[q_impls] Use int32 dtype for linear offset correction

nod-ai / sharktank

SHARK Inference Modeling and Serving

Apache License 2.0

7 stars 9 forks source link

Closed aviator19941 closed 1 month ago

aviator19941 commented 1 month ago

Use int32 dtype for linear offset correction to avoid warp reduction bit width error for int64 during compilation on MI300

Signed-off-by: aviator19941 avinash.sharma@amd.com