SkThreadPool ~~> SkTaskGroup

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

Committed: https://skia.googlesource.com/skia/+/406654be7a930b484159f5bca107d3b11d8a9ede

[mtklein_C, 2014-09-02 08:24:06]

mtklein@chromium.org changed reviewers:
+ bungeman@google.com, caryclark@google.com

[mtklein, 2014-09-03 19:00:17]

mtklein@google.com changed reviewers:
+ bsalomon@google.com, mtklein@google.com

[mtklein, 2014-09-03 19:06:28]

mtklein@google.com changed reviewers:
+ reed@google.com

[reed1, 2014-09-03 19:16:22]

api lgtm

https://codereview.chromium.org/531653002/diff/40001/src/utils/SkRunnable.h
File src/utils/SkRunnable.h (right):

https://codereview.chromium.org/531653002/diff/40001/src/utils/SkRunnable.h#n...
src/utils/SkRunnable.h:11: struct SkRunnable {
SkLambda ? :)

[bsalomon, 2014-09-03 19:21:06]

On 2014/09/03 19:06:29, mtklein wrote:

On windows
time dm --nogms

before: 3m58.878s
after: 1m41.974s

[caryclark, 2014-09-03 19:38:40]

lgtm

[mtklein, 2014-09-03 20:54:02]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-09-03 20:55:00]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/531653002/60001

[mtklein, 2014-09-03 20:57:25]

Jim, FYI, there's a very good chance this CL could cause some bot chaos.  It
changes how DM runs pretty significantly (for the better of course).

Most likely things to break: TSAN bot, Valgrind bot.  I think I've got
intentional memory leaks suppressed for Valgrind, but might have messed that up
and need to iterate.  If TSAN fails, it's probably just uncovering new existing
problems, not something wrong with this CL.  I will investigate those.

If this causes anything else to fail, we should revert.

[commit-bot: I haz the power, 2014-09-03 21:06:54]

Committed patchset #4 (id:60001) as 9c7207b5dc71dc5a96a2eb107d401133333d5b6f

[mtklein, 2014-09-03 21:16:19]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/533393002/ by mtklein@google.com.

The reason for reverting is: Leaks, leaks, leaks..

[mtklein, 2014-09-03 21:57:31]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-09-03 21:58:34]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/531653002/80001

[commit-bot: I haz the power, 2014-09-03 22:34:43]

Committed patchset #5 (id:80001) as 406654be7a930b484159f5bca107d3b11d8a9ede

[jvanverth1, 2014-09-04 13:25:45]

A revert of this CL (patchset #5 id:80001) has been created in
https://codereview.chromium.org/540013002/ by jvanverth@google.com.

The reason for reverting is: It's blocking the DEPS roll.  Error in
linux_chromium_gn_rel is:

ERROR at //skia/BUILD.gn:330:3: Item not found
  "//third_party/skia/src/utils/SkThreadPool.h",
  ^--------------------------------------------
You were trying to remove "//third_party/skia/src/utils/SkThreadPool.h"
from the list but it wasn't there..