There is a sort of corner case regression caused by this patch, if the number of tests are less than 50 they are effectively always run serially. This can have a pretty big performance impact, especially if any of those tests takes a long time to run. Reducing this to 1 results in a very slight performance reduction of very long runs (~2 seconds difference on quick.tests), but the trade off for short runs is much more significant.
Signed-off-by: Dylan Baker <[email protected]> --- framework/core.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/framework/core.py b/framework/core.py index a193a2d..a335033 100644 --- a/framework/core.py +++ b/framework/core.py @@ -547,7 +547,7 @@ class TestProfile: # The default value of pool is the number of virtual processor cores single = multiprocessing.dummy.Pool(1) multi = multiprocessing.dummy.Pool() - chunksize = 50 + chunksize = 1 if env.concurrent == "all": multi.imap(test, self.test_list.iteritems(), chunksize) -- 1.8.5.3 _______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
