You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Could you share a sample ptx file which has mma in it so we can see how it is currently handled? Feel free to include it in src/ptx_files/ as a commit, or share it here in a comment.
Nvidia has reserved 132 instruction keywords, including mma. I made one attempt to rework the operation field with a fat enum and a lot of hard-coding. Perhaps this is an avenue for future consideration.
Here is the output parse_body_example.txt of cargo test parse_body_example -- --nocapture.
I made ptx-parser in service of another WIP crate ptx-viz. Features will trickle in as I need them for viz. Feel free to suggest others.
Hi, thanks for working on this! 👍
I could not find mma instruction support in the codebase. Maybe my search is too quick, I'm simply grepping for
grep mma * -R -i
.Is mma really missing, if yes - do you plan to add it? A README note about feature completion table would be really helpful!
The text was updated successfully, but these errors were encountered: