Status meeting May 31, 2012¶
Geometry on GPGPU¶
Soon is ready for testing. Geometry and navigation classes. Using some of the tools from John A’s student last summer. He wants a representation of the current geometry on the GPU. He observed the pointers in the construction of the geometry tree. Hard to understand how much work is needed to synchronize the host geometry with the GPU geometry using the student tools (the student tools flatten or do something with the pointers to make them copy correction to the GPU).
Measured the step sizes (100GeV pions). Where are the results written up?
What needs to be completed before we move on?
When does the question of what to do next need to be answered?
Dongwook looks at what the EM physics process is doing for Bramsstralung. Physics might be explained in the Geant4 physics guide. The geometry might be a problem. Suggestions were to implement process as defined in the physics manual for Geant4. Daniel asked about why reimplement instead of coding based on the current implementation.
Key is to learn about how the EM physics works.
What next? (continue)¶
Soon talked with Philippe and asked where we were headed - said we do not know where we are going and to get input from what we are doing (limits of hardware).
Geant4 collaboration meeting on the 1st of September - good goal for write-up.
Daniel asked for people to report when their vacation plans are. (next meeting)