Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

huff0: Speed up compression of short blocks #744

Merged
merged 1 commit into from
Jan 20, 2023
Merged

huff0: Speed up compression of short blocks #744

merged 1 commit into from
Jan 20, 2023

Commits on Jan 20, 2023

  1. huff0: Speed up compression of short blocks 

    Tells the compiler to merge loads and stores of nodeElts.

name                                      old speed      new speed      delta
Compress4XReuseNone/digits-8               433MB/s ± 1%   434MB/s ± 0%     ~     (p=0.393 n=10+10)
Compress4XReuseNone/gettysburg-8           245MB/s ± 1%   254MB/s ± 1%   +3.55%  (p=0.000 n=10+10)
Compress4XReuseNone/twain-8                363MB/s ± 0%   363MB/s ± 0%     ~     (p=0.360 n=8+10)
Compress4XReuseNone/low-ent.10k-8          465MB/s ± 0%   467MB/s ± 0%   +0.32%  (p=0.008 n=9+10)
Compress4XReuseNone/superlow-ent-10k-8     305MB/s ± 2%   304MB/s ± 1%     ~     (p=0.143 n=10+10)
Compress4XReuseNone/case1-8               13.1MB/s ± 1%  14.6MB/s ± 2%  +11.44%  (p=0.000 n=10+10)
Compress4XReuseNone/case2-8               10.9MB/s ± 1%  12.3MB/s ± 2%  +12.28%  (p=0.000 n=10+9)
Compress4XReuseNone/case3-8               11.5MB/s ± 1%  12.9MB/s ± 2%  +12.37%  (p=0.000 n=10+10)
Compress4XReuseNone/pngdata.001-8          300MB/s ± 0%   298MB/s ± 2%     ~     (p=0.762 n=8+10)
Compress4XReuseNone/normcount2-8          31.5MB/s ± 1%  33.7MB/s ± 2%   +6.87%  (p=0.000 n=9+10)
Compress4XReuseAllow/digits-8              435MB/s ± 1%   432MB/s ± 2%     ~     (p=0.043 n=10+10)
Compress4XReuseAllow/gettysburg-8          272MB/s ± 1%   281MB/s ± 1%   +3.19%  (p=0.000 n=10+10)
Compress4XReuseAllow/twain-8               363MB/s ± 0%   361MB/s ± 1%     ~     (p=0.150 n=9+10)
Compress4XReuseAllow/low-ent.10k-8         469MB/s ± 1%   469MB/s ± 0%     ~     (p=0.616 n=10+10)
Compress4XReuseAllow/superlow-ent-10k-8    305MB/s ± 0%   306MB/s ± 0%   +0.44%  (p=0.001 n=9+9)
Compress4XReuseAllow/case1-8              15.8MB/s ± 0%  18.0MB/s ± 2%  +13.88%  (p=0.000 n=9+10)
Compress4XReuseAllow/case2-8              13.3MB/s ± 0%  15.4MB/s ± 1%  +15.86%  (p=0.000 n=9+10)
Compress4XReuseAllow/case3-8              14.1MB/s ± 0%  15.8MB/s ± 1%  +12.15%  (p=0.000 n=10+8)
Compress4XReuseAllow/pngdata.001-8         302MB/s ± 1%   304MB/s ± 0%   +0.68%  (p=0.000 n=9+10)
Compress4XReuseAllow/normcount2-8         40.2MB/s ± 1%  43.8MB/s ± 2%   +8.86%  (p=0.000 n=10+10)
Compress4XReusePrefer/digits-8             436MB/s ± 0%   436MB/s ± 1%     ~     (p=0.560 n=9+9)
Compress4XReusePrefer/gettysburg-8         422MB/s ± 1%   421MB/s ± 2%     ~     (p=0.579 n=10+10)
Compress4XReusePrefer/twain-8              363MB/s ± 1%   365MB/s ± 0%     ~     (p=0.018 n=10+10)
Compress4XReusePrefer/low-ent.10k-8        473MB/s ± 0%   472MB/s ± 0%     ~     (p=0.021 n=7+8)
Compress4XReusePrefer/superlow-ent-10k-8   312MB/s ± 1%   312MB/s ± 0%     ~     (p=0.278 n=10+9)
Compress4XReusePrefer/case1-8              134MB/s ± 1%   134MB/s ± 1%     ~     (p=0.780 n=9+10)
Compress4XReusePrefer/case2-8              122MB/s ± 2%   122MB/s ± 0%     ~     (p=1.000 n=10+8)
Compress4XReusePrefer/case3-8              129MB/s ± 3%   129MB/s ± 0%     ~     (p=0.698 n=10+10)
Compress4XReusePrefer/pngdata.001-8        313MB/s ± 1%   313MB/s ± 0%     ~     (p=0.481 n=10+10)
Compress4XReusePrefer/normcount2-8         179MB/s ± 2%   182MB/s ± 0%   +1.96%  (p=0.000 n=10+9)
Compress4XSizes/digits-100-8              57.6MB/s ± 1%  61.7MB/s ± 1%   +7.11%  (p=0.000 n=9+10)
Compress4XSizes/digits-200-8               100MB/s ± 2%   108MB/s ± 2%   +7.90%  (p=0.000 n=10+10)
Compress4XSizes/digits-500-8               191MB/s ± 1%   201MB/s ± 1%   +5.55%  (p=0.000 n=10+9)
Compress4XSizes/digits-1000-8              273MB/s ± 1%   283MB/s ± 1%   +3.75%  (p=0.000 n=10+10)
Compress4XSizes/digits-5000-8              416MB/s ± 0%   418MB/s ± 1%     ~     (p=0.165 n=10+10)
Compress4XSizes/digits-10000-8             434MB/s ± 1%   437MB/s ± 1%   +0.65%  (p=0.002 n=10+10)
Compress4XSizes/digits-50000-8             434MB/s ± 0%   434MB/s ± 0%     ~     (p=0.604 n=10+9)
[Geo mean]                                 152MB/s        157MB/s        +3.33%
    @greatroar
    greatroar committed Jan 20, 2023
    Configuration menu
    Copy the full SHA
    292f2d2 View commit details
    Browse the repository at this point in the history