-
Notifications
You must be signed in to change notification settings - Fork 19
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
ref: Generalize propagation benchmark functionality #404
ref: Generalize propagation benchmark functionality #404
Conversation
fe6d7da
to
5ff2e0e
Compare
Is this still in the development? Because I saw openmp macro in the benchmark code |
True, the openMP part made it into main, but I would like to revisit the refactoring of the benchmark done here |
441ce27
to
484644b
Compare
Close this one? |
Please leave this open, I have been working on this recently to benchmark different scenarios (e.g. with/without covarance transport/material maps). I just need to find a way to integrate it with the recent benchmark monitoring... |
484644b
to
207bb0d
Compare
5d9b666
to
18ac817
Compare
7758f81
to
fe1b1c9
Compare
This comment was marked as outdated.
This comment was marked as outdated.
0206094
to
3eef7e4
Compare
3eef7e4
to
52ccd54
Compare
5c6ac31
to
531da1f
Compare
The GPU benchmarks can now also be run the same way. In particular, with and without covariance transport
CPU (array plugin):
|
cf3e004
to
2550f43
Compare
2550f43
to
ddd3719
Compare
|
This finished, but needs a MR in the detray-benchmark repository first. |
The gitlab benchmark MR was merged, so this PR needs to be the next in line, otherwise the corresponding benchmark run will fail. Could someone please do a review? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good
Generalize the benchmark functionality to different detectors and actor setups. Also splits some common benchmark functionality into a detray benchmark library, like e.g. the generation of track samples.
The CPU benchmarks have also been switched to use dynamic scheduling for load balancing.
Also adds the charge conjugation operation to the pdg particle, so that the charge hypothesis can be updated when test tracks are generated with randomized charge (this was triggering an assertion in the benchmarks otherwise)