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
diff --git a/tools/skpdiff/SkDiffContext.cpp b/tools/skpdiff/SkDiffContext.cpp
index 42d20de..78d8400 100644
--- a/tools/skpdiff/SkDiffContext.cpp
+++ b/tools/skpdiff/SkDiffContext.cpp
@@ -12,7 +12,7 @@
 #include "SkSize.h"
 #include "SkStream.h"
 #include "SkTDict.h"
-#include "SkThreadPool.h"
+#include "SkTaskGroup.h"
 
 // from the tools directory for replace_char(...)
 #include "picture_utils.h"
@@ -24,7 +24,6 @@
 SkDiffContext::SkDiffContext() {
     fDiffers = NULL;
     fDifferCount = 0;
-    fThreadCount = SkThreadPool::kThreadPerCore;
 }
 
 SkDiffContext::~SkDiffContext() {
@@ -87,7 +86,7 @@
 }
 
 static SkString get_combined_name(const SkString& a, const SkString& b) {
-    // Note (stephana): We must keep this function in sync with 
+    // Note (stephana): We must keep this function in sync with
     // getImageDiffRelativeUrl() in static/loader.js (under rebaseline_server).
     SkString result = a;
     result.append("-vs-");
@@ -238,7 +237,7 @@
         return;
     }
 
-    SkThreadPool threadPool(fThreadCount);
+    SkTaskGroup tg;
     SkTArray<SkThreadedDiff> runnableDiffs;
     runnableDiffs.reset(baselineEntries.count());
 
@@ -253,13 +252,11 @@
         if (sk_exists(testFile.c_str()) && !sk_isdir(testFile.c_str())) {
             // Queue up the comparison with the differ
             runnableDiffs[x].setup(this, baselineFile, testFile);
-            threadPool.add(&runnableDiffs[x]);
+            tg.add(&runnableDiffs[x]);
         } else {
             SkDebugf("Baseline file \"%s\" has no corresponding test file\n", baselineFile.c_str());
         }
     }
-
-    threadPool.wait();
 }
 
 
@@ -284,16 +281,14 @@
         return;
     }
 
-    SkThreadPool threadPool(fThreadCount);
+    SkTaskGroup tg;
     SkTArray<SkThreadedDiff> runnableDiffs;
     runnableDiffs.reset(baselineEntries.count());
 
     for (int x = 0; x < baselineEntries.count(); x++) {
         runnableDiffs[x].setup(this, baselineEntries[x], testEntries[x]);
-        threadPool.add(&runnableDiffs[x]);
+        tg.add(&runnableDiffs[x]);
     }
-
-    threadPool.wait();
 }
 
 void SkDiffContext::outputRecords(SkWStream& stream, bool useJSONP) {