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
SKIMS in Conf data processing 75 #10227
SKIMS in Conf data processing 75 #10227
Conversation
please test |
The tests are being triggered in jenkins. |
A new Pull Request was created by @franzoni (Giovanni Franzoni) for CMSSW_7_5_X. SKIMS in Conf data processing 75 It involves the following packages: Configuration/DataProcessing @cmsbuild, @franzoni, @davidlange6 can you please review it and eventually sign? Thanks. |
+1 |
This pull request is fully signed and it will be integrated in one of the next CMSSW_7_5_X IBs once checked with relvals in the development release cycle of CMSSW or unless it breaks tests. This pull request requires discussion in the ORP meeting before it's merged. @davidlange6, @Degano, @smuzaffar |
This pull request is fully signed and it will be integrated in one of the next CMSSW_7_5_X IBs once checked with relvals in the development release cycle of CMSSW (tests are also fine). This pull request requires discussion in the ORP meeting before it's merged. @davidlange6, @Degano, @smuzaffar |
Hi @franzoni , |
The second comment I have comes from my notes on the conversation I had with Dirk about this. You've done part of what he suggested except what he called "PromptSkims" you call "PhysicsSkims" which is fine. What I don't see (and maybe I just missed it), is the place where the filter sequence you define is used to configure an output module that uses that filter sequence to select events. That's probably because that code is in the T0 code here: https://github.com/hufnagel/T0/blob/master/src/python/T0/RunConfig/RunConfigAPI.py#L432 |
hello @sextonkennedy @hufnagel ,
The ConfigBulder,
In light of what the ConfigBulder does, is this:
actually necessary ? I'd imagine not. I can find some more time to work on this today. I'd like to be done by tomorrow morning, for what is needed by the PdmV/PPD side,. Can we iterate, if needed ? Cheers, |
That doesn't work for the Tier0. Tier0 is not WMAgent. WMAgent scans the configuration in ConfigCache and extract output module definitions. Tier0 constructs the config at runtime based on the information it passed into Config.DP. Anything that happens behind the scene in ConfigBuilder to add "magic" output module parameters is a no-go for the Tier0. I need to know upfront what the output module parameters are going to be, at least at the level of the outputs dictionary syntax passed to ConfigBuilder supports. |
After looking at the code, adding a physicskim step should NOT add any output modules. I need to add them myself from the Tier0 side by passing them in via the outputs dict that is passed to ConfigBuilder. That is the only supported way in the Tier0. Which means there needs to be some set if ConfigBuilder output parameters (dataTier0, eventContent, selectEvents) that gives me what we want for this output. |
@hufnagel this is funny: we implement this to circumvent Tier0 shortcomings and stumble in new ones. |
+1 |
…_7_5_0 SKIMS in Conf data processing 75
FW port of #10226
Includes a new key (PhysicsSkims) to allow inclusion of SKIM: in the prompt processing, in the same step as RECO
added a new test which probes such new key for "ppRun2" "ppRun2B0T" scenarios
@cerminar
@hufnagel : can you please take a look and comment ? The integration with T0 needs to follow the same logic of the ALCA, using the skim matrix with @PRIMARYDATASET syntax. The committed example in run_CfgTest.sh demostrates the principle
@sextonkennedy : we spoke about this, do you have comments ?