tilelang.intrinsics.mma_sm70_layout

Functions

shared_16x4_to_mma_a_32x4_layout(row, col, rep)
shared_4x16_to_mma_b_32x4_layout(row, col, rep)
shared_16x4_to_mma_b_32x4_layout_trans(row, col, rep)
mma_32x8_to_shared_16x16_layout_fp32(thread_id, local_id)
mma_32x8_to_shared_16x16_layout_fp16(thread_id, local_id)
mma_load_a_32x4_to_shared_16x4_layout(thread_id, local_id)
mma_load_b_32x4_to_shared_16x4_layout_trans(thread_id, ...)
mma_load_b_32x4_to_shared_4x16_layout(thread_id, local_id)

Module Contents

tilelang.intrinsics.mma_sm70_layout.shared_16x4_to_mma_a_32x4_layout(row, col, rep)

tilelang.intrinsics.mma_sm70_layout.shared_4x16_to_mma_b_32x4_layout(row, col, rep)

tilelang.intrinsics.mma_sm70_layout.shared_16x4_to_mma_b_32x4_layout_trans(row, col, rep)

tilelang.intrinsics.mma_sm70_layout.mma_32x8_to_shared_16x16_layout_fp32(thread_id, local_id)

tilelang.intrinsics.mma_sm70_layout.mma_32x8_to_shared_16x16_layout_fp16(thread_id, local_id)

tilelang.intrinsics.mma_sm70_layout.mma_load_a_32x4_to_shared_16x4_layout(thread_id, local_id)

tilelang.intrinsics.mma_sm70_layout.mma_load_b_32x4_to_shared_16x4_layout_trans(thread_id, local_id)

tilelang.intrinsics.mma_sm70_layout.mma_load_b_32x4_to_shared_4x16_layout(thread_id, local_id)