This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 6bc94d0050 [FFI][REFACTOR] Establish ffi.Module in python (#18214)
add 1163aa0bcd [FFI][REFACTOR] Establish Stream Context in ffi (#18216)
add 216e9e99c1 [FFI] AudoDLPack compatible with torch stream context
(#18217)
No new revisions were added by this update.
Summary of changes:
ffi/CMakeLists.txt | 1 +
ffi/include/tvm/ffi/extra/c_env_api.h | 33 ++++++++
ffi/scripts/benchmark_dlpack.py | 70 +++++++++++++++-
ffi/src/ffi/extra/stream_context.cc | 81 +++++++++++++++++++
include/tvm/runtime/device_api.h | 2 +-
python/tvm/contrib/cutlass/attention_operation.py | 12 +--
python/tvm/contrib/cutlass/conv2d_operation.py | 3 +-
python/tvm/contrib/cutlass/gemm_operation.py | 13 +--
python/tvm/contrib/cutlass/gen_tensor_op.py | 2 +-
python/tvm/contrib/cutlass/layer_norm_operation.py | 3 +-
python/tvm/contrib/cutlass/rms_norm_operation.py | 3 +-
python/tvm/ffi/cython/base.pxi | 8 ++
python/tvm/ffi/cython/function.pxi | 92 ++++++++++++++++++++--
src/contrib/msc/plugin/tvm_codegen.cc | 4 +-
src/runtime/contrib/cublas/cublas.cc | 11 +--
src/runtime/contrib/cublas/cublas_json_runtime.cc | 16 ++--
src/runtime/contrib/cublas/cublas_utils.cc | 12 ++-
src/runtime/contrib/cublas/cublas_utils.h | 4 +-
src/runtime/contrib/cudnn/conv_backward.cc | 12 ++-
src/runtime/contrib/cudnn/conv_forward.cc | 8 +-
.../contrib/cudnn/cudnn_frontend/attention.cc | 8 +-
src/runtime/contrib/cudnn/cudnn_json_runtime.cc | 10 ++-
src/runtime/contrib/cudnn/cudnn_utils.cc | 16 ++--
src/runtime/contrib/cudnn/cudnn_utils.h | 2 +-
src/runtime/contrib/cudnn/softmax.cc | 5 +-
src/runtime/contrib/cutlass/fp16_group_gemm.cuh | 2 +-
src/runtime/contrib/cutlass/fp8_gemm.cu | 7 +-
src/runtime/contrib/cutlass/fp8_group_gemm_sm90.cu | 5 +-
.../contrib/cutlass/fp8_groupwise_scaled_gemm.cuh | 9 +--
.../fp8_groupwise_scaled_group_gemm_sm100.cu | 4 +-
src/runtime/contrib/hipblas/hipblas.cc | 4 +-
.../contrib/hipblas/hipblas_json_runtime.cc | 11 ++-
src/runtime/contrib/hipblas/hipblas_utils.cc | 9 ++-
src/runtime/contrib/miopen/conv_forward.cc | 8 +-
src/runtime/contrib/miopen/miopen_utils.cc | 11 ++-
src/runtime/contrib/miopen/softmax.cc | 2 +-
src/runtime/contrib/msc/tensorrt_runtime.cc | 4 +-
src/runtime/cuda/cuda_common.h | 4 -
src/runtime/cuda/cuda_device_api.cc | 30 ++++---
src/runtime/cuda/cuda_module.cc | 3 +-
src/runtime/cuda/l2_cache_flush.cc | 5 +-
src/runtime/device_api.cc | 8 +-
src/runtime/metal/metal_common.h | 2 -
src/runtime/metal/metal_device_api.mm | 11 ---
src/runtime/rocm/rocm_device_api.cc | 28 ++++---
src/runtime/rocm/rocm_module.cc | 3 +-
src/runtime/vm/cuda/cuda_graph_builtin.cc | 20 +++--
src/runtime/vulkan/vulkan_device_api.cc | 6 --
src/runtime/vulkan/vulkan_device_api.h | 2 -
web/emcc/webgpu_runtime.cc | 4 -
50 files changed, 468 insertions(+), 165 deletions(-)
create mode 100644 ffi/src/ffi/extra/stream_context.cc
