-
Notifications
You must be signed in to change notification settings - Fork 97
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
Create README.md for TensileLite #1466
Conversation
|
||
``` | ||
make co TENSILE_OUT=tensile-out ARCH="gfx942:xnack-" WAVEFRONTSIZE="-mwavefrontsize64" | ||
``` |
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.
For gfx942 example, it is better to remove WAVEFRONTSIZE="-mwavefrontsize64" since it is default setting
add another example for Wave32?
WAVE=32
example: make co TENSILE_OUT=tensile-out ARCH="gfx1201" WAVE=32
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.
Maybe not, sometimes these settings aren't automatically set correctly
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.
Which condition would make setting incorrect? We may list that in the README as well.
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.
@ellosel do you find out why sometimes the architecture cannot be detected correctly?
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.
ARCH should use gfx942 (not gfx942:xnack-)
@KKyang Should we be adding this to the ROCm documentation? This is inside the tensilelite directory, so out of our scope, but if you want me to review it, let me know. |
It's for hipBLASLt developers. Currently we have no plan to add it to the ROCm documentation. |
No description provided.