Skip to content

Conversation

@f3sch
Copy link
Collaborator

@f3sch f3sch commented Jul 12, 2025

No description provided.

Signed-off-by: Felix Schlepper <felix.schlepper@cern.ch>
@github-actions
Copy link
Contributor

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1
async-2024-PbPb-apass2
async-2023-PbPb-apass5

@mconcas
Copy link
Collaborator

mconcas commented Jul 12, 2025

I would mot remove the files for the smoother. You can remove includes if you want, but please leave it there.

@f3sch
Copy link
Collaborator Author

f3sch commented Jul 13, 2025

Please correct if I am wrong but I think you referring to #14504?

@mconcas
Copy link
Collaborator

mconcas commented Jul 14, 2025

Please correct if I am wrong but I think you referring to #14504?

definitely

Copy link
Collaborator

@mconcas mconcas left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just for your reference: the computeVerticesInRofs (and similia) are the corresponding CPU version of what was done on the GPU version of the vertexer at the time we had the possibility to fragment the single TF for processing it on the device.

This was due to the fact that on GPU we were re-using pre-scaled memory allocations, that more often than not were over-sized, then making it impossible to process the whole TF in some cases.

This CPU code is now not useful, as the vertexer did not get updated with the refactorisation of the gpu part, so it is unused.

The current approach we have with the memory is much simpler and gauged (we would pseudo-allocate just the memory we need, as you know) not to require this strategy anymore for Run 3, even when the GPU vertexer will be restored.

I am fine to remove these.

@mconcas mconcas merged commit 0c0f2b9 into AliceO2Group:dev Jul 14, 2025
13 checks passed
@f3sch f3sch deleted the its/vtx_cleanup branch July 14, 2025 07:51
mhemmer-cern pushed a commit to mhemmer-cern/AliceO2 that referenced this pull request Sep 9, 2025
Signed-off-by: Felix Schlepper <felix.schlepper@cern.ch>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

2 participants