Speed up GM ~50% by NOT running most renderModes by default

Speed up GM ~50% by NOT running most renderModes by default
BUG=skia:1566
R=borenet@google.com, bsalomon@google.com

Committed: https://code.google.com/p/skia/source/detail?r=11265

[epoger, 2013-09-13 19:00:37]

https://codereview.chromium.org/23889030/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/23889030/diff/1/gm/gmmain.cpp#newcode1370
gm/gmmain.cpp:1370: DEFINE_bool(deferred, false, "Exercise the deferred
rendering test pass.");
I have changed all of the renderModes from true to false by default.

https://codereview.chromium.org/23889030/diff/1/gm/gmmain.cpp#newcode1374
gm/gmmain.cpp:1374: DEFINE_string(gpuCacheSize, "", "<bytes> <count>: Limit the
gpu cache to byte size or "
Here's what I ran to make sure that the GM results still match what they did
before, and to measure the speedup.

# Before making any changes, wrote out expectations:
EXPECTATIONS_DIR=/tmp/gm-expectations
rm -rf $EXPECTATIONS_DIR
mkdir -p $EXPECTATIONS_DIR
out/Debug/gm --match x --writePath $EXPECTATIONS_DIR

# Ran this before and after making the code change:
START=$(date +%s) && \
  out/Debug/gm --match x --readPath $EXPECTATIONS_DIR && \
  END=$(date +%s) && echo $(($END - $START)) seconds

Before the change:
GM: Ran 450 tests: NoGpuContext=0 IntentionallySkipped=2 RenderModeMismatch=0
ExpectationsMismatch=0 MissingExpectations=0 WritingReferenceImage=0
GM: [*] 0 NoGpuContext:
GM: [ ] 2 IntentionallySkipped
GM: [*] 0 RenderModeMismatch:
GM: [*] 0 ExpectationsMismatch:
GM: [ ] 0 MissingExpectations
GM: [*] 0 WritingReferenceImage:
GM: (results marked with [*] will cause nonzero return value)
71 seconds

After the change:
GM: Ran 135 tests: NoGpuContext=0 IntentionallySkipped=2 RenderModeMismatch=0
ExpectationsMismatch=0 MissingExpectations=0 WritingReferenceImage=0
GM: [*] 0 NoGpuContext:
GM: [ ] 2 IntentionallySkipped
GM: [*] 0 RenderModeMismatch:
GM: [*] 0 ExpectationsMismatch:
GM: [ ] 0 MissingExpectations
GM: [*] 0 WritingReferenceImage:
GM: (results marked with [*] will cause nonzero return value)
37 seconds

https://codereview.chromium.org/23889030/diff/1/gm/tests/outputs/add-config-p...
File gm/tests/outputs/add-config-pdf/output-expected/stdout (right):

https://codereview.chromium.org/23889030/diff/1/gm/tests/outputs/add-config-p...
gm/tests/outputs/add-config-pdf/output-expected/stdout:10: GM: ...  and  0 modes
  []
Note the smaller set of operations that are being exercised now.

[epoger, 2013-09-13 19:01:40]

Er, I guess this is a ~100% speedup.  Anyway, it takes half as long as it used
to.

[borenet, 2013-09-13 19:04:54]

Code change LGTM.  CC Rob and Mike to make sure we really don't care about these
modes.

https://codereview.chromium.org/23889030/diff/1/gm/tests/outputs/add-config-p...
File gm/tests/outputs/add-config-pdf/output-expected/stdout (right):

https://codereview.chromium.org/23889030/diff/1/gm/tests/outputs/add-config-p...
gm/tests/outputs/add-config-pdf/output-expected/stdout:9: GM: ... over  3
configs ["8888", "565", "pdf"]
Do we care about 565?

[epoger, 2013-09-13 19:14:58]

On 2013/09/13 19:04:54, borenet wrote:
> Do we care about 565?

FYI, 565 doesn't seem to take very long.  I tried taking that out too, and it
went down from 37 sec to 32 sec.

GM: Ran 90 tests: NoGpuContext=0 IntentionallySkipped=1 RenderModeMismatch=0
ExpectationsMismatch=0 MissingExpectations=0 WritingReferenceImage=0
GM: [*] 0 NoGpuContext:
GM: [ ] 1 IntentionallySkipped
GM: [*] 0 RenderModeMismatch:
GM: [*] 0 ExpectationsMismatch:
GM: [ ] 0 MissingExpectations
GM: [*] 0 WritingReferenceImage:
GM: (results marked with [*] will cause nonzero return value)
32 seconds

[bsalomon, 2013-09-13 19:53:10]

lgtm

[epoger, 2013-09-13 19:56:24]

Committed patchset #1 manually as r11265 (presubmit successful).