Skip to content

GitLab

Commit d79f7731 authored by Matthew Curtis's avatar Matthew Curtis
Browse files

Fix compilation errors generated by forthcoming Clang changes 

Without this change, the following diagnostic is generated:
  a template argument list is expected after a name prefixed by the template
  keyword [-Wmissing-template-arg-list-after-template-kw]

See C++17 spec [temp.names] p5.
parent 294cb823
Hide whitespace changes
Inline Side-by-side
Showing with 20 additions and 16 deletions
include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_v1_ab_scale.hpp
View file @ d79f7731
...@@ -308,7 +308,7 @@ struct BlockwiseGemmXdlops_pipeline_v1_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -308,7 +308,7 @@ struct BlockwiseGemmXdlops_pipeline_v1_ab_scale<BlockGemmPipelineScheduler::Intr
typename vector_type<ComputeDataType, typename vector_type<ComputeDataType,
xdlops_gemm.K1PerXdlops>::type; xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run( xdlops_gemm.template Run<>(
a_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
b_thread_vec.template AsType<mfma_input_type>(), b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); c_thread_buf_per_scale.GetVectorTypeReference(I0));
...@@ -390,9 +390,10 @@ struct BlockwiseGemmXdlops_pipeline_v1_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -390,9 +390,10 @@ struct BlockwiseGemmXdlops_pipeline_v1_ab_scale<BlockGemmPipelineScheduler::Intr
using mfma_input_type = using mfma_input_type =
typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type; typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run(a_thread_vec.template AsType<mfma_input_type>(), xdlops_gemm.template Run<>(
b_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0));
}); });
static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) { static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) {
constexpr index_t c_offset = constexpr index_t c_offset =
......
include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_v2_ab_scale.hpp
View file @ d79f7731
...@@ -350,7 +350,7 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -350,7 +350,7 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr
typename vector_type<ComputeDataType, typename vector_type<ComputeDataType,
xdlops_gemm.K1PerXdlops>::type; xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run( xdlops_gemm.template Run<>(
a_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
b_thread_vec.template AsType<mfma_input_type>(), b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); c_thread_buf_per_scale.GetVectorTypeReference(I0));
...@@ -443,7 +443,7 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -443,7 +443,7 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr
typename vector_type<ComputeDataType, typename vector_type<ComputeDataType,
xdlops_gemm.K1PerXdlops>::type; xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run( xdlops_gemm.template Run<>(
a_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
b_thread_vec.template AsType<mfma_input_type>(), b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); c_thread_buf_per_scale.GetVectorTypeReference(I0));
...@@ -518,9 +518,10 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -518,9 +518,10 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr
using mfma_input_type = using mfma_input_type =
typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type; typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run(a_thread_vec.template AsType<mfma_input_type>(), xdlops_gemm.template Run<>(
b_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0));
}); });
static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) { static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) {
constexpr index_t c_offset = constexpr index_t c_offset =
...@@ -575,9 +576,10 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -575,9 +576,10 @@ struct BlockwiseGemmXdlops_pipeline_v2_ab_scale<BlockGemmPipelineScheduler::Intr
using mfma_input_type = using mfma_input_type =
typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type; typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run(a_thread_vec.template AsType<mfma_input_type>(), xdlops_gemm.template Run<>(
b_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0));
}); });
static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) { static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) {
constexpr index_t c_offset = constexpr index_t c_offset =
......
include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_v3_ab_scale.hpp
View file @ d79f7731
...@@ -427,7 +427,7 @@ struct BlockwiseGemmXdlops_pipeline_v3_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -427,7 +427,7 @@ struct BlockwiseGemmXdlops_pipeline_v3_ab_scale<BlockGemmPipelineScheduler::Intr
typename vector_type<ComputeDataType, typename vector_type<ComputeDataType,
xdlops_gemm.K1PerXdlops>::type; xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run( xdlops_gemm.template Run<>(
a_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
b_thread_vec.template AsType<mfma_input_type>(), b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); c_thread_buf_per_scale.GetVectorTypeReference(I0));
...@@ -504,9 +504,10 @@ struct BlockwiseGemmXdlops_pipeline_v3_ab_scale<BlockGemmPipelineScheduler::Intr ...@@ -504,9 +504,10 @@ struct BlockwiseGemmXdlops_pipeline_v3_ab_scale<BlockGemmPipelineScheduler::Intr
using mfma_input_type = using mfma_input_type =
typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type; typename vector_type<ComputeDataType, xdlops_gemm.K1PerXdlops>::type;
xdlops_gemm.template Run(a_thread_vec.template AsType<mfma_input_type>(), xdlops_gemm.template Run<>(
b_thread_vec.template AsType<mfma_input_type>(), a_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0)); b_thread_vec.template AsType<mfma_input_type>(),
c_thread_buf_per_scale.GetVectorTypeReference(I0));
}); });
static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) { static_for<0, xdlops_gemm.GetRegSizePerXdlops(), 1>{}([&](auto t) {
constexpr index_t c_offset = constexpr index_t c_offset =
......
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