sphenix-tracking-l AT lists.bnl.gov

Subject: sPHENIX tracking discussion

List archive

From: Christof Roland <christof.roland AT cern.ch>
To: Hugo Pereira Da Costa via sPHENIX-tracking-l <sphenix-tracking-l AT lists.bnl.gov>
Subject: [Sphenix-tracking-l] More magfield memory surprises...
Date: Thu, 5 May 2022 16:49:48 +0200

Hi Everybody,

I replaced the big magnetic field in a few more modules other than the MakeActsGeometry.

It looks like we kept a few more copies of the full field than we thought.

The tracking on cluster files from hijing min bias + 50kHz pileup seems to run well under 2GB now.

I hd a bief look at the efficiency and it looks ok, more jobs are running.

If this turns out to be correct it may change our strategy quite a bit.

I added the changes necessary to run this into the clusterizer pull request (#1481) to

get a quick answer from jenkins if this runs ok.

If you could also have a look at this to help validate, this would help a lot.

I still assume this is too good to be true.

We should redo our benchmarks with this setup if things hold.

thanks

Christof

[Sphenix-tracking-l] More magfield memory surprises..., Christof Roland, 05/05/2022