Skip to content
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

[BACKPORT]: Fix ptx usage to account for PTX ISA availability (#1359) #1421

Merged
merged 7 commits into from
Feb 26, 2024

Conversation

miscco
Copy link
Collaborator

@miscco miscco commented Feb 21, 2024

Currently we only guard those instructions based on the available architecture. However, it is also valid to compile with an old toolkit for a new machine.

Consequently we need to strengthen our checks against available PTX ISA

@miscco miscco requested review from a team as code owners February 21, 2024 09:50
@miscco miscco added libcu++ For all items related to libcu++ bug: functional 2.3.2 Targeted for 2.3.2 patch release labels Feb 21, 2024
Copy link
Contributor

@ahendriksen ahendriksen left a comment

Choose a reason for hiding this comment

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

LGTM

Currently we only guard those instructions based on the available architecture. However, it is also valid to compile with an old toolkit for a new machine.

Consequently we need to strengthen our checks against available PTX ISA
VLAs are a compiler extension and are correctly errored out by some compilers. As we always know the exact size of the array anyway just swtich to a `cuda::std::array`

Fixes nvbug4476664
Authored-by: Allard Hendriksen <ahendriksen@nvidia.com>
@jrhemstad jrhemstad merged commit df93a01 into NVIDIA:branch/2.3.x Feb 26, 2024
538 checks passed
@miscco miscco deleted the backport_1359 branch March 12, 2024 15:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
2.3.2 Targeted for 2.3.2 patch release bug: functional libcu++ For all items related to libcu++
Projects
Archived in project
Development

Successfully merging this pull request may close these issues.

4 participants