Age | Commit message (Collapse) | Author |
|
Change-Id: I6d404919d281423e636fa92d47aa3614d8d252c6
|
|
Change-Id: I8792ea1eef38e52c5159e02610198d6b6f24d13c
|
|
Change-Id: Id92f1e97d7e5032cdb5b91874119f11459f321d2
|
|
output-bmk-results
Change-Id: I9a2cf890425cba9e5b022462911c2ae0f9ac188f
|
|
Change-Id: I17779d0ab850dcd4fb6041ac3dc3536671f09951
|
|
Change-Id: I5c3438ef570c9727e88ee4329de6ad3cfe2d34f6
|
|
Change-Id: If844d1ae1da36a0f85297a57c360b6e67e4a67c9
|
|
base-artifacts branch names
Change-Id: I98a37ee40d482b2e546c76f64e9a66da2ac3e535
|
|
Change-Id: I54417bbd5c9f125a76aa61c23c682d970ffe0e43
|
|
purpose of this upgrade is to update lnt results.
Change-Id: I5474f960dbb087ba9adac7279a5b90cb3dc83534
|
|
We upgrade the version because we need to regenerate the
lnt_results.json
The base-artifacts rewrite failed because it now newly detects
some [justified] regression, and stops rewriting. To continue
this rewrite process further, a upgrade of the major is necessary.
Change-Id: I6843d2c700de11297bd407e308984afd18c01635
|
|
.. for a subset of the projects (tcwg_bmk-code_speed-spec2k6*).
The purpose of this is to update the lnt_reports.json, and push to lnt server
Change-Id: I72ab235d7bf4be09ad940dceebcf7a8bda28295a
|
|
negative score.
When encountering a score-based regression, leading to a negative
score, we consider it as an EXTERNAL_FAIL (infrastructure problem)
This will mark the job as unstable as long as the score is regressing
to a negative score. The effect is no reduction to components in that
cases. Then no attempt to update other components.
Change-Id: I34ff5fb0615bdc0769d19e0c0353084715ded178
|
|
Change-Id: I9834fc0b53dcbc2333fd4b78a20b498178aaa42b
|
|
Change-Id: Icfefa30e2916d7c33ddd9de82066dac7d3eac42a
|
|
Change-Id: Id411349d8a46e6c705681d5e625cc9beb5f1ee97
|
|
Pass correct metric_id to output-bmk-results.py ("sample", not "time")
and do not ignore failure from the script.
Also add a few FIXMEs to cleanup code.
Change-Id: I8b3a1c42a6be3dca30dab5adc8d6a8484535e107
|
|
Change-Id: Id3f00d18bcc23cfd7ebe991b0047f1cfa52b47b6
|
|
Change-Id: Ifff6f65d5c2822852c3f4719683559aed0f2bc01
|
|
Change-Id: I009b39d6489f3e3c0b48b554d50750425b855f17
|
|
round-robin-notify stage when necessary
1) Reorganize the regeneration of the details
- regenerating details are now done in xxx_generate_extra_details()
- generate_extra_details() is called as soon as we know we start full
mode.
2) Add generation of the details in case of bmk jobs.
by directly calling output-bmk-results.py in verbose mode
Change-Id: I35f226d28c1e5b9bc854b04ff3cee8137654974c
|
|
Change-Id: I061cfd24226478de89c5a8884cc2e024cb642f42
|
|
Change-Id: Ib5e83b0f4e265c2fcd2bde1773c7364648729422
|
|
Use jenkins json API to detect completion of tcwg-benchmark job
(instead of parsing console output). In rare cases a build can fail
without printing out "Finished: FAILURE" to the console.
Change-Id: Ifdd746a71c6452ba8482a0edba957a1da3b2c30c
|
|
Change-Id: Iaf5a463f17f6e88d0cd2f83c54a722c68264cde0
|
|
... which fixes failure to set CPU frequency on tcwg-bmk-qc-* boards.
Change-Id: Ib72407a68ba85e336a54265a3a80e2c176aa6808
|
|
To avoid committing unwanted files into git (e.g., raw benchmarking
data) we implement "annex" support. Files are base-artifacts/annex
can be either symlinks to directories or regular files containing
an rsync-able url.
Here we convert directory symlinks into tarballs, upload to bkp-01
and replace symlinks with files pointing to their uploaded location.
In round-robin.sh:reset_artifacts() we do the opposite: download and
extract tarball into a temporary directory, and replace the file with
a symlink to that directory.
Change-Id: I78062a5147c944f18d6cd6eb5129f0d9fb74098c
|
|
Change-Id: I930622fbd2c77e65e11663583bd17135a9bfa6d4
|
|
Change-Id: I9c7bd0f100f06a520a6111a25dd41564a47f5a67
|
|
First baseline run (successful) needs to be used as a reference.
but the results, are not pushed to bkp-01.
In that case, in tested runs, we want to use reference bmk results
directly from base-artifacts directory
Change-Id: If9500bfd9a040ee36289ca31cfa98c37e9cef7d0
|
|
Change-Id: Ic9ebfea312bb05c6e78fa061ecfe617db66e4bed
|
|
The rsync was not transfering links, therefore internal perf record
directory was uncomplete. (i.e, .build-id sym links).
This rsync should now also copies these relative symlinks.
These links pointing on the elf file are important for having funtion
names in the generated perf report.
Change-Id: I292947fa037f8af8b32b2c695923fed4237d6b35
|
|
Change-Id: Ib825c79d67ed76b5e7681464ebfb069eb99201f9
|
|
bmk-results getting&storing
tcwg-benchmark-results: -input/-output results are content of the
results_id files
round-robin-baseline: fix .gitignore
Change-Id: Ia695f951fc725c6f11e30c874bb0b51b50300828
|
|
Change-Id: Ia72e5a31b311bed9e44b7b317f8cd4d41f527fe3
|
|
.. now store in /home/tcwg-benchmark/results-xxx only if successful (in
round-robin-baseline.sh)
Change-Id: Ic700c3c8dbfefff84cb386259f54cbd70147c3f7
|
|
.. and specify remote in git annex init command
Change-Id: Iba4196e92e3eafe00878ca6e8668c969a060e732
|
|
Change-Id: I695d50628ff99f6fedc83c6bf08c0c29b436cc16
|
|
/home/tcwg-buildslave
Change-Id: I6cb051e735d9fd1c057930ba2ec478908ce1015b
|
|
Change-Id: Id1da2c9d7250a681879872d4fe95c2e37df57369
|
|
- remove ssh:// from results_dest,
- use absolute path,
- create a dir in the workspace where tcwg-benchmark can write,
- sync it in artifacts/annex,
- and remove it afterwards.
Change-Id: I1831cd2a4d42ca001fa78f89805953dcd9c306b9
|
|
Change-Id: I9de549918eced19d8aaebc177d0122a0b3b0db8f
|
|
results in artifacts
round-robin.sh initialize git-annex as pointing on
bkp-01:/home/tcwg-benchmark/base-artifacts, and consider the "annex/" directory
as annex datas to be saved with git-annex. Right now, only tcwg_bmk is using it.
Taking bare results to the artifacts & baseline simplifies a lot the
compare_results routine.
tcwg-benchmark-results.sh no more takes results_id string as inputs,
but directly takes the directories paths of each bmk results.
It takes an additional parameter hw_tag, to avoid precedent results_id
parsing. This means this script currently can only treat inputs
belonging to the same hw_tag.
bmk results are stored in :
artifacts/annex/bmk-data
We are still keeping results_id file because it is useful to find the
benchmark job id that were used to run the benchmarks.
Change-Id: I78ef05585be8f1587257f9d8950023b70bff18bc
|
|
bare bmk results in artifacts"
This reverts commit 4b36216d70e60de56e321d0b403020d27c2d44fe.
Reason for revert: pb with ci job deployment make it bancal
Change-Id: I89f94a89a0267780c7e119d03a0ab17908625310
|
|
This reverts commit 45c0646915879740a9f834b13345ae685595a703.
Reason for revert: pb with ci job deployment make it bancal
Change-Id: I84e2b3c2bad1894535d3a4db8c427e6fb42ce524
|
|
Change-Id: I4baf0e2f7f900383af51a70d5954b27ab1bdc969
|
|
results in artifacts
round-robin.sh initialize git-annex as pointing on
bkp-01:/home/tcwg-benchmark/base-artifacts, and consider the "annex/" directory
as annex datas to be saved with git-annex. Right now, only tcwg_bmk is using it.
Taking bare results to the artifacts & baseline simplifies a lot the
compare_results routine.
tcwg-benchmark-results.sh no more takes results_id string as inputs,
but directly takes the directories paths of each bmk results.
It takes an additional parameter hw_tag, to avoid precedent results_id
parsing. This means this script currently can only treat inputs
belonging to the same hw_tag.
bmk results are stored in :
artifacts/annex/bmk-data
We are still keeping results_id file because it is useful to find the
benchmark job id that were used to run the benchmarks.
Change-Id: Ibac5ed3508962f1a32ad3361e429f20c66f7eae3
|
|
Change-Id: I48c7b55797389fd73674924039b3aefdcd5a0072
|
|
- output-bmk-results.py will raise a regression only in the watched
metric.
- Also ends with an error if output-bmk-results fails, instead of
creating a error-code-$? file
Change-Id: I90396862a4c5f9fe0033177cd88cd9ffe5e1ed3c
|
|
notify stage
Change-Id: I756d09f190ded50c10fc41ed88f70f9c2925acc4
|