Closed daggarwa closed 5 years ago
@daggarwa you should put some information about this new option in readme file.
@daggarwa you should put some information about this new option in readme file.
@monidzik Sounds good! I have added the cmake-args info for ROS2PollingSub
here like Cyclone DDS
and others : https://github.com/ApexAI/performance_test/commit/fa25cd1a7aa72acf8e8928737799a9fcda81b04a#diff-fcc41d9bd699e1a53818d508382f0e0aR37
I know you are working on correcting the qos settings so I can compile your branch, so let me check your code again when it's ready to run and CI is passing.
@monidzik I have made the requested changes (most of them except one : https://github.com/ApexAI/performance_test/pull/88#discussion_r328612243) need your input on that. Also the CI passes now. I believe the PR is ready for a second round of review. I have assigned it to you again. Thanks!
@monidzik I guess I have addressed all your review comments. Thanks so much for the review. As discussed I am now assigning the PR to @deeplearningrobotics for further review.
Ok, you still need to add the readme information with the option description, otherwise I have only couple of small remarks.
@monidzik I forgot to respond on this earlier. I can do the update to the readme now and push that.
@monidzik: Please also review my comments.
@monidzik: Please also review my comments.
@deeplearningrobotics @monidzik I have addressed's all of @deeplearningrobotics 's comments. Please review once. Thanks.
@monidzik: Please also review my comments.
@deeplearningrobotics @monidzik I have addressed's all of @deeplearningrobotics 's comments. Please review once. Thanks.
@daggarwa ok, next time please just assign to me, I didn't see this message, sorry!
@deeplearningrobotics Divya has some comments for you to resolve, so leaving them for you!
@monidzik: Please also review my comments.
@deeplearningrobotics @monidzik I have addressed's all of @deeplearningrobotics 's comments. Please review once. Thanks.
@daggarwa ok, next time please just assign to me, I didn't see this message, sorry!
@monidzik Okay I will do that from next time. Thanks for reviewing this.
@deeplearningrobotics As discussed in the morning, I have committed changes to resolve https://github.com/ApexAI/performance_test/pull/88#discussion_r329707857 and https://github.com/ApexAI/performance_test/pull/88#discussion_r329718116
Can you please review the changes here : https://github.com/ApexAI/performance_test/pull/88/commits/aa11e1d46ca92f5ec3fd146a752e459cc7290e41 and continue your review of the PR. Thanks
@deeplearningrobotics I think I have addressed all your comments now. Can you please review once. Assigning back to you.
@monidzik: I reviewed the changes and the code looks good to me. Please test all the communication plugins and check if everything is still functional and then approve.
@daggarwa: FYI
@daggarwa: Can you also run it trough pref to make sure we did not introduce performance bottlenecks.
@daggarwa: Can you also run it trough pref to make sure we did not introduce performance bottlenecks.
@deeplearningrobotics is "pref" a typo? Did you mean https://perf.wiki.kernel.org/index.php/Tutorial?
It is possible that @daggarwa will not understand what is meant by running it through perf - can you point her to what explicitly should be done?
@daggarwa: Can you also run it trough pref to make sure we did not introduce performance bottlenecks.
@deeplearningrobotics is "pref" a typo? Did you mean https://perf.wiki.kernel.org/index.php/Tutorial?
It is possible that @daggarwa will not understand what is meant by running it through perf - can you point her to what explicitly should be done?
@dejanpan: Yes, I mean perf
sorry.
@daggarwa: Lets look together at the perf
graphs when you are back at the office so I can show you what to look out for.
@daggarwa: Can you also run it trough pref to make sure we did not introduce performance bottlenecks.
@deeplearningrobotics is "pref" a typo? Did you mean https://perf.wiki.kernel.org/index.php/Tutorial? It is possible that @daggarwa will not understand what is meant by running it through perf - can you point her to what explicitly should be done?
@dejanpan: Yes, I mean
perf
sorry.@daggarwa: Lets look together at the
perf
graphs when you are back at the office so I can show you what to look out for.
@dejanpan Thank you for the clarifying question. @deeplearningrobotics Okay sounds good.
@daggarwa looks good to me, I checked also other communication means and everything works fine. The only thing is that you are still not mentioning anything about this option in readme - did you decide not to write there anything?
@monidzik I am planning to add ROS2 Waitset
like here : https://github.com/ApexAI/performance_test/pull/88/commits/09fd145426e6cb3ca501d7ac9a4b0d84d30121e5 to be consistent with other implemented plugins
@daggarwa looks good to me, I checked also other communication means and everything works fine. The only thing is that you are still not mentioning anything about this option in readme - did you decide not to write there anything?
@monidzik I am planning to add
ROS2 Waitset
like here : 09fd145 to be consistent with other implemented plugins
@monidzik I have added the changes above to the readme. Kindly review.
@deeplearningrobotics The profiling with perf tool is done. There are no performance impacts if we add this plugin in. Also , rebased the branch off master and cleaned up the commits. Please let me know if anything else is required before it gets merged in! Thanks
Added polling subscription plugin for ROS 2 to the performance test. It can be used by specifying the argument
-c [ --communication ]
arg Communication plugin to use asROS2PollingSub
when running performance test .To turn on the option for using Polling Subscription Plugin we need to compile the performance test with cmake args
PERFORMANCE_TEST_POLLING_SUBSCRIPTION_ENABLED
and value asON
:Sample usage for the plugin can be :
I have verified that this plugin builds and runs successfully on ApexOS version :
0d34c574917d
This change is