diff --git a/CHANGELOG.md b/CHANGELOG.md --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,5 +1,14 @@ # bz2 +## 1.0.0.2 + + * Change buffering parameters for speed + * Add benchmarks relative to `bzlib`, `bzip-pipes`, `bzlib-conduit` + +## 1.0.0.1 + + * Pass `-O3` to `cc-options` + ## 1.0.0.0 * Remove `Codec.Compression.BZip.Foreign` diff --git a/Makefile b/Makefile --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ .PHONY: clean SHELL := bash -MAKEFLAGS += --warn-undefined-variables --no-builtin-rules -j +MAKEFLAGS += --warn-undefined-variables --no-builtin-rules .DELETE_ON_ERROR: setup: valgrind-3.15.0.tar diff --git a/bench/Bench.cpphs b/bench/Bench.cpphs new file mode 100644 --- /dev/null +++ b/bench/Bench.cpphs @@ -0,0 +1,76 @@ +{-# LANGUAGE PackageImports #-} + +module Main (main) where + +-- TODO: bench bzlib? +import qualified "bz2" Codec.Compression.BZip as BZ2 +#ifdef WITH_BZLIB +import qualified "bzlib" Codec.Compression.BZip as BZlib +#endif +import Conduit (runResourceT) +import Criterion.Main +import qualified Data.ByteString.Lazy as BSL +import Data.Conduit (runConduit, (.|)) +import qualified Data.Conduit.BZlib as C +import Data.Conduit.Combinators (sinkFile, sourceFile) +import Pipes (runEffect, (>->)) +import qualified Pipes.ByteString as P +import qualified Pipes.BZip as P +import Pipes.Safe (runSafeT) +import System.FilePath (()) +import System.IO (IOMode (ReadMode, WriteMode), withFile) +import System.IO.Temp (withSystemTempDirectory) + +decompressDump :: (BSL.ByteString -> BSL.ByteString) -> IO () +decompressDump go = withSystemTempDirectory "bz2" $ + \fp -> BSL.writeFile (fp "valgrind-3.15.0.tar") =<< + (go <$> BSL.readFile "valgrind-3.15.0.tar.bz2") + +compressDump :: (BSL.ByteString -> BSL.ByteString) -> IO () +compressDump go = withSystemTempDirectory "bz2" $ + \fp -> BSL.writeFile (fp "valgrind-3.15.0.tar.bz2") =<< + (go <$> BSL.readFile "valgrind-3.15.0.tar") + +decompressDumpPipes :: IO () +decompressDumpPipes = withSystemTempDirectory "bz2" $ \fp -> + withFile "valgrind-3.15.0.tar.bz2" ReadMode $ \hIn -> + withFile (fp "valgrind-3.15.0.tar") WriteMode $ \hOut -> + runSafeT $ runEffect $ P.bunzip2 (P.fromHandle hIn) >-> P.toHandle hOut + +decompressDumpConduit :: IO () +decompressDumpConduit = withSystemTempDirectory "bz2" $ \fp -> + runResourceT $ runConduit $ sourceFile "valgrind-3.15.0.tar.bz2" .| C.bunzip2 .| sinkFile (fp "valgrind-3.15.0.tar") + +decompressFile :: FilePath -> IO BSL.ByteString +decompressFile = fmap BZ2.decompress . BSL.readFile + +compressFile :: FilePath -> IO BSL.ByteString +compressFile = fmap BZ2.compress . BSL.readFile + +main :: IO () +main = + defaultMain [ bgroup "decompressDump" + [ bench "decompress + write to file (bz2)" $ nfIO (decompressDump BZ2.decompress) +#ifdef WITH_BZLIB + , bench "decompress + write to file (bzlib)" $ nfIO (decompressDump BZlib.decompress) +#endif + , bench "decompress + write to file (pipes-bzip)" $ nfIO decompressDumpPipes + , bench "decompress + write to file (bzlib-conduit)" $ nfIO decompressDumpConduit + ] + , bgroup "compressDump" + [ bench "compress + write to file (bz2)" $ nfIO (compressDump BZ2.compress) +#ifdef WITH_BZLIB + , bench "compress + write to file (bzlib)" $ nfIO (compressDump BZlib.compress) +#endif + ] + , bgroup "decompress" + [ bench "decompress file" $ nfIO (decompressFile "test/data/sample1.bz2") + , bench "decompress file" $ nfIO (decompressFile "test/data/sample2.bz2") + , bench "decompress file" $ nfIO (decompressFile "test/data/sample3.bz2") + ] + , bgroup "compress" + [ bench "compress file" $ nfIO (compressFile "test/data/sample1.ref") + , bench "compress file" $ nfIO (compressFile "test/data/sample2.ref") + , bench "compress file" $ nfIO (compressFile "test/data/sample3.ref") + ] + ] diff --git a/bench/Bench.hs b/bench/Bench.hs deleted file mode 100644 --- a/bench/Bench.hs +++ /dev/null @@ -1,35 +0,0 @@ -module Main (main) where - --- TODO: bench bzlib? -import qualified Codec.Compression.BZip as BZ2 -import Criterion.Main -import qualified Data.ByteString.Lazy as BSL -import System.FilePath (()) -import System.IO.Temp (withSystemTempDirectory) - -decompressDump :: IO () -decompressDump = withSystemTempDirectory "bz2" $ - \fp -> BSL.writeFile (fp "valgrind-3.15.0.tar") =<< - (BZ2.decompress <$> BSL.readFile "valgrind-3.15.0.tar.bz2") - -decompressFile :: FilePath -> IO BSL.ByteString -decompressFile = fmap BZ2.decompress . BSL.readFile - -compressFile :: FilePath -> IO BSL.ByteString -compressFile = fmap BZ2.compress . BSL.readFile - -main :: IO () -main = - defaultMain [ bgroup "decompressDump" - [ bench "decompress + write to file" $ nfIO decompressDump ] - , bgroup "decompress" - [ bench "decompress file" $ nfIO (decompressFile "test/data/sample1.bz2") - , bench "decompress file" $ nfIO (decompressFile "test/data/sample2.bz2") - , bench "decompress file" $ nfIO (decompressFile "test/data/sample3.bz2") - ] - , bgroup "compress" - [ bench "compress file" $ nfIO (compressFile "test/data/sample1.ref") - , bench "compress file" $ nfIO (compressFile "test/data/sample2.ref") - , bench "compress file" $ nfIO (compressFile "test/data/sample3.ref") - ] - ] diff --git a/bz2.cabal b/bz2.cabal --- a/bz2.cabal +++ b/bz2.cabal @@ -1,6 +1,6 @@ cabal-version: 1.18 name: bz2 -version: 1.0.0.1 +version: 1.0.0.2 license: BSD3 license-file: LICENSE copyright: Copyright: (c) 2020 Vanessa McHale @@ -8,7 +8,7 @@ author: Vanessa McHale bug-reports: https://hub.darcs.net/vmchale/bz2/issues synopsis: Bindings to libbz2 -description: High-level bindings to libbz2 using c2hs and ByteString +description: High-level bindings to libbz2 via ByteString category: Compression build-type: Simple data-files: @@ -36,6 +36,10 @@ default: False manual: True +flag with-bzlib + description: + Bench against [bzlib](http://hackage.haskell.org/package/bzlib) + library exposed-modules: Codec.Compression.BZip cc-options: -O3 @@ -63,7 +67,7 @@ install-includes: cbits/bzlib.h ghc-options: -Wall build-depends: - base >=4.4 && <5, + base >=4.7 && <5, bytestring -any if !flag(cross) @@ -106,18 +110,32 @@ ghc-options: -Wcpp-undef benchmark bz2-bench - type: exitcode-stdio-1.0 - main-is: Bench.hs - hs-source-dirs: bench - default-language: Haskell2010 - ghc-options: -threaded -rtsopts -with-rtsopts=-N -O3 -Wall + type: exitcode-stdio-1.0 + main-is: Bench.hs + build-tool-depends: cpphs:cpphs -any + hs-source-dirs: bench + default-language: Haskell2010 + other-extensions: PackageImports + ghc-options: -threaded -rtsopts -with-rtsopts=-N -O3 -Wall build-depends: base -any, bz2 -any, criterion -any, filepath -any, temporary -any, + pipes-bzip, + pipes-bytestring, + pipes, + pipes-safe, + bzlib-conduit, + conduit >=1.3.0, bytestring -any + + if flag(with-bzlib) + build-depends: bzlib -any + + if flag(with-bzlib) + cpp-options: -DWITH_BZLIB if impl(ghc >=8.0) ghc-options: diff --git a/src/Codec/Compression/BZip/Pack.chs b/src/Codec/Compression/BZip/Pack.chs --- a/src/Codec/Compression/BZip/Pack.chs +++ b/src/Codec/Compression/BZip/Pack.chs @@ -89,12 +89,12 @@ keepBytesAlive :: Ptr BzStream -> Maybe BS.ByteString -> [BS.ByteString] -> (BZAction -> IO BZError) -> IO (BZError, Maybe BS.ByteString, [BS.ByteString]) keepBytesAlive _ Nothing [] act = (, Nothing, []) <$> act BzFinish keepBytesAlive _ Nothing bs' act = (, Nothing, bs') <$> act BzRun - keepBytesAlive _ passFwd@(Just b) [] act = do - BS.unsafeUseAsCStringLen b $ \_ -> do + keepBytesAlive _ passFwd@(Just b) [] act = + BS.unsafeUseAsCStringLen b $ \_ -> (, passFwd, []) <$> act BzFinish - keepBytesAlive _ passFwd@(Just b) bs' act = do - BS.unsafeUseAsCStringLen b $ \_ -> do + keepBytesAlive _ passFwd@(Just b) bs' act = + BS.unsafeUseAsCStringLen b $ \_ -> (, passFwd, bs') <$> act BzRun @@ -109,7 +109,7 @@ (, Just b, bs') <$> act BzRun bufSz :: Integral a => a -bufSz = 32 * 1024 +bufSz = 16 * 1024 bzCompressInit :: CInt -> CInt -> ForeignPtr BzStream -> IO () bzCompressInit blkSize wf ptr' = do diff --git a/src/Codec/Compression/BZip/Unpack.chs b/src/Codec/Compression/BZip/Unpack.chs --- a/src/Codec/Compression/BZip/Unpack.chs +++ b/src/Codec/Compression/BZip/Unpack.chs @@ -48,24 +48,24 @@ withForeignPtr pForeign $ \p -> withForeignPtr bufOutForeign $ \bufOut -> do - let act = do + let act = do - {# set bz_stream.avail_out #} p bufSz - {# set bz_stream.next_out #} p (castPtr bufOut) + {# set bz_stream.avail_out #} p bufSz + {# set bz_stream.next_out #} p (castPtr bufOut) - bZ2BzDecompress ptr' + bZ2BzDecompress ptr' - (ret, keepAlive, bs'') <- step p passFwd bs' act + (ret, keepAlive, bs'') <- step p passFwd bs' act - szOut <- fromIntegral <$> {# get bz_stream->avail_out #} p + szOut <- fromIntegral <$> {# get bz_stream->avail_out #} p - let bytesAvail = bufSz - szOut + let bytesAvail = bufSz - szOut - newBSAp <- if bytesAvail /= 0 - then (:) <$> BS.packCStringLen (castPtr bufOut, bytesAvail) - else pure id + newBSAp <- if bytesAvail /= 0 + then (:) <$> BS.packCStringLen (castPtr bufOut, bytesAvail) + else pure id - pure (ret, szOut, newBSAp, bs'', keepAlive) + pure (ret, szOut, newBSAp, bs'', keepAlive) let step' = if szOut == 0 then keepBytesAlive @@ -78,7 +78,7 @@ keepBytesAlive :: Ptr BzStream -> Maybe BS.ByteString -> [BS.ByteString] -> IO BZError -> IO (BZError, Maybe BS.ByteString, [BS.ByteString]) keepBytesAlive _ Nothing bs' act = (, Nothing, bs') <$> act keepBytesAlive _ passFwd@(Just b) bs' act = do - BS.unsafeUseAsCStringLen b $ \_ -> do + BS.unsafeUseAsCStringLen b $ \_ -> (, passFwd, bs') <$> act @@ -93,12 +93,13 @@ (, Just b, bs') <$> act bufSz :: Integral a => a -bufSz = 32 * 1024 +bufSz = 64 * 1024 bzDecompressInit :: ForeignPtr BzStream -> IO () bzDecompressInit ptr' = do withForeignPtr ptr' $ \p -> do + {# set bz_stream.next_in #} p nullPtr {# set bz_stream.avail_in #} p 0