Ticket #1345 (closed: duplicate)
CUDA-based event processing
Reported by: | Janik Zikovsky | Owned by: | Janik Zikovsky |
---|---|---|---|
Priority: | trivial | Milestone: | Release 2.0 |
Component: | Mantid | Keywords: | |
Cc: | pf9@… | Blocked By: | |
Blocking: | Tester: | Nick Draper |
Description
Explore the possibility of using CUDA-enabled hardware to do parallel processing on event lists, pixel-by-pixel. I am envisaging one algorithm that would handle several features at once (thereby minimizing back-and-forth memory transfer):
- Histogram events pixel-by-pixel
- Filter events (by wall-clock time, pulseID, others?)
- Perform time-focussing
The algorithm could obviously do only part of this, i.e. histogram without filtering.
Change History
comment:1 Changed 10 years ago by Janik Zikovsky
- Milestone changed from Iteration 24 to Iteration 25
comment:2 Changed 10 years ago by Nick Draper
- Priority changed from major to trivial
Interesting, but priority lower than scientific functionality requests
comment:4 Changed 10 years ago by Janik Zikovsky
- Component set to Mantid
- Milestone changed from Iteration 25 to Unassigned
comment:5 Changed 9 years ago by Janik Zikovsky
- Milestone changed from Unassigned to Iteration 30
Moved my tickets that won't get finished by iteration 29
comment:6 Changed 9 years ago by Nick Draper
- Milestone changed from Iteration 30 to Iteration 31
Bulk move of tickets to iteration 31 at the iteration 30 code freeze
comment:7 Changed 9 years ago by Janik Zikovsky
- Status changed from new to verify
- Resolution set to duplicate
See Ticket #3635. OpenCL is more attractive than CUDA since it is not vendor-specific.
comment:8 Changed 9 years ago by Nick Draper
- Status changed from verify to verifying
- Tester set to Nick Draper
comment:9 Changed 9 years ago by Nick Draper
- Status changed from verifying to closed
agreed, we only need one of cuda or open cl
comment:10 Changed 5 years ago by Stuart Campbell
This ticket has been transferred to github issue 2192