Commit Graph

12 Commits

Author SHA1 Message Date
mtklein
279c786409 If we swap its arguments, SkTaskGroup::batch() _is_ sk_parallel_for.
Why have two names if we can get away with one?

This kills off sk_parallel_for_thread_count(), which was only used to avoid forcing a deadlock in OncePtrTest on multicore machines in singlethreaded mode... a really niche use case.  Instead just don't explicitly force a race.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1552093002

Review URL: https://codereview.chromium.org/1552093002
2016-01-04 19:13:19 -08:00
halcanary
385fe4d4b6 Style Change: SkNEW->new; SkDELETE->delete
DOCS_PREVIEW= https://skia.org/?cl=1316123003

Review URL: https://codereview.chromium.org/1316123003
2015-08-26 13:07:49 -07:00
mtklein
00b621cfc0 Add sk_parallel_for()
This should be a drop-in replacement for most for-loops to make them run in parallel:
   for (int i = 0; i < N; i++) { code... }
   ~~~>
   sk_parallel_for(N, [&](int i) { code... });

This is just syntax sugar over SkTaskGroup to make this use case really easy to write.
There's no more overhead that we weren't already forced to add using an interface like batch(),
and no extra heap allocations.

I've replaced 3 uses of SkTaskGroup with sk_parallel_for:
  1) My unit tests for SkOnce.
  2) Cary's path fuzzer.
  3) SkMultiPictureDraw.
Performance should be the same.  Please compare left and right for readability. :)

BUG=skia:

No public API changes.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/1184373003
2015-06-17 15:26:15 -07:00
mtklein
406654be7a SkThreadPool ~~> SkTaskGroup
SkTaskGroup is like SkThreadPool except the threads stay in
one global pool.  Each SkTaskGroup itself is tiny (4 bytes)
and its wait() method applies only to tasks add()ed to that
instance, not the whole thread pool.

This means we don't need to bring up new thread pools when
tests themselves want to use multithreading (e.g. pathops,
quilt).  We just create a new SkTaskGroup and wait for that
to complete.  This should be more efficient, and allow us
to expand where we use threads to really latency sensitive
places.  E.g. we can probably now use these in nanobench
for CPU .skp rendering.

Now that all threads are sharing the same pool, I think we
can remove most of the custom mechanism pathops tests use
to control threading.  They'll just ride on the global pool
with all other tests now.

This (temporarily?) removes the GPU multithreading feature
from DM, which we don't use.

On my desktop, DM runs a little faster (57s -> 55s) in
Debug, and a lot faster in Release (36s -> 24s).  The bots
show speedups of similar proportions, cutting more than a
minute off the N4/Release and Win7/Debug runtimes.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/9c7207b5dc71dc5a96a2eb107d401133333d5b6f

R=caryclark@google.com, bsalomon@google.com, bungeman@google.com, mtklein@google.com, reed@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/531653002
2014-09-03 15:34:37 -07:00
mtklein
2460bbdfbb Revert of SkThreadPool ~~> SkTaskGroup (patchset #4 id:60001 of https://codereview.chromium.org/531653002/)
Reason for revert:
Leaks, leaks, leaks.

Original issue's description:
> SkThreadPool ~~> SkTaskGroup
>
> SkTaskGroup is like SkThreadPool except the threads stay in
> one global pool.  Each SkTaskGroup itself is tiny (4 bytes)
> and its wait() method applies only to tasks add()ed to that
> instance, not the whole thread pool.
>
> This means we don't need to bring up new thread pools when
> tests themselves want to use multithreading (e.g. pathops,
> quilt).  We just create a new SkTaskGroup and wait for that
> to complete.  This should be more efficient, and allow us
> to expand where we use threads to really latency sensitive
> places.  E.g. we can probably now use these in nanobench
> for CPU .skp rendering.
>
> Now that all threads are sharing the same pool, I think we
> can remove most of the custom mechanism pathops tests use
> to control threading.  They'll just ride on the global pool
> with all other tests now.
>
> This (temporarily?) removes the GPU multithreading feature
> from DM, which we don't use.
>
> On my desktop, DM runs a little faster (57s -> 55s) in
> Debug, and a lot faster in Release (36s -> 24s).  The bots
> show speedups of similar proportions, cutting more than a
> minute off the N4/Release and Win7/Debug runtimes.
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/9c7207b5dc71dc5a96a2eb107d401133333d5b6f

R=caryclark@google.com, bsalomon@google.com, bungeman@google.com, reed@google.com, mtklein@chromium.org
TBR=bsalomon@google.com, bungeman@google.com, caryclark@google.com, mtklein@chromium.org, reed@google.com
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Author: mtklein@google.com

Review URL: https://codereview.chromium.org/533393002
2014-09-03 14:17:48 -07:00
mtklein
9c7207b5dc SkThreadPool ~~> SkTaskGroup
SkTaskGroup is like SkThreadPool except the threads stay in
one global pool.  Each SkTaskGroup itself is tiny (4 bytes)
and its wait() method applies only to tasks add()ed to that
instance, not the whole thread pool.

This means we don't need to bring up new thread pools when
tests themselves want to use multithreading (e.g. pathops,
quilt).  We just create a new SkTaskGroup and wait for that
to complete.  This should be more efficient, and allow us
to expand where we use threads to really latency sensitive
places.  E.g. we can probably now use these in nanobench
for CPU .skp rendering.

Now that all threads are sharing the same pool, I think we
can remove most of the custom mechanism pathops tests use
to control threading.  They'll just ride on the global pool
with all other tests now.

This (temporarily?) removes the GPU multithreading feature
from DM, which we don't use.

On my desktop, DM runs a little faster (57s -> 55s) in
Debug, and a lot faster in Release (36s -> 24s).  The bots
show speedups of similar proportions, cutting more than a
minute off the N4/Release and Win7/Debug runtimes.

BUG=skia:
R=caryclark@google.com, bsalomon@google.com, bungeman@google.com, mtklein@google.com, reed@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/531653002
2014-09-03 14:06:48 -07:00
commit-bot@chromium.org
8cb1daaa1e fix minor skp-found bugs
remove globals from pathops_unittest

BUG=skia:2460
TBR=mtklein

Author: caryclark@google.com

Review URL: https://codereview.chromium.org/239563004

git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
caryclark@google.com
570863f2e2 path ops work in progress
path ops work in progress

BUG=

Review URL: https://codereview.chromium.org/21359002

git-svn-id: http://skia.googlecode.com/svn/trunk@11291 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-09-16 15:55:01 +00:00
caryclark@google.com
a5e55925ea path ops -- fix skp bugs
This fixes a series of bugs discovered by running
the small set of Skia skp files through pathops
to flatten the clips.
Review URL: https://codereview.chromium.org/14798004

git-svn-id: http://skia.googlecode.com/svn/trunk@9042 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-05-07 18:51:31 +00:00
caryclark@google.com
9515f09b6c path ops : remove countdown overkill
Review URL: https://codereview.chromium.org/13958005

git-svn-id: http://skia.googlecode.com/svn/trunk@8756 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-04-18 19:50:01 +00:00
skia.committer@gmail.com
391ca66276 Sanitizing source files in Skia_Periodic_House_Keeping
git-svn-id: http://skia.googlecode.com/svn/trunk@8608 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-04-11 07:01:45 +00:00
caryclark@google.com
66089e4ec4 Make parallel unit testing work on windows
Review URL: https://codereview.chromium.org/14072002

git-svn-id: http://skia.googlecode.com/svn/trunk@8594 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-04-10 15:55:37 +00:00