Skip to content

[Common] Enable arch support for Rubin #3275

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

Open
denera wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[Common] Enable arch support for Rubin #3275

Show file tree
Hide file tree
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
3 changes: 3 additions & 0 deletions transformer_engine/common/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,9 @@ if(NOT arch_100_index EQUAL -1)
if(CUDAToolkit_VERSION VERSION_GREATER_EQUAL 12.9)
list(APPEND NVTE_SPECIFIC_ARCHS "103a")
endif()
if(CUDAToolkit_VERSION VERSION_GREATER_EQUAL 13.4)
list(APPEND NVTE_SPECIFIC_ARCHS "107a")
endif()
endif()

# Check for architecture 101 (if we see this we are in toolkit <= 12.9)
Expand Down
2 changes: 0 additions & 2 deletions ...common/hadamard_transform/graph_safe_group_row_cast_col_hadamard_transform_cast_fusion.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,6 @@
#include <transformer_engine/hadamard_transform.h>

#include <cuda/barrier>
#include <cute/algorithm/gemm.hpp>
#include <cute/arch/cluster_sm90.hpp>
#include <cute/tensor.hpp>

#include "common/common.h"
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/common/hadamard_transform/group_hadamard_transform_cast_fusion.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,6 @@
#include <transformer_engine/hadamard_transform.h>

#include <cuda/barrier>
#include <cute/algorithm/gemm.hpp>
#include <cute/arch/cluster_sm90.hpp>
#include <cute/tensor.hpp>

#include "common/common.h"
Expand Down
2 changes: 0 additions & 2 deletions ...mer_engine/common/hadamard_transform/group_row_cast_col_hadamard_transform_cast_fusion.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,6 @@
#include <transformer_engine/hadamard_transform.h>

#include <cuda/barrier>
#include <cute/algorithm/gemm.hpp>
#include <cute/arch/cluster_sm90.hpp>
#include <cute/tensor.hpp>

#include "common/common.h"
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/common/hadamard_transform/hadamard_transform_cast_fusion.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,6 @@
#include <transformer_engine/hadamard_transform.h>

#include <cuda/barrier>
#include <cute/algorithm/gemm.hpp>
#include <cute/arch/cluster_sm90.hpp>
#include <cute/tensor.hpp>

#include "common/common.h"
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/common/hadamard_transform/row_cast_col_hadamard_transform_cast_fusion.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,6 @@
#include <transformer_engine/hadamard_transform.h>

#include <cuda/barrier>
#include <cute/algorithm/gemm.hpp>
#include <cute/arch/cluster_sm90.hpp>
#include <cute/tensor.hpp>

#include "common/common.h"
Expand Down
2 changes: 1 addition & 1 deletion transformer_engine/common/util/ptx.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,7 +124,7 @@ constexpr bool is_supported_arch() {
NVTE_CUDA_ARCH_MATCHES(ptx::FamilySpecific<100>, ptx::FamilySpecific<110>, \
ptx::FamilySpecific<120>)
#define ARCH_HAS_STOCHASTIC_ROUNDING \
NVTE_CUDA_ARCH_MATCHES(ptx::ArchSpecific<100>, ptx::ArchSpecific<103>)
NVTE_CUDA_ARCH_MATCHES(ptx::ArchSpecific<100>, ptx::ArchSpecific<103>, ptx::ArchSpecific<107>)

// https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#parallel-synchronization-and-communication-instructions-mbarrier-init
__device__ __forceinline__ void mbarrier_init(uint64_t *mbar, const uint32_t count) {
Expand Down
Loading