summaryrefslogtreecommitdiffstats
path: root/webkit
diff options
context:
space:
mode:
authorojan@google.com <ojan@google.com@0039d316-1c4b-4281-b951-d872f2087c98>2009-04-08 01:32:00 +0000
committerojan@google.com <ojan@google.com@0039d316-1c4b-4281-b951-d872f2087c98>2009-04-08 01:32:00 +0000
commit3d7845d63164ac46d396f586bf041d7a4a70e435 (patch)
treec7c917454231c891b58e099632aced88dec9e928 /webkit
parent19142c82dce312511883768af6c4d4695f3f408b (diff)
downloadchromium_src-3d7845d63164ac46d396f586bf041d7a4a70e435.zip
chromium_src-3d7845d63164ac46d396f586bf041d7a4a70e435.tar.gz
chromium_src-3d7845d63164ac46d396f586bf041d7a4a70e435.tar.bz2
Use cpus instead of cpus+1 for webkit tests.
This changes the runtime from 7mins to 8m20s on my quad-core mac. Still running the test on my dual-core windows to see the perf hit. Hopefully this will reduce flakiness. Once we flesh out more of the flakiness we can up this back to cpus+1. Review URL: http://codereview.chromium.org/63105 git-svn-id: svn://svn.chromium.org/chrome/trunk/src@13328 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'webkit')
-rwxr-xr-xwebkit/tools/layout_tests/run_webkit_tests.py5
1 files changed, 2 insertions, 3 deletions
diff --git a/webkit/tools/layout_tests/run_webkit_tests.py b/webkit/tools/layout_tests/run_webkit_tests.py
index e1414ed..476a96c 100755
--- a/webkit/tools/layout_tests/run_webkit_tests.py
+++ b/webkit/tools/layout_tests/run_webkit_tests.py
@@ -811,9 +811,8 @@ def main(options, args):
elif sys.platform in ('darwin'): # OSX:
cpus = int(os.popen2("sysctl -n hw.ncpu")[1].read())
- # TODO: Do timing tests 1/2/4/8 cores and mac, linux, win to see what
- # the optimal number here is.
- options.num_test_shells = cpus + 1
+ # TODO(ojan): Use cpus+1 once we flesh out the flakiness.
+ options.num_test_shells = cpus
else:
options.num_test_shells = 1