On Fri, 2014-09-26 at 07:22 -0700, Tom Stellard wrote:
> On Thu, Sep 25, 2014 at 10:19:16PM -0400, Jan Vesely wrote:
> > Didn't notice somebody dropped Tom from the cc list.
> > 
> > On Thu, 2014-09-04 at 10:28 -0400, Jan Vesely wrote:
> > > On Wed, 2014-09-03 at 20:20 -0700, Dylan Baker wrote:
> > > > I'd caught this error in a branch as well, but noticed there were no 
> > > > users
> > > > of this function so it didn't seem pressing to fix it.
> > > > 
> > > > If you wanted a little less code:
> > > > group[name] = PiglitTest(args, run_concurrent=True)
> > > 
> > > yeah, it looks nicer.
> > > 
> > > > 
> > > > 
> > > > On Wed, Sep 3, 2014 at 4:35 PM, Ilia Mirkin <imir...@alum.mit.edu> 
> > > > wrote:
> > > > 
> > > > > On Wed, Sep 3, 2014 at 7:28 PM, Tom Stellard <t...@stellard.net> 
> > > > > wrote:
> > > > > > On Wed, Sep 03, 2014 at 07:20:10PM -0400, Jan Vesely wrote:
> > > > > >> Signed-off-by: Jan Vesely <jan.ves...@rutgers.edu>
> > > > > >
> > > > > > Reviewed-by: Tom Stellard <thomas.stell...@amd.com>
> > > 
> > > thanks, can you push it too? I don't have commit access. Or Dylan's
> > > version if it is preferable.
> > > 
> > > jan
> > > 
> > > > > >
> > > > > >> ---
> > > > > >>
> > > > > >> seemed wrong. Wonder if we can use concurrent tests.
> > > > > >>
> > > > > >
> > > > > > If you have render-nodes enabled, then you can run the cl tests
> > > > > > concurrently by adding the -c flag when running piglit.
> > 
> > It works nicely (I get linear speedup to number of cores with -c).
> > 
> > I think in this case we can remove add_concurrent_test entirely, since
> > we are not going to have a mix of concurrent and non-concurrent tests.
> > 
> 
> Hi,
> 
> I've pushed this patch with Dylan's suggest change.  I left the 
> add_concurrent_test
> function in, because we will use it eventually once more people are using
> render nodes.

I thought in that case we would switch all tests to concurrent, and drop
add_plain_test.
or we can do something like:

run_concurrent = glob.glob('/dev/dri/render*')

so we don't need to check kernel version or platform (I assume it
returns empty list on mac/win)

jan

> 
> -Tom
> 
> > jan
> > 
> > > > > >
> > > > > > Render-nodes were enabled by default in the 3.16 kernel, so
> > > > > > I think it may be too early to enable concurrent tests by default.
> > > > >
> > > > > Conversely, you can use -1 to force single-threaded runs... seems like
> > > > > it's not for piglit to worry about that. -c forces all tests to run
> > > > > concurrently even if the test author said that it shouldn't run
> > > > > concurrently with other things. Not sure how common that is in the CL
> > > > > world though.
> > > > >
> > > > > >
> > > > > > -Tom
> > > > > >
> > > > > >>  tests/cl.py | 2 +-
> > > > > >>  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > > >>
> > > > > >> diff --git a/tests/cl.py b/tests/cl.py
> > > > > >> index fd698fc..b59368d 100644
> > > > > >> --- a/tests/cl.py
> > > > > >> +++ b/tests/cl.py
> > > > > >> @@ -21,7 +21,7 @@ def add_plain_test(group, name, args):
> > > > > >>  def add_concurrent_test(group, name, args):
> > > > > >>          test = PiglitTest(args)
> > > > > >>          test.run_concurrent = true;
> > > > > >> -        group[name] = PiglitTest(args)
> > > > > >> +        group[name] = test
> > > > > >>
> > > > > >>  def add_plain_program_tester_test(group, name, path):
> > > > > >>          add_plain_test(group, name, ['cl-program-tester', path])
> > > > > >> --
> > > > > >> 1.9.3
> > > > > >>
> > > > > >> _______________________________________________
> > > > > >> Piglit mailing list
> > > > > >> Piglit@lists.freedesktop.org
> > > > > >> http://lists.freedesktop.org/mailman/listinfo/piglit
> > > > > > _______________________________________________
> > > > > > Piglit mailing list
> > > > > > Piglit@lists.freedesktop.org
> > > > > > http://lists.freedesktop.org/mailman/listinfo/piglit
> > > > > _______________________________________________
> > > > > Piglit mailing list
> > > > > Piglit@lists.freedesktop.org
> > > > > http://lists.freedesktop.org/mailman/listinfo/piglit
> > > > >
> > > > _______________________________________________
> > > > Piglit mailing list
> > > > Piglit@lists.freedesktop.org
> > > > http://lists.freedesktop.org/mailman/listinfo/piglit
> > > 
> > 
> > 
> > -- 
> > Jan Vesely <jan.ves...@rutgers.edu>
> 
> 

-- 
Jan Vesely <jan.ves...@rutgers.edu>

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
Piglit mailing list
Piglit@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/piglit

Reply via email to