,0,0.8229248658,0.2745365933,-0.4974176556,0+  ,-0.5404114394,0.6091141441,0.5804613989,0,0.8036581901,-0.2703029469,0.5301601931,0+  ,0.6044318879,0.6832968393,0.4095943388,0,0.06389988817,0.9658208605,-0.2512108074,0+  ,0.1087113286,0.7402471173,-0.6634877936,0,-0.713427712,-0.6926784018,0.1059128479,0+  ,0.6458897819,-0.5724548511,-0.5050958653,0,-0.6553931414,0.7381471625,0.159995615,0+  ,0.3910961323,0.9188871375,-0.05186755998,0,-0.4879022471,-0.5904376907,0.6429111375,0+  ,0.6014790094,0.7707441366,-0.2101820095,0,-0.5677173047,0.7511360995,0.3368851762,0+  ,0.7858573506,0.226674665,0.5753666838,0,-0.4520345543,-0.604222686,-0.6561857263,0+  ,0.002272116345,0.4132844051,-0.9105991643,0,-0.5815751419,-0.5162925989,0.6286591339,0+  ,-0.03703704785,0.8273785755,0.5604221175,0,-0.5119692504,0.7953543429,-0.3244980058,0+  ,-0.2682417366,-0.9572290247,-0.1084387619,0,-0.2322482736,-0.9679131102,-0.09594243324,0+  ,0.3554328906,-0.8881505545,0.2913006227,0,0.7346520519,-0.4371373164,0.5188422971,0+  ,0.9985120116,0.04659011161,-0.02833944577,0,-0.3727687496,-0.9082481361,0.1900757285,0+  ,0.91737377,-0.3483642108,0.1925298489,0,0.2714911074,0.4147529736,-0.8684886582,0+  ,0.5131763485,-0.7116334161,0.4798207128,0,-0.8737353606,0.18886992,-0.4482350644,0+  ,0.8460043821,-0.3725217914,0.3814499973,0,0.8978727456,-0.1780209141,-0.4026575304,0+  ,0.2178065647,-0.9698322841,-0.1094789531,0,-0.1518031304,-0.7788918132,-0.6085091231,0+  ,-0.2600384876,-0.4755398075,-0.8403819825,0,0.572313509,-0.7474340931,-0.3373418503,0+  ,-0.7174141009,0.1699017182,-0.6756111411,0,-0.684180784,0.02145707593,-0.7289967412,0+  ,-0.2007447902,0.06555605789,-0.9774476623,0,-0.1148803697,-0.8044887315,0.5827524187,0+  ,-0.7870349638,0.03447489231,0.6159443543,0,-0.2015596421,0.6859872284,0.6991389226,0+  ,-0.08581082512,-0.10920836,-0.9903080513,0,0.5532693395,0.7325250401,-0.396610771,0+  ,-0.1842489331,-0.9777375055,-0.1004076743,0,0.0775473789,-0.9111505856,0.4047110257,0+  ,0.1399838409,0.7601631212,-0.6344734459,0,0.4484419361,-0.845289248,0.2904925424,0+  ]+{- ORMOLU_ENABLE -}
src/Numeric/Noise/Fractal.hs view
@@ -44,7 +44,8 @@ data FractalConfig a = FractalConfig   { octaves :: Int   -- ^ Number of noise layers to combine. More octaves create more detail-  -- but are more expensive to compute. Must be \( >= 1 \).+  -- but are more expensive to compute. Fewer than 1 octave produces+  -- constant 0.   , lacunarity :: a   -- ^ Frequency multiplier between octaves. Each octave's frequency is   -- the previous octave's frequency multiplied by lacunarity.@@ -54,9 +55,12 @@   -- Values \( < 1 \) create smoother noise, values \( > 1 \) create rougher noise.   , weightedStrength :: a   -- ^ Controls how much each octave's amplitude is influenced by the-  -- previous octave's value. At 0, octaves have independent amplitudes.-  -- At 1, lower-valued areas in previous octaves reduce the amplitude-  -- of subsequent octaves. Range: \( [0, 1] \).+  -- previous octave's value. At 0 (the default), octaves have independent+  -- amplitudes. Range: \( [0, 1] \).+  --+  -- The weighting currently tracks the amplitude-scaled octave value, which+  -- diverges from FastNoiseLite — values near 1 can misbehave (e.g. inverted+  -- ridged octaves). It will align with FNL in 0.3.   }   deriving (Generic, Read, Show, Eq) @@ -118,6 +122,8 @@ -- Ping-pong creates a wave-like pattern by folding the noise values back -- and forth within a range, creating a distinctive undulating appearance. -- The strength parameter controls the intensity of the ping-pong effect.+--+-- Output spans @[0, 1]@; it will align with FNL's @[-1, 1]@ in 0.3. -- -- @ -- waves :: Noise2 Float
src/Numeric/Noise/Internal.hs view
@@ -47,11 +47,13 @@ -- |  'Noise' represents a function from a 'Seed' and coordinates @p@ to a noise -- value @v@. ----- For convenience, dimension-specific type aliases are provided:+-- For convenience, dimension-specific type aliases are provided: 'Noise1',+-- 'Noise2', and 'Noise3', plus primed variants that separate the coordinate+-- and value types. -- -- Use 'warp' to transform coordinates and 'remap' (or 'fmap') to transform values. ----- To evaluate noise functions, use 'noise1At', 'noise2At', or 'noise3At'+-- To evaluate noise functions, use 'noise1At', 'noise2At', or 'noise3At'. -- -- NB: 'Noise' is a lawful 'Profunctor' where 'lmap' = warp and 'rmap' = remap. -- There are some useful implications to this, but pure-noise is committed to@@ -141,23 +143,33 @@   acosh = fmap acosh   atanh = fmap atanh +-- | 1D noise: a single coordinate of type @p@ producing values of type @v@. type Noise1' p v = Noise p v++-- | 1D noise with a single type for coordinates and values. type Noise1 v = Noise1' v v +-- | Build a 1D noise function from a plain @seed -> coordinate -> value@+-- function. Inverse of 'noise1At'. mkNoise1 :: (Seed -> p -> v) -> Noise1' p v mkNoise1 = Noise {-# INLINE mkNoise1 #-}  -- | Evaluate a 1D noise function at the given coordinates with the given seed.--- Currently, you must use a slicing function like 'sliceX' to reduce+-- Currently, you must use a slicing function like 'sliceX2' to reduce -- higher-dimensional noise into 1D noise. noise1At :: Noise1 a -> Seed -> a -> a noise1At = unNoise {-# INLINE noise1At #-} +-- | 2D noise: a pair of coordinates of type @p@ producing values of type @v@. type Noise2' p v = Noise (p, p) v++-- | 2D noise with a single type for coordinates and values. type Noise2 v = Noise2' v v +-- | Build a 2D noise function from a plain @seed -> x -> y -> value@+-- function. Inverse of 'noise2At'. mkNoise2 :: (Seed -> p -> p -> v) -> Noise2' p v mkNoise2 f = Noise (\s (x, y) -> f s x y) {-# INLINE mkNoise2 #-}@@ -175,9 +187,14 @@ noise2At (Noise f) seed x y = f seed (x, y) {-# INLINE noise2At #-} +-- | 3D noise: a triple of coordinates of type @p@ producing values of type @v@. type Noise3' p v = Noise (p, p, p) v++-- | 3D noise with a single type for coordinates and values. type Noise3 v = Noise3' v v +-- | Build a 3D noise function from a plain @seed -> x -> y -> z -> value@+-- function. Inverse of 'noise3At'. mkNoise3 :: (Seed -> p -> p -> p -> v) -> Noise3' p v mkNoise3 f = Noise (\s (x, y, z) -> f s x y z) {-# INLINE mkNoise3 #-}@@ -217,7 +234,7 @@ -- This allows you to scale, rotate, or otherwise modify coordinates before -- they're passed to the noise function: ----- NB: This is 'contramap'+-- NB: This is 'Data.Functor.Contravariant.contramap' -- -- === __Examples__ --@@ -350,7 +367,7 @@ sliceZ3 z = warp (\(x, y) -> (x, y, z)) {-# INLINE sliceZ3 #-} --- | Increment the seed for a 2D noise function. See 'reseed'+-- | Increment the seed for a 2D noise function. See 'reseed'. next2 :: Noise2 a -> Noise2 a next2 = reseed (+ 1) {-# INLINE next2 #-}@@ -365,14 +382,12 @@ const2 = pure {-# INLINE const2 #-} --- | Increment the seed for a 3D noise function. See 'reseed'+-- | Increment the seed for a 3D noise function. See 'reseed'. next3 :: Noise3 a -> Noise3 a next3 = reseed (+ 1) {-# INLINE next3 #-} --- | A noise function that produces the same value everywhere. Alias of 'pure'------ Used to provide the 'Num' instance.+-- | A noise function that produces the same value everywhere. Alias of 'pure'. const3 :: a -> Noise3 a const3 = pure {-# INLINE const3 #-}
src/Numeric/Noise/Internal/Math.hs view
@@ -12,6 +12,7 @@   hermiteInterp,   quinticInterp,   clamp,+  fastRound,   primeX,   primeY,   primeZ,@@ -20,6 +21,7 @@   infinity,   g2,   sqrt3,+  rotate3,   valCoord2,   valCoord3,   gradCoord2,@@ -28,6 +30,7 @@ ) where  import Data.Bits+import Data.Bool (bool) import Data.Int import Data.Primitive.PrimArray import Data.Word@@ -37,6 +40,10 @@ -- Using the same 'Seed' value will produce the same noise pattern, -- allowing for reproducible results. Different seed values produce -- different, independent noise patterns.+--+-- Only the low 32 bits participate in hashing (matching FastNoiseLite's+-- @int@ seed): seeds that differ only in their upper 32 bits generate+-- identical noise. type Seed = Word64  -- | Internal hash value type used in noise calculations.@@ -79,7 +86,7 @@     lerp a b (t * u)   #-} --- | cubic interpolation+-- | Cubic interpolation through four control points. cubicInterp :: (Num a) => a -> a -> a -> a -> a -> a cubicInterp a !b c d !t =   let !c' = c - a@@ -110,7 +117,7 @@     0.125 * (-a + 5 * b + 5 * c - d)   #-} --- | hermite interpolation+-- | Hermite interpolation curve (smoothstep). hermiteInterp :: (Num a) => a -> a hermiteInterp t = t * t * (3 - 2 * t) {-# INLINE [1] hermiteInterp #-}@@ -122,7 +129,7 @@ "hermiteInterp/Double/1" hermiteInterp (1 :: Double) = 1   #-} --- | quintic interpolation+-- | Quintic interpolation curve (smootherstep). quinticInterp :: (Num a) => a -> a quinticInterp t = t * t * t * (t * (t * 6 - 15) + 10) {-# INLINE [1] quinticInterp #-}@@ -158,6 +165,15 @@ clamp l u v = min (max v l) u {-# INLINE clamp #-} +-- | Round half away from zero, matching FastNoiseLite's @FastRound@.+--+-- 'round' rounds half to even and, more importantly, GHC lowers it to an+-- @rintFloat@ FFI call that dominates hot loops; 'truncate' compiles to an+-- inline @float2Int#@.+fastRound :: (RealFrac a) => a -> Hash+fastRound v = truncate (v + bool (-0.5) 0.5 (v >= 0))+{-# INLINE fastRound #-}+ primeX, primeY, primeZ :: Hash primeX = 501125321 {-# INLINE primeX #-}@@ -189,6 +205,13 @@ sqrt3 :: (Fractional a) => a sqrt3 = 1.7320508075688772935274463415059 {-# INLINE sqrt3 #-}++-- | Rotation for OpenSimplex2\/2S+rotate3 :: (Fractional a) => a -> a -> a -> (a, a, a)+rotate3 xo yo zo =+  let !r = (xo + yo + zo) * (2 / 3)+   in (r - xo, r - yo, r - zo)+{-# INLINE rotate3 #-}  valCoord2 :: (RealFrac a) => Seed -> Hash -> Hash -> a valCoord2 seed xPrimed yPrimed =
src/Numeric/Noise/OpenSimplex.hs view
@@ -7,8 +7,13 @@   -- * 2D Noise   noise2,   noise2Base,++  -- * 3D Noise+  noise3,+  noise3Base, ) where +import Data.Bits (complement, shiftR, (.&.)) import Data.Bool (bool) import Numeric.Noise.Internal import Numeric.Noise.Internal.Math@@ -39,11 +44,15 @@       n0 = attenuate a seed i j x0 y0        n1 =-        let cond = bool 0 1 (y0 > x0)-            x1 = (x0 + g2 - 1) + fromIntegral cond-            y1 = (y0 + g2) - fromIntegral cond-            i1 = i + (1 - cond) * primeX-            j1 = j + cond * primeY+        let gt = y0 > x0+            -- Select the constant-folded addend per branch (g2 or g2 - 1),+            -- reproducing FNL's `x0 + (float)G2` / `x0 + ((float)G2 - 1)`+            -- rounding. Arithmetic on the selector, e.g.+            -- (x0 + g2 - 1) + cond, rounds differently when |x0 + g2| < 1.+            x1 = x0 + bool (g2 - 1) g2 gt+            y1 = y0 + bool g2 (g2 - 1) gt+            i1 = i + bool primeX 0 gt+            j1 = j + bool 0 primeY gt             b = 0.5 - x1 * x1 - y1 * y1          in attenuate b seed i1 j1 x1 y1 @@ -65,3 +74,82 @@ normalize :: (RealFrac a) => a -> a normalize = (99.83685446303647 *) {-# INLINE normalize #-}++noise3 :: (RealFrac a) => Noise3 a+noise3 = mkNoise3 noise3Base+{-# INLINE noise3 #-}++noise3Base :: forall a. (RealFrac a) => Seed -> a -> a -> a -> a+noise3Base seed0 xo yo zo =+  let !(x, y, z) = rotate3 xo yo zo++      !i0 = fastRound x+      !j0 = fastRound y+      !k0 = fastRound z+      !x0 = x - fromIntegral i0+      !y0 = y - fromIntegral j0+      !z0 = z - fromIntegral k0++      -- FNL: (int)(-1.0f - x0) | 1, i.e. -1 when the offset is >= 0+      !xns = bool 1 (-1) (x0 >= 0) :: Hash+      !yns = bool 1 (-1) (y0 >= 0) :: Hash+      !zns = bool 1 (-1) (z0 >= 0) :: Hash++      !ax0 = fromIntegral xns * negate x0+      !ay0 = fromIntegral yns * negate y0+      !az0 = fromIntegral zns * negate z0++      !ip = i0 * primeX+      !jp = j0 * primeY+      !kp = k0 * primeZ++      !a0 = (0.6 - x0 * x0) - (y0 * y0 + z0 * z0)++      !v0 = iteration seed0 a0 ip jp kp x0 y0 z0 ax0 ay0 az0 xns yns zns++      -- second lattice: reflect offsets, flip signs, complement the seed+      !ax1 = 0.5 - ax0+      !ay1 = 0.5 - ay0+      !az1 = 0.5 - az0+      !x1 = fromIntegral xns * ax1+      !y1 = fromIntegral yns * ay1+      !z1 = fromIntegral zns * az1+      !a1 = a0 + ((0.75 - ax1) - (ay1 + az1))+      !ip' = ip + ((xns `shiftR` 1) .&. primeX)+      !jp' = jp + ((yns `shiftR` 1) .&. primeY)+      !kp' = kp + ((zns `shiftR` 1) .&. primeZ)++      !v1 = iteration (complement seed0) a1 ip' jp' kp' x1 y1 z1 ax1 ay1 az1 (negate xns) (negate yns) (negate zns)+   in (v0 + v1) * 32.69428253173828125+ where+  -- one loop iteration of the FNL reference: the cell contribution plus one+  -- step along the dominant axis+  iteration+    :: Seed -> a -> Hash -> Hash -> Hash -> a -> a -> a -> a -> a -> a -> Hash -> Hash -> Hash -> a+  iteration !sd !a !i !j !k !x0 !y0 !z0 !ax !ay !az !xns !yns !zns =+    let !va+          | a > 0 = (a * a) * (a * a) * gradCoord3 sd i j k x0 y0 z0+          | otherwise = 0+        !b0 = a + 1+        !vb+          | ax >= ay && ax >= az =+              let !x1 = x0 + fromIntegral xns+                  !b = b0 - fromIntegral (xns * 2) * x1+               in bContrib sd b (i - xns * primeX) j k x1 y0 z0+          | ay > ax && ay >= az =+              let !y1 = y0 + fromIntegral yns+                  !b = b0 - fromIntegral (yns * 2) * y1+               in bContrib sd b i (j - yns * primeY) k x0 y1 z0+          | otherwise =+              let !z1 = z0 + fromIntegral zns+                  !b = b0 - fromIntegral (zns * 2) * z1+               in bContrib sd b i j (k - zns * primeZ) x0 y0 z1+     in va + vb+  {-# INLINE iteration #-}++  bContrib :: Seed -> a -> Hash -> Hash -> Hash -> a -> a -> a -> a+  bContrib !sd !b !i !j !k !x1 !y1 !z1+    | b > 0 = (b * b) * (b * b) * gradCoord3 sd i j k x1 y1 z1+    | otherwise = 0+  {-# INLINE bContrib #-}+{-# INLINE [2] noise3Base #-}
src/Numeric/Noise/SuperSimplex.hs view
@@ -4,15 +4,21 @@ -- Maintainer: Jeremy Nuttall <jeremy@jeremy-nuttall.com> -- Stability: experimental ----- This module implements a variation of OpenSimplex2 noise derived from FastNoiseLite.--- See openSimplex2S+-- This module implements a variation of OpenSimplex2S noise derived from+-- FastNoiseLite, exported from "Numeric.Noise" as 'Numeric.Noise.superSimplex2'+-- and 'Numeric.Noise.superSimplex3'. module Numeric.Noise.SuperSimplex (   -- * 2D Noise   noise2,   noise2Base,++  -- * 3D Noise+  noise3,+  noise3Base, ) where  import Data.Bits+import Data.Bool (bool) import Numeric.Noise.Internal import Numeric.Noise.Internal.Math @@ -117,3 +123,168 @@ normalize :: (RealFrac a) => a -> a normalize = (18.24196194486065 *) {-# INLINE normalize #-}++noise3 :: (RealFrac a) => Noise3 a+noise3 = mkNoise3 noise3Base+{-# INLINE noise3 #-}++noise3Base :: (RealFrac a) => Seed -> a -> a -> a -> a+noise3Base seed xo yo zo =+  let (x, y, z) = rotate3 xo yo zo++      fi = floor x :: Hash+      fj = floor y :: Hash+      fk = floor z :: Hash+      xi = x - fromIntegral fi+      yi = y - fromIntegral fj+      zi = z - fromIntegral fk++      i = fi * primeX+      j = fj * primeY+      k = fk * primeZ+      seed2 = seed + 1293373++      -- FNL: (int)(-0.5f - xi), i.e. -1 when the offset is >= 0.5, else 0+      xnm = bool 0 (-1) (xi >= 0.5) :: Hash+      ynm = bool 0 (-1) (yi >= 0.5) :: Hash+      znm = bool 0 (-1) (zi >= 0.5) :: Hash++      x0 = xi + fromIntegral xnm+      y0 = yi + fromIntegral ynm+      z0 = zi + fromIntegral znm+      a0 = 0.75 - x0 * x0 - y0 * y0 - z0 * z0+      v0 =+        q a0+          * gradCoord3 seed (i + (xnm .&. primeX)) (j + (ynm .&. primeY)) (k + (znm .&. primeZ)) x0 y0 z0++      x1 = xi - 0.5+      y1 = yi - 0.5+      z1 = zi - 0.5+      a1 = 0.75 - x1 * x1 - y1 * y1 - z1 * z1+      v1 = q a1 * gradCoord3 seed2 (i + primeX) (j + primeY) (k + primeZ) x1 y1 z1++      xFlip0 = fromIntegral ((xnm .|. 1) `shiftL` 1) * x1+      yFlip0 = fromIntegral ((ynm .|. 1) `shiftL` 1) * y1+      zFlip0 = fromIntegral ((znm .|. 1) `shiftL` 1) * z1+      xFlip1 = fromIntegral (-2 - (xnm `shiftL` 2)) * x1 - 1.0+      yFlip1 = fromIntegral (-2 - (ynm `shiftL` 2)) * y1 - 1.0+      zFlip1 = fromIntegral (-2 - (znm `shiftL` 2)) * z1 - 1.0++      a2 = xFlip0 + a0+      ~(vX, skip5)+        | a2 > 0 =+            let ~x2 = x0 - fromIntegral (xnm .|. 1)+             in ( q a2+                    * gradCoord3 seed (i + (complement xnm .&. primeX)) (j + (ynm .&. primeY)) (k + (znm .&. primeZ)) x2 y0 z0+                , False+                )+        | otherwise =+            let a3 = yFlip0 + zFlip0 + a0+                ~v3+                  | a3 > 0 =+                      let ~y3 = y0 - fromIntegral (ynm .|. 1)+                          ~z3 = z0 - fromIntegral (znm .|. 1)+                       in q a3+                            * gradCoord3 seed (i + (xnm .&. primeX)) (j + (complement ynm .&. primeY)) (k + (complement znm .&. primeZ)) x0 y3 z3+                  | otherwise = 0+                a4 = xFlip1 + a1+                ~(v4, sk)+                  | a4 > 0 =+                      let ~x4 = fromIntegral (xnm .|. 1) + x1+                       in ( q a4+                              * gradCoord3 seed2 (i + (xnm .&. (primeX * 2))) (j + primeY) (k + primeZ) x4 y1 z1+                          , True+                          )+                  | otherwise = (0, False)+             in (v3 + v4, sk)++      a6 = yFlip0 + a0+      ~(vY, skip9)+        | a6 > 0 =+            let ~y6 = y0 - fromIntegral (ynm .|. 1)+             in ( q a6+                    * gradCoord3 seed (i + (xnm .&. primeX)) (j + (complement ynm .&. primeY)) (k + (znm .&. primeZ)) x0 y6 z0+                , False+                )+        | otherwise =+            let a7 = xFlip0 + zFlip0 + a0+                ~v7+                  | a7 > 0 =+                      let ~x7 = x0 - fromIntegral (xnm .|. 1)+                          ~z7 = z0 - fromIntegral (znm .|. 1)+                       in q a7+                            * gradCoord3 seed (i + (complement xnm .&. primeX)) (j + (ynm .&. primeY)) (k + (complement znm .&. primeZ)) x7 y0 z7+                  | otherwise = 0+                a8 = yFlip1 + a1+                ~(v8, sk)+                  | a8 > 0 =+                      let ~y8 = fromIntegral (ynm .|. 1) + y1+                       in ( q a8+                              * gradCoord3 seed2 (i + primeX) (j + (ynm .&. (primeY `shiftL` 1))) (k + primeZ) x1 y8 z1+                          , True+                          )+                  | otherwise = (0, False)+             in (v7 + v8, sk)++      aA = zFlip0 + a0+      ~(vZ, skipD)+        | aA > 0 =+            let ~zA = z0 - fromIntegral (znm .|. 1)+             in ( q aA+                    * gradCoord3 seed (i + (xnm .&. primeX)) (j + (ynm .&. primeY)) (k + (complement znm .&. primeZ)) x0 y0 zA+                , False+                )+        | otherwise =+            let aB = xFlip0 + yFlip0 + a0+                ~vB+                  | aB > 0 =+                      let ~xB = x0 - fromIntegral (xnm .|. 1)+                          ~yB = y0 - fromIntegral (ynm .|. 1)+                       in q aB+                            * gradCoord3 seed (i + (complement xnm .&. primeX)) (j + (complement ynm .&. primeY)) (k + (znm .&. primeZ)) xB yB z0+                  | otherwise = 0+                aC = zFlip1 + a1+                ~(vC, sk)+                  | aC > 0 =+                      let ~zC = fromIntegral (znm .|. 1) + z1+                       in ( q aC+                              * gradCoord3 seed2 (i + primeX) (j + primeY) (k + (znm .&. (primeZ `shiftL` 1))) x1 y1 zC+                          , True+                          )+                  | otherwise = (0, False)+             in (vB + vC, sk)++      ~v5+        | not skip5+        , a5 <- yFlip1 + zFlip1 + a1+        , a5 > 0 =+            let ~y5 = fromIntegral (ynm .|. 1) + y1+                ~z5 = fromIntegral (znm .|. 1) + z1+             in q a5+                  * gradCoord3 seed2 (i + primeX) (j + (ynm .&. (primeY `shiftL` 1))) (k + (znm .&. (primeZ `shiftL` 1))) x1 y5 z5+        | otherwise = 0++      ~v9+        | not skip9+        , a9 <- xFlip1 + zFlip1 + a1+        , a9 > 0 =+            let ~x9 = fromIntegral (xnm .|. 1) + x1+                ~z9 = fromIntegral (znm .|. 1) + z1+             in q a9+                  * gradCoord3 seed2 (i + (xnm .&. (primeX * 2))) (j + primeY) (k + (znm .&. (primeZ `shiftL` 1))) x9 y1 z9+        | otherwise = 0++      ~vD+        | not skipD+        , aD <- xFlip1 + yFlip1 + a1+        , aD > 0 =+            let ~xD = fromIntegral (xnm .|. 1) + x1+                ~yD = fromIntegral (ynm .|. 1) + y1+             in q aD+                  * gradCoord3 seed2 (i + (xnm .&. (primeX `shiftL` 1))) (j + (ynm .&. (primeY `shiftL` 1))) (k + primeZ) xD yD z1+        | otherwise = 0+   in (v0 + v1 + vX + vY + vZ + v5 + v9 + vD) * 9.046026385208288+ where+  q a = (a * a) * (a * a)+  {-# INLINE q #-}+{-# INLINE [2] noise3Base #-}
test/CellularSpec.hs view
@@ -12,6 +12,8 @@     "Cellular Golden Tests"     [ testGroup "Grid Tests" cellularGridTests     , testGroup "Sparse Tests" cellularSparseTests+    , testGroup "3D Grid Tests" cellular3DGridTests+    , testGroup "3D Sparse Tests" cellular3DSparseTests     ]  -- All combinations of distance functions and result types@@ -34,4 +36,32 @@   , let config = defaultCellularConfig{cellularDistanceFn = distFn, cellularResult = result}   , seed <- cellularSeeds   , let variant = show distFn <> "-" <> show result <> "-seed" ++ show seed+  ]++-- Full 3D image goldens would add ~200 PNGs; sparse JSON covers the whole+-- config matrix, images sample one distance function across result shapes.+cellular3DImageConfigs :: [(CellularDistanceFn, CellularResult)]+cellular3DImageConfigs =+  [ (DistEuclidean, CellValue)+  , (DistEuclidean, Distance)+  , (DistEuclidean, Distance2Add)+  ]++cellular3DGridTests :: [TestTree]+cellular3DGridTests =+  [ goldenImageTest3D "cellular" variant (cellular3 config) seed zOffset+  | (distFn, result) <- cellular3DImageConfigs+  , let config = defaultCellularConfig{cellularDistanceFn = distFn, cellularResult = result}+  , seed <- cellularSeeds+  , (idx, zOffset) <- zip [0 :: Int ..] sliceOffsets3D+  , let variant = "3d-" <> show distFn <> "-" <> show result <> "-seed_" <> show seed <> "-slice_" <> show idx+  ]++cellular3DSparseTests :: [TestTree]+cellular3DSparseTests =+  [ goldenSparseTest3D "cellular" variant (cellular3 config) seed+  | (distFn, result) <- allCellularConfigs+  , let config = defaultCellularConfig{cellularDistanceFn = distFn, cellularResult = result}+  , seed <- cellularSeeds+  , let variant = "3d-" <> show distFn <> "-" <> show result <> "-seed" <> show seed   ]
test/Golden/Util.hs view
@@ -156,7 +156,6 @@   , (-3.2, 4.7)   , -- Edge cases     (1.0e-10, 1.0e-10)-  , (1.0e10, 1.0e10)   , -- More varied points     (2.5, 3.7)   , (-5.2, 8.9)@@ -194,10 +193,6 @@   , -- Large inputs     (10000.0, 10000.0)   , (10000.1, 10000.1)-  , -- Floating-point tomfoolery-    (1 / 0, 1.0) -- infinity-  , (-1 / 0, 1.0) -- negative infinity-  , (0 / 0, 1.0) -- NaN   ]  -- | Strategic test points for 3D noise@@ -231,7 +226,6 @@   , (1.23, -4.56, 7.89)   , -- Edge cases     (1.0e-10, 1.0e-10, 1.0e-10)-  , (1.0e10, 1.0e10, 1.0e10)   , -- Mixed signs     (1.0, -1.0, 1.0)   , (-1.0, 1.0, -1.0)
test/OpenSimplexSpec.hs view
@@ -12,6 +12,8 @@     "OpenSimplex Golden Tests"     [ testGroup "2D Grid Tests" openSimplex2DGridTests     , testGroup "2D Sparse Tests" openSimplex2DSparseTests+    , testGroup "3D Grid Tests" openSimplex3DGridTests+    , testGroup "3D Sparse Tests" openSimplex3DSparseTests     ]  openSimplex2DGridTests :: [TestTree]@@ -19,3 +21,9 @@  openSimplex2DSparseTests :: [TestTree] openSimplex2DSparseTests = golden2DSparseTests "opensimplex" defaultSeeds openSimplex2++openSimplex3DGridTests :: [TestTree]+openSimplex3DGridTests = golden3DImageTests "opensimplex" defaultSeeds openSimplex3++openSimplex3DSparseTests :: [TestTree]+openSimplex3DSparseTests = golden3DSparseTests "opensimplex" defaultSeeds openSimplex3
test/SuperSimplexSpec.hs view
@@ -10,6 +10,8 @@     "SuperSimplex Golden Tests"     [ testGroup "2D Grid Tests" superSimplex2DGridTests     , testGroup "2D Sparse Tests" superSimplex2DSparseTests+    , testGroup "3D Grid Tests" superSimplex3DGridTests+    , testGroup "3D Sparse Tests" superSimplex3DSparseTests     ]  superSimplex2DGridTests :: [TestTree]@@ -17,3 +19,9 @@  superSimplex2DSparseTests :: [TestTree] superSimplex2DSparseTests = golden2DSparseTests "supersimplex" defaultSeeds superSimplex2++superSimplex3DGridTests :: [TestTree]+superSimplex3DGridTests = golden3DImageTests "supersimplex" defaultSeeds superSimplex3++superSimplex3DSparseTests :: [TestTree]+superSimplex3DSparseTests = golden3DSparseTests "supersimplex" defaultSeeds superSimplex3
+ test/TotalitySpec.hs view
@@ -0,0 +1,60 @@+-- | Out-of-domain inputs (beyond the Int32 lattice, non-finite) produce+-- unspecified /values/, but they must stay values: every noise function is+-- total, so one bad coordinate upstream can't crash a whole render. These+-- assertions replace the old out-of-domain golden points, which pinned the+-- exact garbage and so broke on any implementation change.+module TotalitySpec (test_totality) where++import Control.Exception (evaluate)+import Control.Monad (forM_, void)+import Numeric.Noise+import Test.Tasty (TestTree, testGroup)+import Test.Tasty.HUnit (testCase)++-- Beyond the Int32 lattice, past Double integer precision, and non-finite.+badCoords :: [Double]+badCoords = [1.0e10, -1.0e10, 2 ^ (53 :: Int), 1 / 0, -1 / 0, 0 / 0]++noises2 :: [(String, Noise2 Double)]+noises2 =+  [ ("perlin2", perlin2)+  , ("openSimplex2", openSimplex2)+  , ("superSimplex2", superSimplex2)+  , ("value2", value2)+  , ("valueCubic2", valueCubic2)+  , ("cellular2", cellular2 defaultCellularConfig)+  , ("fractal2 perlin2", fractal2 defaultFractalConfig perlin2)+  ]++noises3 :: [(String, Noise3 Double)]+noises3 =+  [ ("perlin3", perlin3)+  , ("openSimplex3", openSimplex3)+  , ("superSimplex3", superSimplex3)+  , ("value3", value3)+  , ("valueCubic3", valueCubic3)+  , ("cellular3", cellular3 defaultCellularConfig)+  , ("fractal3 perlin3", fractal3 defaultFractalConfig perlin3)+  ]++test_totality :: TestTree+test_totality =+  testGroup+    "Totality on out-of-domain inputs"+    [ testGroup+        "2D"+        [ testCase name $+            forM_ badCoords $ \c ->+              forM_ [(c, 0.5), (0.5, c), (c, c)] $ \(x, y) ->+                void $ evaluate (noise2At noise 0 x y)+        | (name, noise) <- noises2+        ]+    , testGroup+        "3D"+        [ testCase name $+            forM_ badCoords $ \c ->+              forM_ [(c, 0.5, 0.5), (0.5, c, 0.5), (0.5, 0.5, c), (c, c, c)] $ \(x, y, z) ->+                void $ evaluate (noise3At noise 0 x y z)+        | (name, noise) <- noises3+        ]+    ]