Commit 5b5a42a3 authored by wangshaojie6's avatar wangshaojie6
Browse files

reuse 256x256 in example

parent 099c470e
...@@ -44,22 +44,22 @@ using DeviceConvBwdWeightInstance = ck::tensor_operation::device:: ...@@ -44,22 +44,22 @@ using DeviceConvBwdWeightInstance = ck::tensor_operation::device::
WeiElementOp, // WeiElementwiseOperation WeiElementOp, // WeiElementwiseOperation
OutElementOp, // OutElementwiseOperation OutElementOp, // OutElementwiseOperation
256, // BlockSize 256, // BlockSize
64, // MPerBlock 256, // MPerBlock
128, // NPerBlock 256, // NPerBlock
8, // K0PerBlock 4, // K0PerBlock
8, // K1 8, // K1
32, // MPerXdl 32, // MPerXdl
32, // NPerXdl 32, // NPerXdl
1, // MXdlPerWave 4, // MXdlPerWave
2, // NXdlPerWave 4, // NXdlPerWave
S<1, 8, 8, 4>, // ABlockTransferThreadClusterLengths_K0_M_K1 S<1, 4, 32, 2>, // ABlockTransferThreadClusterLengths_K0_M_K1
S<0, 3, 1, 2>, // ABlockTransferThreadClusterArrangeOrder S<0, 3, 1, 2>, // ABlockTransferThreadClusterArrangeOrder
S<0, 2, 1, 3>, // ABlockTransferSrcAccessOrder S<0, 2, 1, 3>, // ABlockTransferSrcAccessOrder
2, // ABlockTransferSrcVectorDim 2, // ABlockTransferSrcVectorDim
8, // ABlockTransferSrcScalarPerVector 8, // ABlockTransferSrcScalarPerVector
2, // ABlockTransferDstScalarPerVector_K1 4, // ABlockTransferDstScalarPerVector_K1
true, // ABlockLdsAddExtraM true, // ABlockLdsAddExtraM
S<1, 8, 16, 2>, // BBlockTransferThreadClusterLengths_K0_N_K1 S<1, 4, 32, 2>, // BBlockTransferThreadClusterLengths_K0_N_K1
S<0, 3, 1, 2>, // BBlockTransferThreadClusterArrangeOrder S<0, 3, 1, 2>, // BBlockTransferThreadClusterArrangeOrder
S<0, 2, 1, 3>, // BBlockTransferSrcAccessOrder S<0, 2, 1, 3>, // BBlockTransferSrcAccessOrder
2, // BBlockTransferSrcVectorDim 2, // BBlockTransferSrcVectorDim
......
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