-
Notifications
You must be signed in to change notification settings - Fork 393
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
Segmentation Fault with combineTools using HybridNew in Combine-v10/CMSSW_14_1_0_pre4 #1001
Comments
Hi @sergicastells9, could you please share the datacards with the workspaces? It could be something related to pdfs you are using, but it is hard to say more without workspaces. |
Hi, thanks for the quick response. I've attached my datacard (and the compiled one for completeness) and my workspace for your reference. Let me know if I can provide anything else that might help. For some more context, I use HiggsAnalysis-CombinedLimit with tag v10.0.2 and CombineHarvester with tag v3.0.0. |
Hi @sergicastells9, I think the segfault happens here because you are using |
Hi @anigamova, I'm still learning how everything works with Combine so in my tests I had left nuisance parameters empty. I see now why the |
A quick note that might help understanding why this happens more: I tried running |
In this attempt did you run with combine v9 or v10? I think one of the reasons it might be not working with v9 is because you created |
I've recently upgraded from CMSSW_11_3_4 to CMSSW_14_1_0_pre4 with Combine v10 and am trying to run HybridNew. I am doing a parametric analysis that uses RooPdfs for the signal/background models. My setup used to work with CMSSW_11_3_4, but now, with CMSSW_14_1_0_pre4, I get a segmentation fault when running
combine -M HybridNew -m 125 -d ../combine/15_GeV/h4g_15_GeV_datacard.txt -n 15GeV --LHCmode LHC-limits --saveHybridResult
. I've gone through both the ROOT and Combine changelogs and haven't found anything that might explain this behavior.I have no issues with text2workspace, FitDiagnostics, Significance, or
-H HybridNew
(until it gets to the HybridNew part). The full stack is below. I have also tested runningcombine datacard-2-template-analysis.txt -M HybridNew --LHCmode LHC-limits --rMax 2.0
which did not have any issues, where the datacard/workspace are inHiggsAnalysis/CombinedLimit/data/tutorials/CAT23001/
...Can anyone provide any insight into what might be going wrong? I'm totally at a loss here. If I'm missing any information or can upload my workspace for reference then just let me know. Many thanks in advance.
The text was updated successfully, but these errors were encountered: