benchmark for https://github.com/bos/text/pull/18

bench-decodeLatin1.hs

module Main where

import           Control.DeepSeq
import           Control.Exception     (evaluate)
import           Criterion.Main
import qualified Data.ByteString.Char8 as B
import qualified Data.Text             as T
import qualified Data.Text.Encoding    as TE
import qualified Data.Text.ICU.Convert as TI

test7bit_0  = B.empty
test7bit_16 = B.pack $ take 16 $ cycle ['\0' .. '\x7f']
test7bit_1K = B.pack $ take (1024) $ cycle ['\0' .. '\x7f']
test7bit_1M = B.pack $ take (1024*1024) $ cycle ['\0' .. '\x7f']

decodeL1 = T.pack . B.unpack

ttest7bit_0  = decodeL1 test7bit_0
ttest7bit_16 = decodeL1 test7bit_16
ttest7bit_1K = decodeL1 test7bit_1K
ttest7bit_1M = decodeL1 test7bit_1M

instance NFData B.ByteString

main :: IO ()
main = do
    cvtL1   <- TI.open "ISO-8859-1" Nothing
    cvtUtf8 <- TI.open "UTF8" Nothing
    let decodeIcuL1 = TI.toUnicode cvtL1
        decodeIcuUtf8 = TI.toUnicode cvtUtf8

    evaluate $ rnf [ test7bit_0, test7bit_16, test7bit_1K, test7bit_1M ]
    evaluate $ rnf [ ttest7bit_0, ttest7bit_16, ttest7bit_1K, ttest7bit_1M ]

    defaultMain
        [ bgroup "id"            [ bench "0" $ nf id ttest7bit_0
                                 , bench "16" $ nf id ttest7bit_16
                                 , bench "1K" $ nf id ttest7bit_1K
                                 -- , bench "1M" $ nf id ttest7bit_1M
                                 ]
        , bgroup "decodeIcuUtf8" [ bench "0" $ nf decodeIcuUtf8 test7bit_0
                                 , bench "16" $ nf decodeIcuUtf8 test7bit_16
                                 , bench "1K" $ nf decodeIcuUtf8 test7bit_1K
                                 -- , bench "1M" $ nf decodeIcuUtf8 test7bit_1M
                                 ]

        , bgroup "TE.decodeUtf8" [ bench "0" $ nf TE.decodeUtf8 test7bit_0
                                 , bench "16" $ nf TE.decodeUtf8 test7bit_16
                                 , bench "1K" $ nf TE.decodeUtf8 test7bit_1K
                                 -- , bench "1M" $ nf TE.decodeUtf8 test7bit_1M
                                 ]

        , bgroup "TE.decodeLatin1" [ bench "0" $ nf TE.decodeLatin1 test7bit_0
                                   , bench "16" $ nf TE.decodeLatin1 test7bit_16
                                   , bench "1K" $ nf TE.decodeLatin1 test7bit_1K
                                   -- , bench "1M" $ nf TE.decodeLatin1 test7bit_1M
                                   ]

        , bgroup "decodeIcuL1"   [ bench "0" $ nf decodeIcuL1 test7bit_0
                                 , bench "16" $ nf decodeIcuL1 test7bit_16
                                 , bench "1K" $ nf decodeIcuL1 test7bit_1K
                                 -- , bench "1M" $ nf decodeIcuL1 test7bit_1M
                                 ]

        -- , bgroup "decodeL1(naive)" [ bench "0" $ nf decodeL1 test7bit_0
        --                          , bench "16" $ nf decodeL1 test7bit_16
        --                          , bench "1K" $ nf decodeL1 test7bit_1K
        --                          , bench "1M" $ nf decodeL1 test7bit_1M
        --                          ]
        ]