diff options
author | Anthony Barbier <anthony.barbier@arm.com> | 2017-11-15 13:06:46 +0000 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:35:24 +0000 |
commit | 2f8e077378a757128540428b0a1318f80b6dea75 (patch) | |
tree | e71f94781ad384d8748c57a30f2ca16bed2cf8fe /tests/framework/Framework.cpp | |
parent | 4c2dd54d6983275530ef20f9dbb4ce6080c7307b (diff) | |
download | ComputeLibrary-2f8e077378a757128540428b0a1318f80b6dea75.tar.gz |
COMPMID-556: Don't measure performance for the first run
This messes up badly the opencl_timer if we measure the first run and are using a CLTuner at the same time.
Basically all the tuner combinations appear in the JSON file and this makes it really hard to post process
Change-Id: I59377eb752ca8ddc95b884ab04b1ef3f65389931
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/95860
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Diffstat (limited to 'tests/framework/Framework.cpp')
-rw-r--r-- | tests/framework/Framework.cpp | 10 |
1 files changed, 8 insertions, 2 deletions
diff --git a/tests/framework/Framework.cpp b/tests/framework/Framework.cpp index 39fe1fa00a..38c65fed85 100644 --- a/tests/framework/Framework.cpp +++ b/tests/framework/Framework.cpp @@ -289,7 +289,10 @@ void Framework::run_test(const TestInfo &info, TestCaseFactory &test_factory) for(int i = 0; i < _num_iterations; ++i) { - profiler.start(); + if(_num_iterations > 1 && i != 0) + { + profiler.start(); + } test_case->do_run(); #ifdef ARM_COMPUTE_CL if(opencl_is_available()) @@ -303,7 +306,10 @@ void Framework::run_test(const TestInfo &info, TestCaseFactory &test_factory) GCScheduler::get().sync(); } #endif /* ARM_COMPUTE_GC */ - profiler.stop(); + if(_num_iterations > 1 && i != 0) + { + profiler.stop(); + } } test_case->do_teardown(); |