-
Notifications
You must be signed in to change notification settings - Fork 15
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
hitindex for pLS is a bit confusing #267
Comments
I have a couple of questions on this issue and I would like to also bring @tresreid to the discussion, as he may understand a bit better the situation:
Maybe my confusion comes from the fact that you quoted twice the same piece of code, while you wanted to quote something else in the second case? |
After studying a bit more point 2, I think I reached the following understanding: Situation for OT hits
Situation for IT hits
Some parting thoughts
|
Because hit index is already referenced against hitsInGPU.idxs here:
TrackLooper/SDL/Event.cu
Lines 678 to 681 in 398270a
Downstream it gets confusing.
TrackLooper/SDL/Event.cu
Lines 678 to 681 in 398270a
There are two hit indexs (1) within GPU memory, (2) within CMSSW / tracking Ntuple.
(1) idxs is internally used, and then we obtain (1) -> (2) mapping via hitsInGPU.idxs[].
But for pLSs, where the internal indexs should be saved, it already saved the dereferenced version (1) -> (2) into the internal memory.
So far this has not caused a problem, because in the special case of # of outer tracker hits being > N inner tracker hits used in the pLSs, we are fine.
However, in the case tha N outer tracker hits < inner tracker hits from pLSs, we may be in trouble.
The text was updated successfully, but these errors were encountered: