As a continuation of the count words challenge, let's make Haskell fast!
This repo is ordered primarily after different ways of counting in the src/Count folders
and modules.
Hopefully, this can serve as an experimentation ground for different ways of counting items.
- Download nix with flakes
- Run
nix develop
- This also installs the haskell language server and other nice things for development.
- Run
cabal bench --enable-benchmarks -O2 --benchmark-options '--output index.html --csv stats/output.csv'
Three kinds of benchmarks:
numbers: Counting the same number of numbers as words in kjvbibledata/numbers.txtkjvbible: Counting all words in thedata/kjvbible.txt:10xkjvbible: Counting all words in thedata/kjvbible.txt, 10 times:
| Name | Mean [ms] | Stddev [ms] |
|---|---|---|
| numbers/lengthBaseline | 7.74 | 0.18 |
| numbers/viaIntCounter | 10.73 | 1.04 |
| numbers/viaFinite | 31.10 | 1.61 |
| numbers/viaUnboxedVectorHash | 31.13 | 2.31 |
| numbers/viaVectorHashMap | 50.43 | 2.51 |
| numbers/viaDiscrimination | 80.69 | 4.47 |
| numbers/viaStrictHashMap | 113.74 | 7.63 |
| numbers/viaIntMap | 451.08 | 17.36 |
| numbers/viaStrictMap | 595.45 | 14.78 |
| Name | Mean [ms] | Stddev [ms] |
|---|---|---|
| kjvbible/lengthBaseline | 6.24 | 0.19 |
| kjvbible/viaFinite | 51.56 | 3.43 |
| kjvbible/viaVectorHashMap | 69.18 | 2.60 |
| kjvbible/viaStrictHashMap | 278.48 | 6.07 |
| kjvbible/viaStrictMap | 315.82 | 9.40 |
| kjvbible/viaDiscrimination | 399.38 | 20.89 |
| Name | Mean [ms] | Stddev [ms] |
|---|---|---|
| 10xkjvbible/lengthBaseline | 22.40 | 0.32 |
| 10xkjvbible/viaFinite | 452.61 | 41.70 |
| 10xkjvbible/viaVectorHashMap | 639.77 | 16.51 |
| 10xkjvbible/viaStrictHashMap | 3159.40 | 250.81 |
| 10xkjvbible/viaDiscrimination | 3398.02 | 50.65 |
| 10xkjvbible/viaStrictMap | 3563.63 | 100.53 |
| Name | Mean [ms] | Stddev [ms] |
|---|---|---|
| kjvbible/lengthBaseline | 6.24 | 0.19 |
| numbers/lengthBaseline | 7.74 | 0.18 |
| numbers/viaIntCounter | 10.73 | 1.04 |
| 10xkjvbible/lengthBaseline | 22.40 | 0.32 |
| numbers/viaFinite | 31.10 | 1.61 |
| numbers/viaUnboxedVectorHash | 31.13 | 2.31 |
| numbers/viaVectorHashMap | 50.43 | 2.51 |
| kjvbible/viaFinite | 51.56 | 3.43 |
| kjvbible/viaVectorHashMap | 69.18 | 2.60 |
| numbers/viaDiscrimination | 80.69 | 4.47 |
| numbers/viaStrictHashMap | 113.74 | 7.63 |
| kjvbible/viaStrictHashMap | 278.48 | 6.07 |
| kjvbible/viaStrictMap | 315.82 | 9.40 |
| kjvbible/viaDiscrimination | 399.38 | 20.89 |
| numbers/viaIntMap | 451.08 | 17.36 |
| 10xkjvbible/viaFinite | 452.61 | 41.70 |
| numbers/viaStrictMap | 595.45 | 14.78 |
| 10xkjvbible/viaVectorHashMap | 639.77 | 16.51 |
| 10xkjvbible/viaStrictHashMap | 3159.40 | 250.81 |
| 10xkjvbible/viaDiscrimination | 3398.02 | 50.65 |
| 10xkjvbible/viaStrictMap | 3563.63 | 100.53 |
Are welcome and encouraged!
data/kjvbible.txtis borrowed from the https://github.com/benhoyt/countwords repository.src/IntCounter.hsis borrowed from Jaro Reinders (https://github.com/noughtmare/clutter/blob/main/src/IntCounter.hs)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent