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

[SYCL][CUDA][DOC] State min. CUDA ver. for sm_75 #5642

Merged
merged 1 commit into from Feb 24, 2022

Conversation

pgorlani
Copy link
Contributor

This patch adds a note on the Get Started Guide regarding the minimum CUDA toolkit version required for fully utilize Turing devices (sm_75).

CUDA toolkit version 11.0 introduces PTX7.0. This version supports for the first time the Ampere architecture (sm_80), however some instructions introduced by PTX7.0 (e.g. approximated tanh (#5265) and ex2 for halfs) can be executed also by Turing devices (sm_75), if CUDA 11.0 (or above) is installed.

Compilation on Turing devices is possible also using CUDA 10.2 (the actual version reported as tested), however if one these PTX7.0 instruction is used, it will generate an error.

@pgorlani pgorlani requested a review from a team as a code owner February 23, 2022 14:11
@pvchupin pvchupin merged commit ae94db4 into intel:sycl Feb 24, 2022
codeplay-sycl pushed a commit to codeplaysoftware/intel-llvm-mirror that referenced this pull request Feb 25, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants