From ca47a6cfe2f42a777968ad3f69ad66d21d0a6569 Mon Sep 17 00:00:00 2001
From: rocking5566 <ChunYu.Lai@amd.com>
Date: Tue, 25 Jan 2022 12:44:13 +0800
Subject: [PATCH] Do not hardcode the function parameter, use template instead.
 (#72)

* Do not hardcode the function parameter, use template instead.

* [What] Remove AThreadTransferSrcResetCoordinateAfterRun and BThreadTransferSrcResetCoordinateAfterRun in host API
[Why] "C_Shuffle" version is supposed to be similar to the vanilla one

* Fix typo
Let DeviceGemmXdl_C_Shuffle use kernel_gemm_xdlops_v3r1
---
 .../include/device_gemm_xdl_c_shuffle.hpp      | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/device_operation/include/device_gemm_xdl_c_shuffle.hpp b/device_operation/include/device_gemm_xdl_c_shuffle.hpp
index 2c70e955d7..da19b5ec4f 100644
--- a/device_operation/include/device_gemm_xdl_c_shuffle.hpp
+++ b/device_operation/include/device_gemm_xdl_c_shuffle.hpp
@@ -156,20 +156,20 @@ struct DeviceGemmXdl_C_Shuffle
         MXdlPerWave,
         NXdlPerWave,
         ABlockTransferThreadClusterLengths_K0_M_K1,
-        Sequence<1, 0, 2>, // ABlockTransferThreadClusterArrangeOrder,
-        Sequence<1, 0, 2>, // ABlockTransferSrcAccessOrder,
-        2,                 // ABlockTransferSrcVectorDim,
+        ABlockTransferThreadClusterArrangeOrder,
+        ABlockTransferSrcAccessOrder,
+        ABlockTransferSrcVectorDim,
         ABlockTransferSrcScalarPerVector,
         ABlockTransferDstScalarPerVector_K1,
-        false, // AThreadTransferSrcResetCoordinateAfterRun,
+        false,
         ABlockLdsAddExtraM,
         BBlockTransferThreadClusterLengths_K0_N_K1,
-        Sequence<1, 0, 2>, // BBlockTransferThreadClusterArrangeOrder,
-        Sequence<1, 0, 2>, // BBlockTransferSrcAccessOrder,
-        2,                 // BBlockTransferSrcVectorDim,
+        BBlockTransferThreadClusterArrangeOrder,
+        BBlockTransferSrcAccessOrder,
+        BBlockTransferSrcVectorDim,
         BBlockTransferSrcScalarPerVector,
         BBlockTransferDstScalarPerVector_K1,
-        false, // BThreadTransferSrcResetCoordinateAfterRun,
+        false,
         BBlockLdsAddExtraN,
         CShuffleMXdlPerWavePerShuffle,
         CShuffleNXdlPerWavePerShuffle,
@@ -317,7 +317,7 @@ struct DeviceGemmXdl_C_Shuffle
             }
             else
             {
-                const auto kernel = kernel_gemm_xdlops_v2r3<
+                const auto kernel = kernel_gemm_xdlops_v3r1<
                     GridwiseGemm,
                     ADataType, // TODO: distiguish A/B datatype
                     CDataType,