Commit Graph

1148 Commits

Author SHA1 Message Date
msarett
ab80e35fbd Improved subset benchmarks
I think these changes to the subset benchmarks cover what we discussed yesterday.

I removed the divisor benchmarks (2x2, 3x3) and changed the single subset benchmarks.

Also, we will no longer benchmark subset decodes on small images.

BUG=skia:

Review URL: https://codereview.chromium.org/1188223002
2015-06-17 10:28:22 -07:00
mtklein
bf9e600069 nanobench: split CPU and GPU .SKP tile sizes.
Let's make CPU-bound .SKP benching mimic Chrome's tiles.
Unfortunately, the CPU code also performs a lot better with those big wide tiles...

BUG=skia:

Review URL: https://codereview.chromium.org/1189863002
2015-06-16 10:41:27 -07:00
mtklein
c751ecb368 When --mpd is true, run _mpd variants first, then non-mpd.
This makes it easier to benchmark _mpd variants in a profiler.

E.g.,

<profiler> out/Release/nanobench --images --config 8888 --loops -1 --match sp_desk_nytimes

BUG=skia:

Review URL: https://codereview.chromium.org/1184673006
2015-06-15 08:56:38 -07:00
mtklein
f2fe0e0320 Remove overly-promiscuous SkNx syntax sugar.
I haven't figured out a pithy way to have these apply to only classes
originating from SkNx, so let's just remove them.  There aren't too
many use cases, and it's not really any less readable without them.

Semantically, this is a no-op.

BUG=skia:

Review URL: https://codereview.chromium.org/1167153002
2015-06-10 08:57:28 -07:00
msarett
b23e6aa676 Subset decoding benchmarks
It was my goal to create benchmarks that could measure all
of the use cases that we have identified.  I think single
subsets, translating, and scaling are the important ones.

It might be a good idea to discuss the document in greater
detail as well.  I just wanted to share this to aid the
discussion.
https://docs.google.com/a/google.com/document/d/1OxW96GDMAlw6dnzNXmiNX-F9oDBBlGXzSsgd0DMIkbI/edit?usp=sharing

BUG=skia:

Review URL: https://codereview.chromium.org/1160953002
2015-06-09 13:56:10 -07:00
reed
41e010cb90 Revert[2] SkDraw and all Blitters to use pixmap instead of bitmapi
This reverts commit b3f0ec9f99.

BUG=skia:

Review URL: https://codereview.chromium.org/1168303006
2015-06-09 12:16:53 -07:00
mtklein
1831f990c3 A bunch of little SkMScalar-as-float build fixes.
Landing this before http://crrev.com/1169813006 so the test changes needed
there are clearer.  This just lets things build with floats, and is a no-op for
doubles.

No public API changes we didn't intend to already have...
TBR=reed@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/1173673003
2015-06-09 11:47:01 -07:00
mtklein
257918898a Pump up matrix44_setconcat benches 1000x so they can be timed on Android.
BUG=skia:

Review URL: https://codereview.chromium.org/1174583002
2015-06-09 09:29:12 -07:00
mtklein
c6c6a91ff2 Split SetConcatMatrix44Bench to cover both internal paths.
The existing bench only tests the fast path, but we're looking to speed
up the general case.  It'd be nice to be able to measure that speedup.

BUG=skia:

Review URL: https://codereview.chromium.org/1146953003
2015-06-09 04:44:07 -07:00
reed
b3f0ec9f99 Revert of change SkDraw and all Blitters to use pixmap instead of bitmap (patchset #6 id:100001 of https://codereview.chromium.org/1148793007/)
Reason for revert:
speculative revert to try to unblock DEPS roll

Original issue's description:
> change SkDraw and all Blitters to use pixmap instead of bitmap
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/c31af44336f5eb4a50e83e76e51962d46c3ed458

TBR=scroggo@google.com,jvanverth@google.com,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1164373003
2015-06-08 19:58:07 -07:00
reed
c31af44336 change SkDraw and all Blitters to use pixmap instead of bitmap
BUG=skia:

Review URL: https://codereview.chromium.org/1148793007
2015-06-08 10:47:13 -07:00
caryclark
c58e5326bb add res to stroke bench
R=reed@google.com

Review URL: https://codereview.chromium.org/1158183003
2015-06-01 06:30:06 -07:00
bsalomon
7622863095 Add direct getter for GrCaps to GrContext.
TBR=joshualitt@google.com

Committed: https://skia.googlesource.com/skia/+/9138c46e572085870638b6f7ad7fcdfcdf3cac99

Review URL: https://codereview.chromium.org/1149773005
2015-05-29 08:02:10 -07:00
reed
ff2f423d90 add bench for building strokes
BUG=skia:

Review URL: https://codereview.chromium.org/1160093004
2015-05-29 07:17:16 -07:00
bsalomon
084d1b61bc Revert of Add direct getter for GrCaps to GrContext. (patchset #4 id:60001 of https://codereview.chromium.org/1149773005/)
Reason for revert:
Breaking

Original issue's description:
> Add direct getter for GrCaps to GrContext.
>
> TBR=joshualitt@google.com
>
> Committed: https://skia.googlesource.com/skia/+/9138c46e572085870638b6f7ad7fcdfcdf3cac99

TBR=joshualitt@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1164443002
2015-05-28 08:20:58 -07:00
bsalomon
9138c46e57 Add direct getter for GrCaps to GrContext.
TBR=joshualitt@google.com

Review URL: https://codereview.chromium.org/1149773005
2015-05-28 08:13:16 -07:00
bsalomon
682c269a15 Store context options on caps.
Committed: https://skia.googlesource.com/skia/+/f28cff71db2cbb1ff18a8fbf1e80ca761d1f69bc

Review URL: https://codereview.chromium.org/1158433006
2015-05-22 14:01:46 -07:00
bsalomon
08bf86d1b7 Revert of Store context options on caps. (patchset #3 id:40001 of https://codereview.chromium.org/1158433006/)
Reason for revert:
breaking chromeos build (???)

Original issue's description:
> Store context options on caps.
>
> Committed: https://skia.googlesource.com/skia/+/f28cff71db2cbb1ff18a8fbf1e80ca761d1f69bc

TBR=joshualitt@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1151603005
2015-05-22 12:41:05 -07:00
bsalomon
f28cff71db Store context options on caps.
Review URL: https://codereview.chromium.org/1158433006
2015-05-22 12:25:41 -07:00
kkinnunen
b4a797f3aa Move font loading in gm tests and benches out of constructors
Constructing the gm tests and benches causes many calls to font loads.
This is visible as profiling samples in fontconfig and freetype on Linux
for all profiling runs of nanobench. This complicates analysis of
test-cases that are suspected of being slow due to font-related issues.

Move the font loading to GM::onOnceBeforeDraw and Benchmark::onPreDraw.
This way the code is not executed if the testcase does not match the
nanobench --match filter. This way the samples in font-related code are
more easy to identify as legitimate occurances caused by the testcase.

This should not cause differences in timings, because:
* Benchmark::preDraw / onPreDraw is defined to be run outside the timer
* GM::runAsBench is not enabled for any of the modified testcases. Also
  nanobench untimed warmup round should run the onOnceBeforeDraw.
  (and there are other GM::runAsBench gms already doing loading in
   onOnceBeforeDraw).

Changes the behavior:
In TextBench:
Before, the test would report two different gms with the same name if
the color emoji font was not loaded successfully.
After, the test always reports all tests as individual names.

Generally:
The errors from loading fonts now print inbetween each testcase, as
opposed to printing during construction phase. Sample output:
( 143/145 MB  1872) 14.7ms	8888 gm  quadclosepathResource /fonts/Funkster.ttf not a valid font.
( 160/160 MB  1831) 575µs	8888 gm  surfacenewResource /fonts/Funkster.ttf not a valid font.
( 163/165 MB  1816) 12.5ms	8888 gm  linepathResource /fonts/Funkster.ttf not a valid font.
( 263/411 MB  1493) 118ms	8888 gm  typefacestyles_kerningResource /fonts/Funkster.ttf not a valid font.
( 374/411 MB  1231) 7.16ms	565 gm  getpostextpathResource /fonts/Funkster.ttf not a valid font.
( 323/411 MB  1179) 4.92ms	565 gm  stringartResource /fonts/Funkster.ttf not a valid font.
( 347/493 MB   917) 191ms	565 gm  patch_gridResource /fonts/Funkster.ttf not a valid font.
( 375/493 MB   857) 23.9ms	gpu gm  clipdrawdrawCannot render path (0)
( 393/493 MB   706) 2.91ms	unit test  ParsePath------ png error IEND: CRC error
( 394/493 MB   584) 166ms	gpu gm  hairmodesResource /fonts/Funkster.ttf not a valid font.
Resource /fonts/Funkster.ttf not a valid font.
Resource /fonts/Funkster.ttf not a valid font.
...

Review URL: https://codereview.chromium.org/1144023002
2015-05-21 06:15:28 -07:00
kkinnunen
54b8511189 Make GrResourceCache perf less sensitive to key length change
Make GrResourceCache performance less sensitive to key length change.
The memcmp in GrResourceKey is called when SkTDynamicHash jumps the
slots to find the hash by a index. Avoid most of the memcmps by
comparing the hash first.

This is important because small changes in key data length can cause
big performance regressions. The theory is that key length change causes
different hash values. These hash values might trigger memcmps that
originally weren't there, causing the regression.

Adds few specialized benches to grresourcecache_add to test different
key lengths. The tests are run only on release, because on debug the
SkTDynamicHash validation takes too long, and adding many such delays
to development test runs would be unproductive. On release the tests
are quite fast.

Effect of this patch to the added tests on amd64:
grresourcecache_find_10  738us ->  768us        1.04x
 grresourcecache_find_2  472us ->  476us        1.01x
grresourcecache_find_25  841us ->  845us        1x
 grresourcecache_find_4  565us ->  531us        0.94x
grresourcecache_find_54 1.18ms ->  1.1ms        0.93x
 grresourcecache_find_5  834us ->  749us        0.9x
 grresourcecache_find_3  620us ->  542us        0.87x
 grresourcecache_add_25 2.74ms -> 2.24ms        0.82x
 grresourcecache_add_56 3.23ms -> 2.56ms        0.79x
 grresourcecache_add_54 3.34ms -> 2.62ms        0.78x
  grresourcecache_add_5 2.68ms ->  2.1ms        0.78x
 grresourcecache_add_10  2.7ms -> 2.11ms        0.78x
  grresourcecache_add_2 1.85ms -> 1.41ms        0.76x
    grresourcecache_add 1.84ms ->  1.4ms        0.76x
  grresourcecache_add_4 1.99ms -> 1.49ms        0.75x
  grresourcecache_add_3 2.11ms -> 1.55ms        0.73x
 grresourcecache_add_55   39ms -> 13.9ms        0.36x
grresourcecache_find_55 23.2ms -> 6.21ms        0.27x

On arm64 the results are similar.

On arm_v7_neon, the results lack the discontinuity at 55:
    grresourcecache_add 4.06ms -> 4.26ms        1.05x
  grresourcecache_add_2 4.05ms -> 4.23ms        1.05x
   grresourcecache_find 1.28ms ->  1.3ms        1.02x
grresourcecache_find_56 3.35ms -> 3.32ms        0.99x
 grresourcecache_find_2 1.31ms -> 1.29ms        0.99x
grresourcecache_find_54 3.28ms -> 3.24ms        0.99x
  grresourcecache_add_5 6.38ms -> 6.26ms        0.98x
 grresourcecache_add_55 8.44ms -> 8.24ms        0.98x
 grresourcecache_add_25 7.03ms -> 6.86ms        0.98x
grresourcecache_find_25  2.7ms -> 2.59ms        0.96x
 grresourcecache_find_4 1.45ms -> 1.38ms        0.95x
grresourcecache_find_10 2.52ms -> 2.39ms        0.95x
grresourcecache_find_55 3.54ms -> 3.33ms        0.94x
 grresourcecache_find_5  2.5ms -> 2.32ms        0.93x
 grresourcecache_find_3 1.57ms -> 1.43ms        0.91x

The extremely slow case, 55, is postulated to be due to the index jump
collisions running the memcmp. This is not visible on arm_v7_neon probably due
to hash function producing different results for 32 bit architectures.

This change is needed for extending path cache key in Gr
NV_path_rendering codepath. Extending is needed in order to add dashed
paths to the path cache.

Review URL: https://codereview.chromium.org/1132723003
2015-05-18 22:47:33 -07:00
mtklein
d0a1088bf7 Add _aa variants of Xfermode benches.
Also, allow multiple DEF_BENCH() per line by using __COUNTER__ instead of __LINE__.

BUG=skia:

Review URL: https://codereview.chromium.org/1140643004
2015-05-13 11:54:00 -07:00
reed
80ea19ca4b Revert of Revert of stop calling SkScalarDiv (patchset #1 id:1 of https://codereview.chromium.org/1138263002/)
Reason for revert:
android patched, blink has rolled

Original issue's description:
> Revert of stop calling SkScalarDiv (patchset #4 id:60001 of https://codereview.chromium.org/1135053002/)
>
> Reason for revert:
> need to wait for Blink roll (and patch android)
>
> Original issue's description:
> > stop calling SkScalarDiv
> >
> > BUG=skia:
> > TBR=
> >
> > Committed: https://skia.googlesource.com/skia/+/67d71c898249a7af3523b16c6a69895a63bfae0a
>
> TBR=
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/2629697933b5cc975e45d2a45c48f803fc6cbcec

TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1135693003
2015-05-12 10:37:34 -07:00
reed
2629697933 Revert of stop calling SkScalarDiv (patchset #4 id:60001 of https://codereview.chromium.org/1135053002/)
Reason for revert:
need to wait for Blink roll (and patch android)

Original issue's description:
> stop calling SkScalarDiv
>
> BUG=skia:
> TBR=
>
> Committed: https://skia.googlesource.com/skia/+/67d71c898249a7af3523b16c6a69895a63bfae0a

TBR=
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1138263002
2015-05-12 10:28:06 -07:00
reed
67d71c8982 stop calling SkScalarDiv
BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1135053002
2015-05-12 09:47:22 -07:00
mtklein
44d43d8d6e Add a control benchmark.
I'm thinking of using this in perf with something like:
   ratio(fill(filter("test=foo")), fill(filter("test=control")))

Does that make sense to you?

Not sure that this is really a good control bench on all bots,
but I propose we just run it a bit and find out if it needs work.

BUG=skia:

Review URL: https://codereview.chromium.org/1129823003
2015-05-06 12:42:04 -07:00
reed
71a6cbfc58 remove redundant/deprecated TwoPointRadial gradiet -- use TwoPointConical
This CL derived from https://codereview.chromium.org/1114243005/

BUG=skia:

Review URL: https://codereview.chromium.org/1117423003
2015-05-04 08:32:51 -07:00
bungeman
3ffa126066 Move resource fonts to common location.
Move resource fonts to resources/fonts, add a destortable font for
testing, and clean up how the tests create fonts from resources.

R=joshualitt@google.com

Review URL: https://codereview.chromium.org/1120823002
2015-04-30 17:12:58 -04:00
mtklein
d75c466ef5 DM+nanobench: print both current and max RSS.
out/Debug/dm:
    ( 360/380 MB  2112) 396ms   gpu gm  xfermodes

out/Release/nanobench:
    32/33  MB 1659    2.46µs  2.55µs  2.54µs  2.67µs  2%  ▄▄▄▅▅▃▂▁▂█  gpu gradient_conicalZero_clamp_hicolor

out/Debug/nanobench:
    42/42  MB   desk_css3gradients.skp_1    8888

BUG=skia:

NOTREECHECKS=true

Review URL: https://codereview.chromium.org/1115203002
2015-04-30 07:11:22 -07:00
joshualitt
261c3ad7fd Change to add zoom animations to nanobench
BUG=skia:

Review URL: https://codereview.chromium.org/1061323003
2015-04-27 09:16:57 -07:00
msarett
9d9725c892 Do not crash nanobench in debug mode
Instead print an error message

BUG=skia:

Review URL: https://codereview.chromium.org/1102083002
2015-04-24 11:41:55 -07:00
reed
ca1a22ee53 Revert of Revert of remove unused (by clients) SkPathUtils (patchset #1 id:1 of https://codereview.chromium.org/1060703003/)
Reason for revert:
fix (removal from gypi/gn files) has landed in chrome.

Original issue's description:
> Revert of remove unused (by clients) SkPathUtils (patchset #1 id:1 of https://codereview.chromium.org/1088383003/)
>
> Reason for revert:
> This change is causing the DEPS roll to fail:
>
>
> http://build.chromium.org/p/tryserver.chromium.linux/builders/android_chromium_gn_compile_rel/builds/78771/steps/gn/logs/stdio
>
> Original issue's description:
> > remove unused (by clients) SkPathUtils
> >
> > BUG=skia:
> >
> > Committed: https://skia.googlesource.com/skia/+/aab35d91b8b80acd1902594bbf542083fdfa4bb7
>
> TBR=scroggo@google.com,reed@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/bdb0bf5f8858043878d8a4fa8130c6c87bef3fd4

TBR=scroggo@google.com,jcgregorio@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1091963002
2015-04-16 20:10:55 -07:00
jcgregorio
bdb0bf5f88 Revert of remove unused (by clients) SkPathUtils (patchset #1 id:1 of https://codereview.chromium.org/1088383003/)
Reason for revert:
This change is causing the DEPS roll to fail:

http://build.chromium.org/p/tryserver.chromium.linux/builders/android_chromium_gn_compile_rel/builds/78771/steps/gn/logs/stdio

Original issue's description:
> remove unused (by clients) SkPathUtils
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/aab35d91b8b80acd1902594bbf542083fdfa4bb7

TBR=scroggo@google.com,reed@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

Review URL: https://codereview.chromium.org/1060703003
2015-04-16 09:47:36 -07:00
reed
aab35d91b8 remove unused (by clients) SkPathUtils
BUG=skia:

Review URL: https://codereview.chromium.org/1088383003
2015-04-16 06:16:38 -07:00
reed
9645ce17cd crank up innerloop to make hairlinebench more usable/reliable
BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1087583002
2015-04-13 11:12:06 -07:00
mtklein
9ff378b01b Rewrite memset benches, then use results to add a small-N optimization.
The benches for N <= 10 get around 2x faster on my N7 and N9.  I believe this
is because of the reduced function-call-then-function-pointer-call overhead on
the N7, and additionally because it seems autovectorization beats our NEON code
for small N on the N9.

My desktop is unchanged, though that's probably because N=10 lies well within a
region where memset's performance is essentially constant: N=100 takes only
about 2x as long as N=1 and N=10, which perform nearly identically.

BUG=skia:

Review URL: https://codereview.chromium.org/1073863002
2015-04-09 14:05:17 -07:00
mtklein
f559de4338 Expand bench to cover no-draw SkPictures too.
This looks a lot closer to what Chromium's profiling is showing.

BUG=chromium:470553

Review URL: https://codereview.chromium.org/1063723002
2015-04-06 07:25:04 -07:00
mtklein
649e045170 Add a bench to measure SkPictureRecorder constant overhead.
BUG=chromium:470553

Review URL: https://codereview.chromium.org/1061633002
2015-04-03 13:25:13 -07:00
joshualitt
eef5b3eb12 BUG=skia:
Review URL: https://codereview.chromium.org/1031423002
2015-04-03 08:07:26 -07:00
mtklein
3d626834b4 New names for SkPMFloat methods.
BUG=skia:

Review URL: https://codereview.chromium.org/1055123002
2015-04-03 07:05:20 -07:00
reed
aa4c7a7042 remove useless benches
The colorfilter is applied to a single (paint's) color, so the bench does not
measure the filter at all, but simply the blit of a color.

BUG=skia:
TBR=

Review URL: https://codereview.chromium.org/1055383002
2015-04-02 20:31:17 -07:00
scroggo
2102799419 Test SkCodec to kIndex8 in nanobench.
BUG=skia:3257
BUG=skia:3475

Review URL: https://codereview.chromium.org/1051973002
2015-04-02 13:22:38 -07:00
reed
f92ace90d8 experimental speedup some xfermodes with Sk4f
Old:
   7M	1	11.1ms	11.3ms	11.3ms	11.6ms	1%	▅▄▂▂▁▁▄▄█▇	8888	Xfermode_Screen
   7M	1	10.7ms	10.9ms	10.9ms	11.1ms	1%	▄▄▄▇▃▁█▄▂▅	8888	Xfermode_Modulate
   7M	1	7.86ms	8.03ms	8ms	8.18ms	1%	█▇▅▁▃▃▂▃▆▅	8888	Xfermode_Plus
   7M	1	14.6ms	14.8ms	14.8ms	15.1ms	1%	▄█▆▅▄▁▁▆▄▆	8888	Xfermode_Xor
   7M	1	13ms	13.5ms	13.4ms	13.8ms	2%	▅▃▇▁█▂▃▅▃▅	8888	Xfermode_DstATop
   7M	1	13.1ms	13.4ms	13.3ms	13.6ms	1%	▄▁▁▆▅▄▇▆█▂	8888	Xfermode_SrcATop

New:
   7M	1	6.99ms	7.19ms	7.4ms	8.98ms	8%	▁▂▁▃▂█▁▂▂▂	8888	Xfermode_Screen
   7M	1	5.27ms	5.46ms	5.46ms	5.89ms	3%	▁▁▅▁▂█▄▃▄▃	8888	Xfermode_Modulate
   7M	1	6.8ms	7.04ms	7.27ms	8.53ms	8%	▂▁█▁▁▂▂▂▂▇	8888	Xfermode_Plus
   7M	1	9ms	9.2ms	9.33ms	10.5ms	5%	▁█▃▁▂▁▁▁▅▂	8888	Xfermode_Xor
   7M	1	8.34ms	8.57ms	8.73ms	10.6ms	8%	▁▁▁▂▂▂▂▂▂█	8888	Xfermode_DstATop
   7M	1	8.38ms	8.62ms	8.91ms	10.3ms	8%	▁▃▁▂▇▂▁▂▁█	8888	Xfermode_SrcATop

Need to define SK_SUPPORT_LEGACY_SCALAR_XFERMODES in chrome to suppress change (see https://codereview.chromium.org/1054083002/)

Review URL: https://codereview.chromium.org/1043413002
2015-04-02 12:46:24 -07:00
scroggo
081a8a4f84 nanobench does not need to handle failed rewind.
Now that all SkCodecs can rewind (assuming the stream is rewindable),
we do not need to special case it.

Pointed out by Derek in the code review that added this.

TBR=djsollen

Review URL: https://codereview.chromium.org/1058633002
2015-04-01 14:34:40 -07:00
scroggo
60869a42a1 Add timing SkCodec to nanobench.
CodecBench:
Add new class for timing using SkCodec.

DecodingBench:
Include creating a decoder inside the loop. This is to have a better
comparison against SkCodec. SkCodec's factory function does not
necessarily read the same amount as SkImageDecoder's, so in order to
have a meaningful comparison, read the entire stream from the
beginning. Also for comparison, create a new SkStream from the
SkData each time.
Add a debugging check to make sure we have an SkImageDecoder.
Add include guards.

nanobench.cpp:
Decode using SkCodec.
When decoding using SkImageDecoder, exclude benches where we decoded
to a different color type than requested. SkImageDecoder may decide to
decode to a different type, in which case the name is misleading.

TODOs:
Now that we ignore color types that do not match the desired
color type, we should add Index8. This also means calling the more
complex version of getPixels so CodecBench can support kIndex8.

BUG=skia:3257

Review URL: https://codereview.chromium.org/1044363002
2015-04-01 12:09:17 -07:00
mtklein
0340df5b36 back to Sk4f for SkPMColor
#floats

BUG=skia:
BUG=skia:3592

Review URL: https://codereview.chromium.org/1047823002
2015-03-31 08:17:00 -07:00
mtklein
c9adb05b64 Refactor Sk2x<T> + Sk4x<T> into SkNf<N,T> and SkNi<N,T>
The primary feature this delivers is SkNf and SkNd for arbitrary power-of-two N.  Non-specialized types or types larger than 128 bits should now Just Work (and we can drop in a specialization to make them faster).  Sk4s is now just a typedef for SkNf<4, SkScalar>; Sk4d is SkNf<4, double>, Sk2f SkNf<2, float>, etc.

This also makes implementing new specializations easier and more encapsulated.  We're now using template specialization, which means the specialized versions don't have to leak out so much from SkNx_sse.h  and SkNx_neon.h.

This design leaves us room to grow up, e.g to SkNf<8, SkScalar> == Sk8s, and to grown down too, to things like SkNi<8, uint16_t> == Sk8h.

To simplify things, I've stripped away most APIs (swizzles, casts, reinterpret_casts) that no one's using yet.  I will happily add them back if they seem useful.

You shouldn't feel bad about using any of the typedef Sk4s, Sk4f, Sk4d, Sk2s, Sk2f, Sk2d, Sk4i, etc.  Here's how you should feel:
  - Sk4f, Sk4s, Sk2d: feel awesome
  - Sk2f, Sk2s, Sk4d: feel pretty good

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

BUG=skia:3592

Review URL: https://codereview.chromium.org/1048593002
2015-03-30 10:50:27 -07:00
reed
7da19014fd use Sk4f for matrix math
Need to land SK_SUPPORT_LEGACY_SCALAR_MAPPOINTS in chrome to suppress Affine
version which causes slight differences (which will need to be rebaselined)

BUG=skia:

Review URL: https://codereview.chromium.org/1045493002
2015-03-29 11:58:48 -07:00
reed
d8b34c26c2 use table of procs (and unrolling) to speed up mapPts
BUG=skia:

Review URL: https://codereview.chromium.org/1040783002
2015-03-27 14:00:41 -07:00
tomhudson
573ae01295 Move HWUI boilerplate into utils/android
Duplicate code from the HWUI backends for DM and nanobench
moves into a single place, saving a hundred lines or more of
cut-and-paste.

There's some indication that this increases the incidence of
SkCanvas "Unable to find device for layer." warnings, but no
clear degradation in test results.

R=djsollen@google.com,mtklein@google.com
BUG=skia:3589

Review URL: https://codereview.chromium.org/1036303002
2015-03-27 12:22:01 -07:00