Commit 3e411ef0 authored by Qianfeng Zhang's avatar Qianfeng Zhang
Browse files

Use MakeQDramTileDistribution for q_dram_window

parent ceea2722
...@@ -175,10 +175,11 @@ struct BlockFmhaPipelineQRKSVSAsync ...@@ -175,10 +175,11 @@ struct BlockFmhaPipelineQRKSVSAsync
static_assert(NumKLdsBuffers >= 2); static_assert(NumKLdsBuffers >= 2);
static_assert(NumVLdsBuffers >= 2); static_assert(NumVLdsBuffers >= 2);
auto q_dram_window = make_tile_window(q_dram_block_window_tmp.get_bottom_tensor_view(), auto q_dram_window =
make_tile_window(q_dram_block_window_tmp.get_bottom_tensor_view(),
q_dram_block_window_tmp.get_window_lengths(), q_dram_block_window_tmp.get_window_lengths(),
q_dram_block_window_tmp.get_window_origin(), q_dram_block_window_tmp.get_window_origin(),
Policy::template MakeQRegTileDistribution<Problem>()); Policy::template MakeQDramTileDistribution<Problem>());
auto original_q = load_tile(q_dram_window); auto original_q = load_tile(q_dram_window);
__builtin_amdgcn_sched_barrier(0); __builtin_amdgcn_sched_barrier(0);
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment