Adds BENCH_ALLOWED_NOISE to widen expected bench ranges.
This is especially useful for benches with small absolute value.
Review URL: https://codereview.appspot.com/6842112
git-svn-id: http://skia.googlecode.com/svn/trunk@6633 2bbb7eff-a529-9590-31e7-b0007b416f81
diff --git a/bench/gen_skp_ranges.py b/bench/gen_skp_ranges.py
index e6c4295..575db37 100755
--- a/bench/gen_skp_ranges.py
+++ b/bench/gen_skp_ranges.py
@@ -35,6 +35,11 @@
BENCH_UB = 1.1 # Allow for 10% room for normal variance on the up side.
BENCH_LB = 0.85
+# Further allow for a fixed amount of noise. This is especially useful for
+# benches of smaller absolute value. Keeping this value small will not affect
+# performance tunings.
+BENCH_ALLOWED_NOISE = 10
+
# List of platforms to track.
PLATFORMS = ['Mac_Float_Bench_32',
'Nexus10_4-1_Float_Bench_32',
@@ -56,8 +61,8 @@
combination that we care most about. Platforms that did not run bench_pictures
in the given revision range will not have corresponding outputs.
Please check http://go/skpbench to choose a range that fits your needs.
-BENCH_UB and BENCH_LB can be changed to expand or narrow the permitted bench
-ranges without triggering buidbot alerts.
+BENCH_UB, BENCH_LB and BENCH_ALLOWED_NOISE can be changed to expand or narrow
+the permitted bench ranges without triggering buidbot alerts.
"""
HELP_STRING = """
Outputs expectation picture bench ranges for the latest revisions for the given
@@ -123,8 +128,9 @@
for key in keys:
bench_val = expectation_dic[key]
# Prints out expectation lines.
- print '%s,%.3f,%.3f,%.3f' % (key, bench_val, bench_val * BENCH_LB,
- bench_val * BENCH_UB)
+ print '%s,%.3f,%.3f,%.3f' % (key, bench_val,
+ bench_val * BENCH_LB - BENCH_ALLOWED_NOISE,
+ bench_val * BENCH_UB + BENCH_ALLOWED_NOISE)
def main():
"""Parses flags and outputs expected Skia picture bench results."""