-
Notifications
You must be signed in to change notification settings - Fork 4.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
migrate modules used by the HLT menu to multithreading #10909
migrate modules used by the HLT menu to multithreading #10909
Conversation
- change all configuration parameters into const members - initialise them in the member initializer list - remove unnecessary parameters - remove unnecessary caching of EventSetup products - remove unused methods and member variables - remove usage of exceptions in the FED check loop - change into a global::EDFilter
@cmsbuild please test |
The tests are being triggered in jenkins. |
A new Pull Request was created by @fwyzard (Andrea Bocci) for CMSSW_7_6_X. migrate to modules used by the HLT menu multithreading It involves the following packages: Alignment/LaserAlignment @perrotta, @cmsbuild, @diguida, @cvuosalo, @fwyzard, @cerminar, @Martin-Grunewald, @slava77, @mmusich, @mulhearn can you please review it and eventually sign? Thanks. |
clock_gettime(CLOCK_REALTIME, &start_time); | ||
*/ | ||
|
||
void RPCPointProducer::produce(edm::StreamID, edm::Event& iEvent, const edm::EventSetup& iSetup) const { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this looks way too simple compared to the failed attempt in #9760
DTSegtoRPC etc are using global (static) cache variables
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for pointing it out.
Indeed I have checked for static analysis reports only for the modules themselves, not for the other code that they use.
Is there a simple way to find all thread-safety issues related to a plugin ?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I know only of running helgrind
@Dr15Jones may know better
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Best way to do it is to walk yourself through the code in CMSSW, lets you
catch everything but externals with problems and then really complicated
template code like stuff in reco::parser.
helgrind works but it is a bit noisey and thinks all of the std::atomic<>
variables are causing data races, hard to figure out what exactly is
happening.
On Mon, Aug 24, 2015 at 1:57 AM, Slava Krutelyov [email protected]
wrote:
In RecoLocalMuon/RPCRecHit/src/RPCPointProducer.cc
#10909 (comment):-RPCPointProducer::~RPCPointProducer(){
-}
-void RPCPointProducer::produce(edm::Event& iEvent, const edm::EventSetup& iSetup){
- /*
- struct timespec start_time, stop_time;
- time_t fs;
- time_t fn;
- time_t ls;
- time_t ln;
- clock_gettime(CLOCK_REALTIME, &start_time);
- */
+void RPCPointProducer::produce(edm::StreamID, edm::Event& iEvent, const edm::EventSetup& iSetup) const {
I know only of running helgrind
@Dr15Jones https://github.com/Dr15Jones may know better—
Reply to this email directly or view it on GitHub
https://github.com/cms-sw/cmssw/pull/10909/files#r37712383.
-1 DAS Error 1000.0 step1 DAS Error 1001.0 step1 DAS Error 1003.0 step1 DAS Error you can see the results of the tests here: |
cf42e60
to
eb41cda
Compare
I've dropped the migration of RPCPointProducer for the time being. |
Look at http://cms-sw.github.io/showIB.html and click on the link 'Modules to thread unsafe statics'. The link 'Modules to thread unsafe EventSetup products' isn't accurate because some safe EventSetup products are falsely marked |
edm::EDGetTokenT<EcalRecHitCollection> tok_ee; | ||
edm::EDGetTokenT<EcalRecHitCollection> tok_eb; | ||
edm::EDGetTokenT<trigger::TriggerFilterObjectWithRefs> tok_trigcand; | ||
virtual void beginJob() override; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The beingJob and endJob methods are empty, why not just get rid of them?
From a cursory look, everything seems fine. I also couldn't find any of the modified modules in the most recent 'module to statics' report: |
uhm, a very timely rebase (just seconds after I signed) |
@cmsbuild please test |
The tests are being triggered in jenkins. |
well, I did the rebase simply to fix the typo you spotted, do we really need to test it all again ? |
@fwyzard was it possible without a rebase? |
here is the diff between the version you signed (67f7701 ) and the new version ( 2259c5d ):
|
By the way, if you prefer I can make a separate PR with only the RECO changes ? |
no particular preference, I can figure out which files are RECO-related. |
-1 Andrea, please close this one (or the others) so that we don't have duplicated features |
OK, I've opened #10964 to keep track of all the related pull requests. |
migrate the following modules to be
stream
orglobal
modules: