AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Add name field to calmbench json

Browse Source 
This should let perf.skia.org get the "name" keyword instead of "test".

Bug: skia:7816
Change-Id: Icb2e9f7c6109a0cfb06499ee58f9fd0cc834e5c3
Reviewed-on: https://skia-review.googlesource.com/125540
Auto-Submit: Yuqian Li <liyuqian@google.com>
Commit-Queue: Joe Gregorio <jcgregorio@google.com>
Reviewed-by: Joe Gregorio <jcgregorio@google.com>
This commit is contained in:
Yuqian Li 2018-05-02 17:26:40 -04:00 committed by Skia Commit-Bot
parent 2f62de01d3
commit 9127ea3488
1 changed files with 12 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
tools/calmbench/ab.py
View File

@ -245,6 +245,9 @@ def different_enough(lower1, upper2):
return upper2 < DIFF_T * lower1
# TODO(liyuqian): we used this hacky criteria mainly because that I didn't have
# time to study more rigorous statistical tests. We should adopt a more rigorous
# test in the future.
def get_suspects():
suspects = []
for bench in timesA.keys():
@ -347,7 +350,15 @@ def test():
args.config: {
"signed_regression": normalize_r(r),
"lower_quantile_ms": get_lower_upper(timesA[bench])[0] * 1e-6,
"upper_quantile_ms": get_lower_upper(timesA[bench])[1] * 1e-6
"upper_quantile_ms": get_lower_upper(timesA[bench])[1] * 1e-6,
"options": {
# TODO(liyuqian): let ab.py call nanobench with --outResultsFile so
# nanobench could generate the json for us that's exactly the same
# as that being used by perf bots. Currently, we cannot guarantee
# that bench is the name (e.g., bench may have additional resolution
# information appended after name).
"name": bench
}
}
}