[Webkit-unassigned] [Bug 115722] [sh4] Take advantage of pre-decrement and post-increment opcodes for sh4 base JIT

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue May 7 05:31:43 PDT 2013


https://bugs.webkit.org/show_bug.cgi?id=115722





--- Comment #4 from Julien Brianceau <jbrianceau at nds.com>  2013-05-07 05:30:07 PST ---
I didn't see any regression when testing this patch with:
- Tools/Scripts/run-javascriptcore-tests
- Tools/Scripts/run-fast-jsc


Very little speed improvement according to SunSpider with this patch:
Rev 149664 without patch:
============================================
RESULTS (means and 95% confidence intervals)
--------------------------------------------
Total:                 13461.8ms +/- 0.3%
--------------------------------------------

  3d:                   2751.5ms +/- 0.2%
    cube:                838.5ms +/- 0.7%
    morph:               500.5ms +/- 0.2%
    raytrace:           1412.5ms +/- 0.4%

  access:               1948.0ms +/- 0.1%
    binary-trees:        136.5ms +/- 0.7%
    fannkuch:           1176.8ms +/- 0.4%
    nbody:               454.8ms +/- 1.2%
    nsieve:              180.0ms +/- 1.8%

  bitops:               1051.8ms +/- 0.5%
    3bit-bits-in-byte:    94.8ms +/- 0.8%
    bits-in-byte:        122.0ms +/- 0.0%
    bitwise-and:         202.8ms +/- 2.7%
    nsieve-bits:         632.3ms +/- 0.8%

  controlflow:           105.5ms +/- 4.5%
    recursive:           105.5ms +/- 4.5%

  crypto:               1236.0ms +/- 0.2%
    aes:                 725.0ms +/- 0.8%
    md5:                 328.3ms +/- 1.0%
    sha1:                182.8ms +/- 0.4%

  date:                 1983.0ms +/- 0.7%
    format-tofte:        927.3ms +/- 0.6%
    format-xparb:       1055.8ms +/- 1.8%

  math:                  988.5ms +/- 2.1%
    cordic:              289.8ms +/- 2.0%
    partial-sums:        395.5ms +/- 1.5%
    spectral-norm:       303.3ms +/- 6.0%

  regexp:                563.3ms +/- 0.1%
    dna:                 563.3ms +/- 0.1%

  string:               2834.3ms +/- 0.8%
    base64:              224.5ms +/- 3.3%
    fasta:               459.0ms +/- 9.8%
    tagcloud:            568.3ms +/- 0.9%
    unpack-code:        1242.5ms +/- 4.2%
    validate-input:      340.0ms +/- 2.5%




Rev 149664 with patch:
============================================
RESULTS (means and 95% confidence intervals)
--------------------------------------------
Total:                 13311.0ms +/- 0.5%
--------------------------------------------

  3d:                   2599.8ms +/- 0.4%
    cube:                834.5ms +/- 0.4%
    morph:               446.3ms +/- 0.9%
    raytrace:           1319.0ms +/- 1.0%

  access:               1957.3ms +/- 0.5%
    binary-trees:        137.3ms +/- 3.7%
    fannkuch:           1180.0ms +/- 1.1%
    nbody:               456.5ms +/- 0.9%
    nsieve:              183.5ms +/- 0.5%

  bitops:               1059.3ms +/- 0.4%
    3bit-bits-in-byte:    94.5ms +/- 1.7%
    bits-in-byte:        121.5ms +/- 0.8%
    bitwise-and:         202.0ms +/- 0.0%
    nsieve-bits:         641.3ms +/- 0.8%

  controlflow:           103.3ms +/- 0.8%
    recursive:           103.3ms +/- 0.8%

  crypto:               1236.3ms +/- 0.9%
    aes:                 739.8ms +/- 1.1%
    md5:                 313.5ms +/- 0.5%
    sha1:                183.0ms +/- 3.0%

  date:                 2020.0ms +/- 1.5%
    format-tofte:        949.0ms +/- 1.6%
    format-xparb:       1071.0ms +/- 3.4%

  math:                  987.0ms +/- 2.7%
    cordic:              293.0ms +/- 2.3%
    partial-sums:        410.8ms +/- 7.9%
    spectral-norm:       283.3ms +/- 0.3%

  regexp:                532.5ms +/- 1.2%
    dna:                 532.5ms +/- 1.2%

  string:               2815.8ms +/- 1.1%
    base64:              213.8ms +/- 2.0%
    fasta:               452.5ms +/- 9.2%
    tagcloud:            588.8ms +/- 3.3%
    unpack-code:        1189.3ms +/- 4.8%
    validate-input:      371.5ms +/- 2.9%

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.


More information about the webkit-unassigned mailing list