TICLGeom: do not schedule the @alpaka geometry ES producers in FastSim - #51657
Conversation
|
cms-bot internal usage |
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51657/50521
|
|
A new Pull Request was created by @felicepantaleo for master. It involves the following packages:
@Moanwar, @cmsbuild, @jfernan2, @mandrenguyen, @srimanob can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
|
test parameters:
|
|
@cmsbuild please test |
|
type bugfix |
|
urgent |
|
-1 Failed Tests: RelVals Failed RelVals |
|
the fastsim wf without pileup is fixed. The error in the pileup wf is not connected to the PR |
|
test parameters:
|
|
@cmsbuild please test |
|
+1 Size: This PR adds an extra 20KB to repository Comparison SummaryThere are some workflows for which there are errors in the baseline: Summary:
|
Add the @alpaka TICLGeom EventSetup producers to hgcalLocalRecoTask instead of attaching them to every phase2_hgcal process. Their EventSetup modules are then instantiated only when the reco task is scheduled, which is where the Alpaka accelerator is present, so a process that only loads the reconstruction sequence without scheduling it (FastSim, the tracker geometry dump tests) no longer fails to find TICLGeomESProducer@alpaka at EventProcessor construction.
32848ea to
bf7020b
Compare
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51657/50522
|
|
@cmsbuild please test |
|
+1 Size: This PR adds an extra 24KB to repository Comparison SummarySummary:
|
|
+1 |
|
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will be automatically merged. |
I'm on it. |
Phase2 FastSim is a
phase2_hgcalprocess that loads no Alpaka backend and runs no HGCal reco, so scheduling the@alpakaTICLGeom EventSetup producers made the EventProcessor fail to find the plugin at construction.