Go to file
Mike Klein 4f6e271596 ok, add a bench source
This new source acts like other sources (GMs, SKPs) for benchmarks.  It
times multiple samples (controlled by samples=N, default 20), and each
of those samples uses the same strategy as monobench, growing loops
exponentially until it runs for at least 10ms.

When done it prints the fastest and the two slowest samples.  In
practice the 100th percentile sample is very different from the
next slowest due to caching, and the fastest is always interesting.

Because these benchmarks run in whatever execution engine ok has
selected, on non-Windows platforms you have some real control over the
interaction between benchmarks.  In its default "fork" mode each
benchmark runs independently in its own process, so the 100th
percentiles really stand out.  The other modes "thread" and "serial"
work as you'd expect too.

Here's an example where you can see how the different interactions work:

out/ok bench:samples=100 8888 filter:search=text_16_AA fork
    [text_16_AA_WT] 2.32µs  @0  6.23µs  @99 24.3ms  @100
    [text_16_AA_FF] 2.41µs  @0  5.7µs   @99 23.3ms  @100
    [text_16_AA_88] 2.55µs  @0  5.6µs   @99 24.8ms  @100
    [text_16_AA_BK] 1.97µs  @0  5.44µs  @99 23.2ms  @100

out/ok bench:samples=100 8888 filter:search=text_16_AA thread
    [text_16_AA_FF] 2.45µs  @0  23.5µs  @99 24.8ms  @100
    [text_16_AA_WT] 2.52µs  @0  17.8µs  @99 24.7ms  @100
    [text_16_AA_88] 2.55µs  @0  19.7µs  @99 25.1ms  @100
    [text_16_AA_BK] 1.8µs   @0  14.7µs  @99 25.1ms  @100

out/ok bench:samples=100 8888 filter:search=text_16_AA serial
    [text_16_AA_88] 2.35µs  @0  3.53µs  @99 16.7ms  @100
    [text_16_AA_FF] 2.09µs  @0  2.73µs  @99 2.91µs  @100
    [text_16_AA_BK] 1.75µs  @0  2.46µs  @99 2.65µs  @100
    [text_16_AA_WT] 2.1µs   @0  3.16µs  @99 3.17µs  @100

In the first "fork" case all runs are independent and have roughly
the same profile.  "thread" looks similar except you can see them
contending at the 99th percentile.  In "serial", the first bench
warms up the rest, so their 100th percentiles are all much faster.

Change-Id: I01a9f8c54b540221a9f232b271bb8ef3fda2569c
Reviewed-on: https://skia-review.googlesource.com/33585
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
2017-08-11 15:32:36 +00:00
animations first cut at a checkbox 2009-10-21 19:41:10 +00:00
bench Abort dm/nanobench on bad configs. 2017-08-11 13:12:26 +00:00
bin add a Win/Clang build bot 2017-07-31 19:53:51 +00:00
debugger Add Make[backend] calls for creating GrContexts 2017-07-25 14:33:03 +00:00
dm Abort dm/nanobench on bad configs. 2017-08-11 13:12:26 +00:00
docs fix links and missing constructor 2017-08-04 17:25:24 +00:00
example Add Make[backend] calls for creating GrContexts 2017-07-25 14:33:03 +00:00
experimental Add Make[backend] calls for creating GrContexts 2017-07-25 14:33:03 +00:00
fuzz Remove SkTypeface::Style from fuzzers and lua. 2017-08-08 15:27:19 +00:00
gm Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. 2017-08-11 14:13:26 +00:00
gn Build support for 32-bit armv8-a 2017-08-10 17:31:25 +00:00
include Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. 2017-08-11 14:13:26 +00:00
infra Abort dm/nanobench on bad configs. 2017-08-11 13:12:26 +00:00
platform_tools Enable ios on Raspberry Pi 2017-04-25 16:56:41 +00:00
resources Remove SkTypeface::Style from fuzzers and lua. 2017-08-08 15:27:19 +00:00
samplecode Remove SkTypeface::Style from fuzzers and lua. 2017-08-08 15:27:19 +00:00
site fix Vulkan doc typo 2017-08-07 16:53:19 +00:00
src Remove subpixel positioning implies no bytecode hinting. 2017-08-11 14:21:27 +00:00
tests Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. 2017-08-11 14:13:26 +00:00
third_party link libwebpmux in system-webp builds 2017-08-01 14:33:38 +00:00
tools ok, add a bench source 2017-08-11 15:32:36 +00:00
.clang-format Mark flatennable macros as block beginning/ending in .clang-format 2017-01-09 15:31:36 +00:00
.gitignore clang on windows support 2017-07-31 18:39:23 +00:00
.gn Basic standalone GN configs. 2016-07-21 12:25:45 -07:00
AUTHORS Added support for building for tvOS 2017-03-14 22:55:04 +00:00
BUILD.gn ok, add a bench source 2017-08-11 15:32:36 +00:00
codereview.settings Make uploading to Gerrit the default for Skia 2016-11-09 19:07:56 +00:00
CONTRIBUTING Fix references to https://sites.google.com/site/skiadocs/. 2015-02-03 13:12:54 -02:00
CQ_COMMITTERS Moved committer list to chrome-infra-auth and deleted it from the repo 2015-09-02 13:37:54 -07:00
DEPS Roll skia/third_party/externals/angle2/ d23bcd8eb..c2157a09d (1 commit) 2017-08-11 08:55:56 +00:00
Doxyfile Make the housekeeper upload doxygen to a newer bucket 2016-10-04 13:23:57 -07:00
LICENSE BUG=skia:5602 2016-09-02 11:19:34 -07:00
PRESUBMIT.py Update CQ extra trybots after switch to Debian 2017-06-29 19:35:40 +00:00
public.bzl Fix DM_ARGS in public.bzl. 2017-08-11 13:51:56 +00:00
README Fix references to https://sites.google.com/site/skiadocs/. 2015-02-03 13:12:54 -02:00
README.chromium Update README.chromium. 2015-06-11 13:19:24 -07:00
whitespace.txt Revert "Revert "Make GrAtlasTextOp a non-legacy GrMeshDrawOp"" 2017-07-19 12:17:34 +00:00

Skia is a complete 2D graphic library for drawing Text, Geometries, and Images.

See full details, and build instructions, at https://skia.org.