| c9faddb4a9c94f7ec477fea7f32ce9c4fa34c8d0 |
|
09-Dec-2016 |
Ben Chan <benchan@chromium.org> |
glbench: replace scoped_ptr with std::unique_ptr
scoped_ptr from libchrome has been deprecated in favor of C++11
std::unique_ptr.
BUG=chromium:672314
TEST=Run unit tests.
Change-Id: I0f71641e2be379c2a0efe80e1b4e7b303610bf20
Reviewed-on: https://chromium-review.googlesource.com/418397
Commit-Ready: Ben Chan <benchan@chromium.org>
Tested-by: Ben Chan <benchan@chromium.org>
Reviewed-by: Mike Frysinger <vapier@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 5e738bc5151db76ba00fc2ad9712b0cfe3a7aee6 |
|
18-Feb-2016 |
Ilja H. Friedel <ihf@chromium.org> |
graphics_GLBench: change cooldown heuristics.
Boards run much cooler these days. It makes no sense to cool them
all the way to room temperature. Instead allow them to run a little
warmer, as long as it is below 45'C. This can half test runtimes.
BUG=None.
TEST=Ran on lars.
Before: Test runs 20+ minutes, temperatures at the end
02/17 20:30:21.132 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 10.0s.
02/17 20:30:21.132 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 3.0s.
02/17 20:30:21.133 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 4.0s.
02/17 20:30:21.134 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 9.0s.
02/17 20:30:21.135 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 19.0s.
02/17 20:30:21.135 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 12.0s.
02/17 20:31:47.575 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 15.0s.
02/17 20:31:47.576 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 21.0s.
02/17 20:31:47.576 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 15.0s.
02/17 20:31:47.576 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 9.0s.
02/17 20:31:47.577 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 36.0'C (initial=32.0'C) after waiting 20.0s.
02/17 20:31:48.718 INFO | graphics_GLBench:0078| temperature_3_after_test = 41.000000 degree Celsius
After: Test runs 11 minutes, temperatures at the end
02/17 21:25:22.369 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 1.0s.
02/17 21:25:22.369 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 43.0'C (initial=31.0'C) after waiting 1.0s.
02/17 21:25:22.370 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 1.0s.
02/17 21:25:22.370 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 2.0s.
02/17 21:25:22.371 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 2.0s.
02/17 21:25:59.368 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 6.0s.
02/17 21:25:59.369 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 2.0s.
02/17 21:25:59.369 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 4.0s.
02/17 21:25:59.369 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 4.0s.
02/17 21:25:59.370 DEBUG| base_utils:0268| [stdout] Bench: Cooled down to 44.0'C (initial=31.0'C) after waiting 5.0s.
02/17 21:26:00.573 INFO | graphics_GLBench:0078| temperature_3_after_test = 47.000000 degree Celsius
Change-Id: Ia16cf80837b04b0ecb99e75b79a7ce40ad9f95cb
Reviewed-on: https://chromium-review.googlesource.com/328073
Commit-Ready: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 88959fb037a4a7a8864a671f014b09b0148095fc |
|
24-Dec-2015 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: Add verbose flag and hook it up to print per-iteration times
Add a "verbose" flag to allow printing some more verbose messages when
running the glbench test manually.
Also, add a 'verbose' message that prints per-iteration test times.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=chrome-os-partner:48800
TEST=/usr/local/autotest/deps/glbench/glbench -tests=texture_update -verbose
=> Shows per-iteration run-times
TEST=/usr/local/autotest/deps/glbench/glbench -tests=texture_update
=> Does not show per-iteration run-times (same as before this change)
Change-Id: I2d07787c62e1258753c7a4a52d644239dd88a061
Reviewed-on: https://chromium-review.googlesource.com/319920
Commit-Ready: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 6c04ea4a0f108b126378c3bb09707389714aa0a3 |
|
12-Dec-2014 |
Alex Vakulenko <avakulenko@chromium.org> |
autotest-deps-glbench: Sync sources with the copy from platform2
autotest-deps-glbench source have been moved into platform2, however
it is still being compiled from third_party/autotest/files location
because its ebuild still uses autotest-deponly eclass.
We have a bug pending (see crbug.com/394034) about changing its
ebuild to compile the sources from platform2 but until it's done
we need to make sure both locations are in sync.
The change in platform2's version of glbench is caused by the
uprev of libchrome to r307740.
BUG=chromium:394034
TEST=FEATURES=test emerge-link autotest-deps-glbench
CQ-DEPEND=CL:234440
Change-Id: I21d99c0cd918b94da721ff329a3e8964b274f984
Reviewed-on: https://chromium-review.googlesource.com/234980
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Tested-by: Alex Vakulenko <avakulenko@chromium.org>
Commit-Queue: Alex Vakulenko <avakulenko@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| c37cc9be72f3b2f1e0bd20a757b2267d5b97e1ad |
|
19-Sep-2014 |
Ben Chan <benchan@chromium.org> |
glbench: Update code to include base/files/file_util.h
file_util.h was moved from base to base/files
(https://codereview.chromium.org/468253002). This CL updates glbench
code to include base/files/file_util.h instead of base/file_util.h.
BUG=chromium:411001
TEST=`FEATURES=test emerge-$BOARD autotest-deps-glbench`
Change-Id: I794deea5a23353bb077d7eb3f699442276a38daa
Reviewed-on: https://chromium-review.googlesource.com/219041
Commit-Queue: Ben Chan <benchan@chromium.org>
Tested-by: Ben Chan <benchan@chromium.org>
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| be252420861fb14d6a79a4767c8a75a852643fc5 |
|
17-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: remove unit from .png file names
Prepending the benchmark unit to pixel readback file name was serving
no other purpose than to match the "testname" which also used to have
the unit prepended.
This change looks scarier than it is... it is mostly just file names
change, plus the corresponding .html and .txt imagename indexes, some
documentation, and a one line change to testbase.cc to stop prepending
unit_ to the generated image names.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=test_that graphics_GLBench
Change-Id: I2e0810a4bc5c4b9dcd69413f05f5fa3d4e27eb58
Reviewed-on: https://chromium-review.googlesource.com/202888
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| b722be2b3770034fe8749ab56449eef7dddde8d1 |
|
17-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: remove unit from test name
The test unit instructs graphics_GLBench how to interpret the score, but
instead of prepending it to the testname [and image name], we can just
print it once on the score line, and let graphics_GLBench parse it from
there.
A later patch also removes the unit from the image name, but that is an
independent chage that involves renaming a ton of files, so we leave it
to separate patch.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=test_that graphics_GLBench
Change-Id: Ia9d95b51e2b7297a907c1b52972fb04ac4815273
Reviewed-on: https://chromium-review.googlesource.com/202887
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| aace01b7d49e41d1bd26be3767305f57e036e0d7 |
|
17-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: refactor how tests specify the unit of their measurement
Instead of hardcoding the unit into subtest name strings, let's set the
unit on a per test basis, and have testbase append the unit as required.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=test_that graphics_GLBench
Change-Id: Ieb06ad1dd6b2bee8fb0af039e65dc357aed4df1c
Reviewed-on: https://chromium-review.googlesource.com/202886
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 7dee336595720f67f2312db842667bcb238e4869 |
|
17-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: unify score line printing and update error handling
Currently there is 1 "good" case, and 2 "error" cases (Bench() returns 0.0,
and glGetError() != GL_NO_ERROR).
The final score line is more or less the same in all 3 cases, but changes
over time have led them to diverge slightly (GL error case was missing
@RESULT, and the space between '=' and value differs between good and
error cases.
Let's refactor RunTest() a bit, such that we can have a single unified
printf that just prints both the 'good' case and the error cases,
Previously, we would use '' for the image name in the error cases.
This has the downside that '' matches the final blank line in the
"knownbad" file list, so these kinds of errors would look like known bad
images.
Instead, we now use:
* "no_score" if Bench() returns 0.0
* "glGetError=0x%02x" if glGetError() is not GL_NO_ERROR after the test
Furthermore, we just return -1.0 from glbench in the GL error case rather
than nan, since C/C++ may have trouble representing NAN on some machines.
graphics_GLBench is updated to handle to look for these error conditions.
A perf value of "-3.0" is assigned to the GL Error case, and "0.0" is used
if no perf score is returned.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=test_that graphics_GLBench
Change-Id: I7154b26849099e3530e51381ff202c46a030e35f
Reviewed-on: https://chromium-review.googlesource.com/202885
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| bcad4fb80e1115af4aad4558d93be6dabf7a3f26 |
|
17-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: allow tests to skip MD5 / image save step
Several glbench test cases don't draw anything to the framebuffer, so
reading back pixels, computing an MD5 hash on them, and saving them as a
.png is just wasted effort.
Furthermore, every time a test case is added that doesn't need this pixel
verification must also be added to the graphics_GLBench no_checksum_tests
list.
It's much easier to just let the test request that this stage be skipped,
and make graphics_GLBench just skip verification if no .png is provided.
When a test doesn't write a .png, the "imagefile" field is set to "[none]".
This is nominally done so that graphics_GLBench can distinguish this case
from a test that has 'value=0', which already uses the NULL imagefile
("[]"). Although, in truth, it is done because graphics_GLBench currently
treats NULL "[]" imagefile as a "knownbad" image, since the
glbench_knownbad_images.txt ends with a blank line.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=test_that graphics_GLBench
Change-Id: Ib8fc4b403638199f5e6390105d78dd5da4d1cab3
Reviewed-on: https://chromium-review.googlesource.com/202884
Reviewed-by: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| f899c0b4f6cb7468ff6e98941e0ef4512cc2d06a |
|
03-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: refactor testname formatting
Future test name size increases will involve slightly less effort.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=glbench
=> test results line up nicely, with no WARNING messages
Change-Id: I0e704b6ba2fe691b31780dc9e181a3741dcff6d2
Reviewed-on: https://chromium-review.googlesource.com/202477
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 8290ac901296cbf692df417412ae3546cbcce0a4 |
|
03-Jun-2014 |
Daniel Kurtz <djkurtz@chromium.org> |
glbench: add --notemp option
The temperature checking of glbench is really good in the lab when running
glbench as part of a test suite.
However, it is very annoying when running glbench manually.
The --hasty option turns off temperature checking, but it also reduces the
number of iterations and tests, which makes it much less useful.
Add a --notemp option to just turn off the temperature checking.
Signed-off-by: Daniel Kurtz <djkurtz@chromium.org>
BUG=none
TEST=glbench --notemp
Change-Id: Ieeb26f0ab184948db6c80ce9dcea66e8a8735e7c
Reviewed-on: https://chromium-review.googlesource.com/202476
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Daniel Kurtz <djkurtz@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 3907fd1607fdc5ae5065adad80b35c7733fbbf7a |
|
15-Apr-2014 |
Ilja Friedel <ihf@chromium.org> |
GLBench: add new fillrate test of variable FBO size.
* Add fbofill test: We want to measure memory throughput (in this case
bilinear copy) depending on texture size. This is interesting to study
cache behavior.
* Also changes slightly thermal throttling if we can't read temperatures
(by sleeping for a full 30 seconds between tests to ensure cold state).
* Finally add a command line option called "-hasty". This option will
run each test at at most 512x512 resolution, run it only for a fraction
of the loops we normally do, and not cool down the machine between
tests. The hasty option will be used to verify MD5s in BVT and might
also be useful for debugging test failures. Currently
glbench -save -hasty
runs in less than 30 seconds on daisy.
* This change fixes an old standing bug with the yuv_shader images. The
references for this and the new fbofill test are updated.
* We will run "glbench -hasty" on the BVT suite. It will run a subset of
tests in less than a minute total and verify MD5s of the generated
images. We will not upload any performance results from BVT.
Notice that we might need to collect more references from the lab as the
yuv images differ depending on the GPU architecture. Hence BVT is only
experimental.
BUG=chromium:355202, chromium:361925
TEST=Ran locally on daisy and link. Ran cbuildbot on alex.
Change-Id: I54039cce967664d553c79d16ca2b1361953ab9fe
Reviewed-on: https://chromium-review.googlesource.com/194991
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 3bbf4f178422d0a234a8ac0d090f691ff6344765 |
|
09-Apr-2014 |
Ben Chan <benchan@chromium.org> |
glbench: Update to build against libchrome-242728
BUG=chromium:356431
CQ-DEPEND=CL:193874
TEST=`emerge-$BOARD autotest-tests autotest-deps-glbench`
TEST=Run graphics_GLBench test
Change-Id: I1b40642b8a0ccb5c7a43c7331f8855d875a8098f
Reviewed-on: https://chromium-review.googlesource.com/193882
Reviewed-by: Daniel Erat <derat@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Tested-by: Ben Chan <benchan@chromium.org>
Commit-Queue: Ben Chan <benchan@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 6f72b2bf8d333b52f6808bcc0f1e43239de26fed |
|
11-Apr-2014 |
Ilja Friedel <ihf@chromium.org> |
Revert "GLBench: add new fillrate test of variable FBO size."
Some pictures have changed that should not have. Starting over.
This reverts commit b4efb4a15609c2af3248aa9a256a7d5e957afa78.
Change-Id: I15af5b4c4878f2f95b661b7c575a3fd71b81b075
Reviewed-on: https://chromium-review.googlesource.com/194135
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| b4efb4a15609c2af3248aa9a256a7d5e957afa78 |
|
08-Apr-2014 |
Ilja Friedel <ihf@chromium.org> |
GLBench: add new fillrate test of variable FBO size.
We would like to measure memory throughput (in this case bilinear copy)
depending on texture size. This is interesting to study cache behavior.
Also changes slightly thermal throttling if we can't read temperatures
(by sleeping for a full 30 seconds between tests to ensure cold state).
Finally adding a command line option called "-hasty". This option will
run each test at 512x512 resolution only, run it only for a fraction of
the loops we normally do, and not cool down the machine between test.
The hasty option will be used to verify MD5s in BVT and might also be
useful for debugging test failures.
Currently glbench -save -hasty runs in 15 seconds on daisy.
BUG=chromium:355202
TEST=Ran locally on daisy and link.
Change-Id: I6a21565eaf112f3373b2bab26ac122ed7b1ae5cb
Reviewed-on: https://chromium-review.googlesource.com/193576
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 3b4712cc3f0a77cfa7851fb73fb03fcb0c4e6318 |
|
14-Mar-2014 |
Ilja Friedel <ihf@chromium.org> |
Start monitoring temperature during tests.
There are two parts to this:
1) Added infrastructure to site_utils.py to get machine into defined state
and keep tests honest.
2) Change glbench binary to base its wait times on currently observed
temperatures.
This should make the test faster, as we can cut down on blind sleep times.
But I don't expect on its own to change the numbers.
BUG=chromium:344766
TEST=1) Ran graphics_GLBench on link.
2) Ran hardware_SsdDetection and firmware_TouchMTB on link to ensure
get_board() still works for them.
Change-Id: I3cb908a35d911a3f9ca6ef194aa5a3d1ac3e2152
Reviewed-on: https://chromium-review.googlesource.com/190290
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 8d7913bdbed2083caa203aa497bc6d99342a1406 |
|
25-Feb-2014 |
Ilja H. Friedel <ihf@chromium.org> |
GLBench: idle machine between benchmarks.
This is to physically cool down the machine as suggested by Android.
Also make GLBench an experimental bvt test.
BUG=chromium:344766
TEST=Ran on link.
Change-Id: Id6979270de252865beb8068931b7554c92b94011
Reviewed-on: https://chromium-review.googlesource.com/187758
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 439ea14b8019a05cd4d3f038b4d32bca1d3fe5dc |
|
22-Feb-2014 |
Ilja H. Friedel <ihf@chromium.org> |
GLBench: don't run linear regression anymore.
Instead the test runs larger and larger test problems until it crosses
5 seconds. It uses the last two runs (just below and just above the
threshold) to compute the performance values.
I hope this makes test results less noisy, especially when updating
the drivers.
BUG=chromium:344766
TEST=Ran on lumpy, compared with lab numbers.
Change-Id: Ief9ebaaefd3d7ec65d61a054658e06dc5b6c3a60
Reviewed-on: https://chromium-review.googlesource.com/187498
Reviewed-by: Ilja Friedel <ihf@chromium.org>
Commit-Queue: Ilja Friedel <ihf@chromium.org>
Tested-by: Ilja Friedel <ihf@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 88548a134aa5938a5494c2ff634c4027799963d6 |
|
28-Jan-2014 |
Ben Chan <benchan@chromium.org> |
glbench: Replace deprecated scoped_array<T> with scoped_ptr<T[]>
BUG=chromium:338610
TEST=emerge-{x86,amd64,arm}-generic autotest-tests
Change-Id: I5bed0a426993c2b6b34336bfdfc042c6fd375e2f
Reviewed-on: https://chromium-review.googlesource.com/184058
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Tested-by: Ben Chan <benchan@chromium.org>
Commit-Queue: Ben Chan <benchan@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 7f840e726dd0455d4ef338de4de89adcf2118534 |
|
19-Dec-2012 |
Simon Que <sque@chromium.org> |
glbench: refactor EGL/GLX API into its own class
- Don't leave it hanging in main.h.
- Create GLInterface class to contain the common EGL/GLX API.
- Define derived EGLInterface and GLXInterface classes in
egl/glx_stuff.h, respectively.
- In egl/glx_stuff.cc, the functions are mostly the same, just members
of the EGL/GLXInterface class.
- There is a common global pointer g_main_gl_interface that contains a
globally used GLInterface object. Calls to the EGL/GLX functions are
made through this pointer.
BUG=chromium-os:36746,chromium-os:36750
TEST=emerge autotest-deps-glbench, run glbench on link and daisy
Change-Id: I123a8005c960783df7e4c75ea5126ec8b5d8bbd5
Signed-off-by: Simon Que <sque@chromium.org>
Reviewed-on: https://gerrit.chromium.org/gerrit/39970
Reviewed-by: Stuart Abercrombie <sabercrombie@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| d5dfe5f92baa50ad3cf6a7381af6a73e11901740 |
|
16-Aug-2012 |
Stuart Abercrombie <sabercrombie@chromium.org> |
Get glbench passing.
Add clears to ensure we don't rely on previous test rendering.
Remove reference images for cases that were correct only without clears.
Add reference images for new results with clears.
Remove known bad images that coincidentally reflect the new norm.
Set the clear color to to preserve green frame edges in YUV shader tests.
BUG=chrome-os-partner:11731
TEST=Passes on i915, Sandy Bridge and Arm.
Change-Id: Iaaa0b73d623d25bcfcfc041907028986126c913c
Reviewed-on: https://gerrit.chromium.org/gerrit/30495
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Commit-Ready: Stuart Abercrombie <sabercrombie@chromium.org>
Tested-by: Stuart Abercrombie <sabercrombie@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| c33878b0587a1b46090e7249a87d32e7f88ed9ae |
|
25-Jul-2012 |
Stuart Abercrombie <sabercrombie@chromium.org> |
Switch GLBench glDrawElements calls to GL_UNSIGNED_SHORT.
We were using GL_UNSIGNED_INT which is not part of the standard GL ES spec.
BUG=chromium-os:32455
TEST=Same results on i915. Sane results for affected tests on Arm.
Change-Id: I6e055f42b11a17129c80679bf1fb5e7afc550f0c
Reviewed-on: https://gerrit.chromium.org/gerrit/28417
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Tested-by: Stuart Abercrombie <sabercrombie@chromium.org>
Commit-Ready: Stuart Abercrombie <sabercrombie@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| dd29ce1cf6d57858a2b06fdec627abc39030ab81 |
|
13-Jun-2012 |
Stuart Abercrombie <sabercrombie@chromium.org> |
Fix i915 GLBench regression.
Some tests were hitting the new non-regression-based test case and giving unrepresentative results. To prevent this, perform another dummy run before deciding whether to do regression.
BUG=chromium-os:31758
TEST=i915 results are back to the old levels
Change-Id: I3582d5ff58237be018d245d69a0e6ced214a6dd0
Reviewed-on: https://gerrit.chromium.org/gerrit/25253
Tested-by: Stuart Abercrombie <sabercrombie@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Commit-Ready: Stuart Abercrombie <sabercrombie@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| ddd7fae897257e3c689c3d1e46f47ed97312b51c |
|
04-Jun-2012 |
Stuart Abercrombie <sabercrombie@chromium.org> |
Get GLBench passing on Arm Mesa softpipe.
Reduce test run count for long tests, instead of returning failure.
Stop the texture update test from needlessly retaining pixel array allocations.
Fix EGL context leak that led to out of memory failures.
Return failure when context initialization fails.
Add softpipe reference images.
TEST=GLBench now passes on Arm softpipe. Still passes on i915.
Change-Id: I0b79676c1833d2a82cd625bf822985415692153a
Reviewed-on: https://gerrit.chromium.org/gerrit/24407
Tested-by: Stuart Abercrombie <sabercrombie@chromium.org>
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Commit-Ready: Stuart Abercrombie <sabercrombie@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 28cc0255071eb4ac79ca1fe7607163d0e712b0ad |
|
13-May-2011 |
Chris Masone <cmasone@chromium.org> |
[glbench] Roll forward to new libchrome
BUG=chromium-os:14304
TEST=build, unit tests
Change-Id: Iec1c1d95c666a57bdb0d7ed2c03f4ffe3b9d51a4
Reviewed-on: http://gerrit.chromium.org/gerrit/862
Reviewed-by: Chris Masone <cmasone@chromium.org>
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Tested-by: Chris Masone <cmasone@chromium.org>
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 8faad30e5a44271d9e50c4d75e1271018d7439b5 |
|
26-Apr-2011 |
Ilja H. Friedel <ihf@google.com> |
This change looks scary, but it shouldn't be once you ignore all the image file changes. Making the browser full screen also helps.
This change does not
a) change any of the existing tests
b) in particular it does not change any timing.
This change should do the following
1) write all images as .png
2) encode pixel MD5 in the image filename
These two things combined make it possible to easily examine images with standard viewers and do a check to classify an image by simply looking it up in a directory of known good/bad images. Also lists of good/bad images can be generated simply by doing an ls on a directory. This means checksums.txt can go away.
3) give more verbose output in case of warning/failures
4) make glbench buildable outside of chroot again controlled via #define
5) copy images into result tree in autotest so they can be easily be looked at when something breaks
6) create a human friendly summary.txt in the result tree for easy failure analysis
The beef of the changes is in testbase.cc and graphics_GLBench.py
The change was tested on mario and tegra.
BUG=chromium-os:14282
TEST=
Review URL: http://codereview.chromium.org/6880025
Change-Id: I480fb7c9c058594c5ccd32b9251a42778296db7f
/external/autotest/client/deps/glbench/src/testbase.cc
|
| a1a86cbb14725f46b095671263c04289935bc8e7 |
|
18-Jan-2011 |
Kenneth Waters <kwaters@chromium.org> |
graphics_GLBench: Fix the window manager compositing test for GLES.
- Remove legacy fixed function calls. These call weren't doing anything and
are not leagal in GLES.
- Expand varyings to vec4 to be compatible with the I915 workaround.
- Update list of tests that don't draw.
- Add gold checksums for the new test.
- Add clear to triangle rate tests.
- Work around second X server issue on tegra2.
BUG=None
TEST=Run on x86-generic and ensured that the test runs at the same speed as
before. Run on tegra2_seaboard
Review URL: http://codereview.chromium.org/5976011
Change-Id: I8c3b9de6d73322c6bc2508fdb3471349437dc85f
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 52e113a50fc80e4e359d521cf971a57091ef1150 |
|
06-Dec-2010 |
Kenneth Waters <kwaters@chromium.org> |
graphics_GLBench: Fix OpenGL-ES incompatibilites.
- Add 'NVIDIA Corporation - NVIDIA Tegra' golds.
- Don't checksum tests that rely on screen garbage.
- Increase timeout so readpixels test will run on Tegra
- Fix iteration count in ReadPixels test.
BUG=7736
TEST=Run on tegra2_seaboard and x86-generic
Change-Id: I1a323a85eaa8416d33a65afab6ee38569107b528
Review URL: http://codereview.chromium.org/5587003
/external/autotest/client/deps/glbench/src/testbase.cc
|
| b70421fcf85db631d875f3a69277a0b6e90ac6c7 |
|
13-Jul-2010 |
Alexey Marinichev <amarinichev@chromium.org> |
Converted fill rate test to GLES2.
Review URL: http://codereview.chromium.org/2913008
/external/autotest/client/deps/glbench/src/testbase.cc
|
| f50ecb1b44db52db39768f55a03280e1fd8eeac7 |
|
15-Jun-2010 |
Alexey Marinichev <amarinichev@chromium.org> |
Added gflags switches: duration, tests, save, out.
Added saving images after a testcase is finished.
Updated power_Draw control file.
Checking against golden images will be done outside from a python script.
Review URL: http://codereview.chromium.org/2794002
/external/autotest/client/deps/glbench/src/testbase.cc
|
| e735df6f6d4d0822971ede2afea45b01b5315332 |
|
25-May-2010 |
Alexey Marinichev <amarinichev@chromium.org> |
Fixes for GLES systems.
Review URL: http://codereview.chromium.org/2122020
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 9c891efd1e01010c2d9cfb1f682acb0b793e54e2 |
|
22-May-2010 |
Alexey Marinichev <amarinichev@chromium.org> |
Added texture update test.
Review URL: http://codereview.chromium.org/2115012
/external/autotest/client/deps/glbench/src/testbase.cc
|
| 9f9b873ae65c0a087ed3cd9856121a6d41e3410a |
|
21-May-2010 |
Alexey Marinichev <amarinichev@chromium.org> |
Split tests into individual files. Got rid of globals by converting them to classes.
Review URL: http://codereview.chromium.org/2123013
/external/autotest/client/deps/glbench/src/testbase.cc
|