Add BlockFmhaFwdSplitKVPipelineQRKSVSAsync (just copy files)

include/ck_tile/ops/fmha.hpp

@@ -30,6 +30,8 @@
 #include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_combine_pipeline.hpp"
 #include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_combine_pipeline_default_policy.hpp"
 #include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_pipeline_qr_ks_vs.hpp"
+#include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_pipeline_qr_ks_vs_async.hpp"
+#include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_pipeline_qr_ks_vs_async_default_policy.hpp"
 #include "ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_pipeline_qr_ks_vs_default_policy.hpp"
 #include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_enum.hpp"
 #include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_problem.hpp"

include/ck_tile/ops/fmha/pipeline/block_fmha_fwd_splitkv_pipeline_qr_ks_vs_async_default_policy.hpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#pragma once
+
+#include "ck_tile/core.hpp"
+#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qx_ks_vs_custom_policy.hpp"
+
+namespace ck_tile {
+
+// This pipeline is q in register; kv all located in LDS
+using BlockFmhaFwdSplitKVPipelineQRKSVSAsyncDefaultPolicy =
+    BlockFmhaPipelineQXKSVSCustomPolicy</* QLoadOnce = */ true,
+                                        /* AsyncCopyK = */ true,
+                                        /* AsyncCopyV = */ false,
+                                        /* NumPrefetchK = */ 3,
+                                        /* NumPrefetchV = */ 3>;
+
+} // namespace ck_tile