Skip to content

Update GPU and FST documentation #14340

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.

Sign up for GitHub

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

Merged
davidrohr merged 1 commit into from
May 28, 2025
Merged

Update GPU and FST documentation #14340

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions GPU/documentation/build-standalone.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,6 +71,8 @@ For a reaslistic benchmark using GPU 0 on the EPNs, please use:
ROCR_VISIBLE_DEVICES=0 numactl --membind 0 --cpunodebind 0 ./ca -e o2-pbpb-100 --gpuType HIP --memSize 15000000000 --inputMemory 6000000000 --outputMemory 10000000000 --sync --runs 10 --RTCenable --setO2Settings --PROCdoublePipeline
```

Note that on the MI50 nodes, we use only <16 GB of memory, since there is a performance regression when using the upper half of the 32 GB. In order to fit in the 16 GB, we have reduced the time frame length to 32 orbits from 2024 onwards.

# Generating a dataset

The standalone benchmark supports running on Run2 data exported from AliRoot, or to run on Run3 data from O2. This document covers only the O2 case.
Expand Down
3 changes: 3 additions & 0 deletions prodtests/full-system-test/documentation/full-system-test-setup.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,9 @@ This will use 4 GPU with the HIP backend and allocate 22 GB of scratch memory on
SHM_MANAGER_SHMID=1 TFDELAY=2.8457 NTIMEFRAMES=8 $O2_ROOT/prodtests/full-system-test/start-tmux.sh dd
```

Note that as of 2024, we are running with a TF length of 32 orbits.
To simulate time frames accordingly, please set `NEVENTS=200 NHBPERTF=32` for full-system-test.sh.

---

# Remarks for running with distortions:
Expand Down
4 changes: 4 additions & 0 deletions prodtests/full-system-test/documentation/full-system-test.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,10 @@ For a simulation of a full 128 orbit time frame, run
```
NEvents=650 NEventsQED=30000 SHMSIZE=128000000000 TPCTRACKERSCRATCHMEMORY=30000000000 $O2_ROOT/prodtests/full_system_test.sh
```
As of 2024, we are using 32 orbit time frames. Fo doing so, please set:
```
NEVENTS=200 NHBPERTF=32
```
To simulate collisions with an embedded signal one can set 'D0_EMBEDDING=1' while also supplying a config file with the desired settings using 'FST_EMBEDDING_CONFIG'. For an example configuration see '/prodtests/full-system-test/pythia8.cfg' and to generate a specific configuration one can use '${O2DPG_ROOT}/MC/config/common/pythia8/utils/mkpy8cfg.py'. Additional examples can be found in '/run/SimExamples/'
```
DO_EMBEDDING=1 NEvents=5 NEventsQED=100 $O2_ROOT/prodtests/full_system_test.sh
Expand Down