Age | Commit message (Collapse) | Author |
|
Crosperf was assuming the user always has an NFS home directory,
and was looking for the cros_scratch directory there. This makes
crosperf use the user's home directory, whether it's NFS or local.
Also change --share_users, which also assumed NFS home directories,
to --share_cache, which requires specification of the cache path(s)
to use.
BUG=None
TEST=Tested this with my new local homedir.
Change-Id: I28d15467dfbd25b8b237cefa2979d3fd9464a7a7
Reviewed-on: https://chrome-internal-review.googlesource.com/179978
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Crosperf dies handling the results of the graphics test,
because the 'graph' field in the results is null and crosperf
was expecting it to be non-null. This fixes that problem.
BUG=chromium:424623
TEST=Ran graphics test with this fix and it no longer crashes.
Change-Id:I6764e7c38db5d1913109948d5f4798bdf5d6fb7c
Reviewed-on: https://chrome-internal-review.googlesource.com/180212
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
BUG=None
TEST=Ran regression tests successfully. Ran 'normal' crosperf tests
successfully as well.
Change-Id: Ib3387cb63f1050cc0d561cc676d91be2334e0c79
Reviewed-on: https://chrome-internal-review.googlesource.com/166133
Reviewed-by: Caroline Tice <cmtice@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
This updates the output messages to actually tell users where the
results files are stored, especially the perf.data and perf.report files.
BUG=None
TEST=Ran Crosperf with these changes.
Change-Id: I4030d8fbf47c0a313e0ab349da9c4d201f946d11
Reviewed-on: https://chrome-internal-review.googlesource.com/161486
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Crosperf's help was completely out of date and incorrect. Fixed
that. Also, I cleaned up the options, removed some that were no
longer appropriate or not ever used, and made some options global-only.
Also removed all references to pyauto. We no longer use or run pyauto
tests.
Option changes: Completely removed outlier_range (it was not actually doing
anything), key_results_only (applied only to pyauto tests), use_test_that
(not needed now that we're not using pyauto) and md5sum (never used).
Made the following options global-only: rm_chroot_tmp, perf_args, logging_level,
show_all_results, board. Removed 'build' from global settings and made
it an image-only setting (it made no sense as a global setting).
I also adjusted the unittests to work with these changes.
BUG=None
TEST=I tested all the options in all their various positions; I tested with and
wtihout caching; and I tested local and official images. I also ran the unittests. Everything works.
Change-Id: I7653663f65c0a3363a1904c0220f1066e72c7f9d
Reviewed-on: https://chrome-internal-review.googlesource.com/160758
Reviewed-by: Bhaskar Janakiraman <bjanakiraman@google.com>
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
The current implementation of the option that masks unimportant results
discards the uninteresting results before caching the test results. After
further thought, that was not the right design choice. Among other things it
can cause confusing results, such as seen in issue 357346. We should always
cache all the results, and do the result filtering during report generation.
This CL makes that change.
BUG=357346,357343
TEST=Ran crosperf tests with and without cache hits, and with and without entreies in the json file. It all seems to work as expected.
Change-Id: I778e5614c73bf751ebaa2d4606af636275247c60
Reviewed-on: https://chrome-internal-review.googlesource.com/159108
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
This is a temporary fix, until we can figure out why the 'normal' version
of 'perf' is not properly symbolizing ChromeOS perf data. This CL
commits a version of the 'perf' binary (based on perf 3.13) that does
properly symbolize ChromeOS perf data files. It also updates crosperf to
copy this working perf binary into the chroot (in /tmp) and to use it
when generating the perf report. Work to track down and fix the real
problem is ongoing, but in the meantime this allows Crosperf users to
get meaningful perf reports for ChromeOS.
BUG=None
TEST=Tested Crosperf using this patch and generating perf reports.
Change-Id: Ia8dd157b88d37a897eb9becf8fa444de41e88873
Reviewed-on: https://chrome-internal-review.googlesource.com/157756
Reviewed-by: Caroline Tice <cmtice@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Add --logging_level={verbose, average, quiet} option to crosperf,
optionally reducing the amount of logging output. "verbose" gives
the same output Crosperf always gave; "quiet" gives minimal output, and
"average" is in the middle. The default for now is "verbose". Doing this
also involved adding logging levels to the command_executer, and to
image_chromeos (again, with the default being "verbose").
BUG=None
TEST=Tested all three levels of logging, with and without having to
reimage the DUT. It all worked properly.
Change-Id: Icf8c9a6831fe25202adcb624c6c7e1d1a7ac25a5
Reviewed-on: https://chrome-internal-review.googlesource.com/156883
Reviewed-by: Caroline Tice <cmtice@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
The call to self._ProcessResults is supposed to take the show_all_results
flag as an argument. This was overlooked in the read-it-out-of-the-cache
path when this change was first introduced. This CL fixes that.
BUG=None
TEST=I tested cached and non-cached runs, with and without the
--show_all_results option. Everything worked as it should.
Change-Id: Ibf8f206406b0a4535d48b598709e62135289f993
Reviewed-on: https://chrome-internal-review.googlesource.com/156367
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
|
|
BUG=None
TEST=Tested with old and new chroots; seems to work.
Change-Id: Id1a7fc9d6b1dea0b96fe51032bb25d4722dd8b48
Reviewed-on: https://chrome-internal-review.googlesource.com/156381
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
This patch updates Crosperf to add the 'build' tag in the
experiment file, and to allow that field to contain xbuddy
syntax for using trybot and/or official builds in the test runs.
It also adds a bit more checking to make sure we have everything
necessary for running 'cros flash' before attempting to use it.
BUG=None
TEST=I have run this using an experiment file that compares a local
image, a trybot image and an official image against each other. It
all worked.
Change-Id: Ia896799061508fb5718a3201b1599d8622de0b3f
Reviewed-on: https://chrome-internal-review.googlesource.com/154097
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Fix crosperf so that it doesn't try to look for a default result
(from the json file) if the telemetry test failed altogether. Also
fix it so that if it can't find the default result (on a successful run),
it returns all the results, rather than exiting with a backtrace. Also
add units to results.
BUG=None
TEST=Tested these fixes on my system. They appear to work as desired now.
Change-Id: I86099a1174c4d91d7d92a5aeea0792c68fa8dbbb
Reviewed-on: https://chrome-internal-review.googlesource.com/154114
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Some Telemetry tests return results that have the same 'description'
name, but different 'graph' names. We were only using the description
names, so we were losing some Telemetry measurements in our reports.
This CL fixes that.
BUG=None
TEST=Ran this on with smoothness.top_25. Was missing some results
before this change; after, the results are all there.
Change-Id: Ibcbe31da5f2860c494cbc8364b0a720d1c50169c
Reviewed-on: https://chrome-internal-review.googlesource.com/150215
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
When running telemetry_Crosperf, use the json file of report defaults to
determine which results to actually include in the report. This CL
also adds a flag to Crospref, --show_all_results=True, that allows the
user to override the defaults and generate a report showing all the values.
BUG=None
TEST=Tested this both with and without the --show_all_results flag. It
works as expected for me.
Change-Id: I91364572ebdf3d54bb8b408fbd5a2580923f6829
Reviewed-on: https://chrome-internal-review.googlesource.com/149457
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Sometimes _GetDataMeasurementsFiles returns the file with the full
directory path, and sometimes it doesn't. We were assuming it always
did not. This patch checks to see if we can already find the file
before attempting to prepend the directory path.
BUG=None
TEST=Tested this in the nightly infrastructure tests area, where we
were seeing the problem. This fixed that problem. (Before this fix
we were not seeing the successful test results; after it we are.)
Change-Id: Ic0a2df4c9aff7b820f652f7ecad56976ed97a4b0
Reviewed-on: https://chrome-internal-review.googlesource.com/149256
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
run_benchmark.
BUG=None
TEST=I have tested this with my own crosperf, and test_that is passing the
arguments along properly. The telemetry_Crosperf autotest also needs to
be updated before this completely works.
Change-Id: Ic887380b297d966229ac21ba4e7e909e7248fcb0
Reviewed-on: https://chrome-internal-review.googlesource.com/148485
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
The latest version of telemetry_Crosperf generates a perf_measurements
file, containing the test results, rather than a keyval file. This CL
checks for perf_measurements, and reads it, if it is there. This
change is backwards compatible with the older versions of telemetry_Crosperf
that used the keyval file.
BUG=None
TEST=Tested this with the results only in a keyval file and only in a
perf_measurements file. Got acceptable test reports either way.
Change-Id: If10a62b2f7fbe89d35c1ebe789a81764c9d134e4
Reviewed-on: https://chrome-internal-review.googlesource.com/148295
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
If rerun=True, the data in the cache is useless, so remove it first
to avoid data corrupture.
BUG=None
TEST=The cache dir is removed.
Change-Id: I59b219c57cc99aaea7a734ef6e7fd91007cedfd3
Reviewed-on: https://chrome-internal-review.googlesource.com/146065
Reviewed-by: Luis Lozano <llozano@chromium.org>
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Yunlian Jiang <yunlian@google.com>
Tested-by: Yunlian Jiang <yunlian@google.com>
|
|
Add a new flag, --use_test_that=True/False, to crosperf, to allow users to
tell crosperf to use the test_that script rather than the run_remote_tests.sh
script for running their tests.
BUG=None
TEST=Tested with and without the flag, to verify that it does the right thing.
Change-Id: I1a040d23adbf364b4a47320d9fd759a2d46e6689
Reviewed-on: https://gerrit-int.chromium.org/43109
Reviewed-by: Luis Lozano <llozano@chromium.org>
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
This is the same as the previous CL; I had to abandon the previous CL
because it erroneously depended on a different abandoned CL, so it refused
to commit.
BUG=none
TEST=I ran the telemetry tests with the modification and now ALL of them
give good output (before only about half of them did).
Change-Id: I567f51176c522bc1c41676f6dc560f3827b9cef4
Reviewed-on: https://gerrit-int.chromium.org/42930
Reviewed-by: Bhaskar Janakiraman <bjanakiraman@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
Fix issues causing errors in nightly tester: Attempting to use
machine_id_checksum when it hasn't been initialized; and attempting to
unlock an already unlocked machine causes a fatal error.
Change-Id: Ia149953d4e25c7abf97a189bb5fdd129de865c34
Reviewed-on: https://gerrit-int.chromium.org/42489
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Caroline Tice <cmtice@google.com>
Tested-by: Caroline Tice <cmtice@google.com>
|
|
BUG=None
TEST=run crosperf with pyauto/telemetry test with/without cache.
all pass.
Change-Id: If07ac020a9107a79d5780a58fd6dcc924d07f07f
Reviewed-on: https://gerrit-int.chromium.org/36594
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Yunlian Jiang <yunlian@google.com>
Tested-by: Yunlian Jiang <yunlian@google.com>
Reviewed-on: https://gerrit-int.chromium.org/39241
Reviewed-by: Yunlian Jiang <yunlian@google.com>
|
|
So others can open the cache dir
BUG=None
TEST=run crosperf, the permission of the cache dir shown as drwx--x---
Change-Id: Iadc5519679f2820f047e17c10429205636e65f5b
Reviewed-on: https://gerrit-int.chromium.org/34614
Reviewed-by: Luis Lozano <llozano@chromium.org>
Commit-Queue: Yunlian Jiang <yunlian@google.com>
Tested-by: Yunlian Jiang <yunlian@google.com>
|
|
This bug happens when we move the invocation of generate_test_report
into chroot. The source dir did not change accordingly, so it cannot
remove the /tmp stuff anymore.
BUG=None
TEST=The /tmp/... stuff is removed on both cache and non cached case.
Change-Id: I29dac68eecf0ff1d265f2a3f004627c30962e228
Reviewed-on: https://gerrit-int.chromium.org/34407
Reviewed-by: asharif <asharif@google.com>
Commit-Queue: Yunlian Jiang <yunlian@google.com>
Tested-by: Yunlian Jiang <yunlian@google.com>
|
|
Got rid of stale copies of some tools like "crosperf" and moved all
files under v14 directory (that came from perforce) into the top
directory.
BUG=None
TEST=None
Change-Id: I408d17a36ceb00e74db71403d2351fd466a14f8e
Reviewed-on: https://gerrit-int.chromium.org/33887
Tested-by: Luis Lozano <llozano@chromium.org>
Reviewed-by: Yunlian Jiang <yunlian@google.com>
Commit-Queue: Luis Lozano <llozano@chromium.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|