diff options
| author | Frank Barchard <fbarchard@google.com> | 2017-09-22 11:54:55 -0700 |
|---|---|---|
| committer | Frank Barchard <fbarchard@google.com> | 2017-09-22 18:58:11 +0000 |
| commit | 0b0a891cb29384bc8cd74df3019ff1b8d41d3e6a (patch) | |
| tree | 5e444a0d5d19011a7e6e3096a8b5c9531e16de22 /BUILD.gn | |
| parent | efbf15754a1efb8e378b1873eb213d5c8931a3ba (diff) | |
| download | libyuv-0b0a891cb29384bc8cd74df3019ff1b8d41d3e6a.tar.gz | |
Change TestScaleSumSamples_C test to allow for some float error in sum.
The sum of floats can optimize differently with vectorization, producing
a different result between NEON and C.
Adjust the unittest to allow for some difference in the sum.
The NEON version is 8 samples at a time, so the test now rounds up
the number of values to multiple of 8.
TBR=kjellander@chromium.org
Bug: libyuv:717
Test: LibYUVPlanarTest.TestScaleSumSamples_Opt
Change-Id: I2a0783780c7e0f240f7a8e4700b2a4d3e6b52d87
Reviewed-on: https://chromium-review.googlesource.com/673708
Reviewed-by: Cheng Wang <wangcheng@google.com>
Diffstat (limited to 'BUILD.gn')
| -rw-r--r-- | BUILD.gn | 3 |
1 files changed, 2 insertions, 1 deletions
@@ -160,7 +160,7 @@ static_library("libyuv_internal") { # To enable AVX2 or other cpu optimization, pass flag here # cflags = [ "-mavx2", "-mpopcnt", "-mavx2", "-mfma" ] if (!is_win) { - cflags = [ "-ffp-contract=fast" ] # Enable fma vectorization for NEON. + cflags = [ "-ffp-contract=fast" ] # Enable fma vectorization for NEON. } } if (libyuv_use_neon) { @@ -185,6 +185,7 @@ if (libyuv_use_neon) { configs -= [ "//build/config/compiler:default_optimization" ] # Enable optimize for speed (-O2) over size (-Os). + # TODO(fbarchard): Consider optimize_speed which is O3. configs += [ "//build/config/compiler:optimize_max" ] } |