Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[inference] fix cutlass sm75 build bug #60583

Merged
merged 1 commit into from
Jan 5, 2024
Merged

[inference] fix cutlass sm75 build bug #60583

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 10 additions & 4 deletions ...phi/kernels/fusion/cutlass/cutlass_extensions/gemm/threadblock/default_dq_mma_pipelined.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,7 +78,7 @@ template <
/// Instruction-level tile size (concept: GemmShape)
typename InstructionShape,
/// Operation performed by GEMM
typename Operator>
typename Operator_>
struct DqMma<
ElementA,
LayoutA,
Expand All @@ -97,9 +97,12 @@ struct DqMma<
WarpShape,
InstructionShape,
2,
Operator,
Operator_,
SharedMemoryClearOption::kNone,
typename platform::enable_if<(ArchTag::kMinComputeCapability < 80)>::type> {
using OperatorInfo = arch::DetagOperator<Operator_>;
using Operator = typename OperatorInfo::Operator;

static_assert(platform::is_same<ElementA, half_t>::value ||
platform::is_same<ElementA, bfloat16_t>::value,
"Element A must be fp16 or bf16");
Expand Down Expand Up @@ -227,7 +230,7 @@ template <
/// Instruction-level tile size (concept: GemmShape)
typename InstructionShape,
/// Operation performed by GEMM
typename Operator,
typename Operator_,
///
int RowsPerTile,
///
Expand All @@ -250,9 +253,12 @@ struct DqMma<
WarpShape,
InstructionShape,
2,
Operator,
Operator_,
SharedMemoryClearOption::kNone,
typename platform::enable_if<(ArchTag::kMinComputeCapability < 80)>::type> {
using OperatorInfo = arch::DetagOperator<Operator_>;
using Operator = typename OperatorInfo::Operator;

static_assert(platform::is_same<ElementA, half_t>::value ||
platform::is_same<ElementA, bfloat16_t>::value,
"Element A must be fp16 or bf16");
Expand Down