Fixed encoding options for tensormap_encode_tiled. More...
#include <tensormap.h>
Inheritance diagram for tvm::backend::cuda::TensorMapEncodeTiledAttr:
Collaboration diagram for tvm::backend::cuda::TensorMapEncodeTiledAttr:
Public Member Functions | |
| TVM_FFI_DECLARE_OBJECT_INFO_FINAL ("tirx.cuda.TensorMapEncodeTiledAttr", TensorMapEncodeTiledAttr, AttrsNode) | |
 Public Member Functions inherited from tvm::AttrsNode | |
| TVM_FFI_DECLARE_OBJECT_INFO ("ir.Attrs", AttrsNode, ffi::Object) | |
Static Public Member Functions | |
| static void | RegisterReflection () |
Public Attributes | |
| DLDataType | descriptor_dtype |
| int64_t | rank |
| int64_t | interleave |
| int64_t | swizzle |
| int64_t | l2_promotion |
| int64_t | oob_fill |
| int64_t | force_cu_dtype |
Additional Inherited Members | |
| Static Public Attributes inherited from tvm::AttrsNode | |
| static constexpr TVMFFISEqHashKind | _type_s_eq_hash_kind = kTVMFFISEqHashKindTreeNode |
Detailed Description
Fixed encoding options for tensormap_encode_tiled.
Member Function Documentation
◆ RegisterReflection()
|
static |
◆ TVM_FFI_DECLARE_OBJECT_INFO_FINAL()
| tvm::backend::cuda::TensorMapEncodeTiledAttr::TVM_FFI_DECLARE_OBJECT_INFO_FINAL | ( | "tirx.cuda.TensorMapEncodeTiledAttr" | , |
| TensorMapEncodeTiledAttr | , | ||
| AttrsNode | |||
| ) |
Member Data Documentation
◆ descriptor_dtype
| DLDataType tvm::backend::cuda::TensorMapEncodeTiledAttr::descriptor_dtype |
◆ force_cu_dtype
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::force_cu_dtype |
◆ interleave
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::interleave |
◆ l2_promotion
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::l2_promotion |
◆ oob_fill
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::oob_fill |
◆ rank
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::rank |
◆ swizzle
| int64_t tvm::backend::cuda::TensorMapEncodeTiledAttr::swizzle |
The documentation for this struct was generated from the following file:
- include/tvm/backend/cuda/op/tensormap.h
