[Piglit] [PATCH] Mark timer tests as not concurrent

Nicolai Hähnle nhaehnle at gmail.com
Thu Sep 15 16:11:11 UTC 2016

On 13.09.2016 15:38, Emil Velikov wrote:
> On 12 September 2016 at 16:31, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
>> From: Nicolai Hähnle <nicolai.haehnle at amd.com>
>> This reduces the chance of false positives caused by system load.
> From my experience the {arb,ext}-timer-query tests have always been
> inconsistent. Even on a consecutive, non-concurrent, runs outside
> piglit. System load no affect here, afaict.
> What is your observation before/after the patch, please add some
> numbers to the commit message.

Going to add this to the commit message, but posting here as well for 
the record:

On an 8-core test machine with an AMD Polaris, I've observed about a 50% 
false positive rate on these tests when running concurrently. On the 
same system, I have not observed a single failure in 10 runs when 
running non-concurrently.

> As it helps at least some setups:
> Acked-by: Emil Velikov <emil.velikov at collabora.com>



> Thanks
> Emil

More information about the Piglit mailing list