Mojo module
mha_rdna
RDNA-specific MHA kernel configurations and entry points.
This module provides attention computation configurations optimized for AMD RDNA consumer GPUs (Radeon RX 7000/8000 series, gfx11xx/gfx12xx).
Key features:
- Wave32 execution model
- 16x16x16 WMMA shape (only supported shape for RDNA)
- Optimized shared memory management with K/P reuse
Structs
-
MHAAttentionConfigRDNA: RDNA-specific attention configuration for Wave32 WMMA.
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!