MCPcopy Create free account

hub / github.com/NVIDIA/cutlass / functions

Functions19,553 in github.com/NVIDIA/cutlass

⨍ Functions 19,553◇ Types & classes 10,397↳ Endpoints 26
↓ 4 callersMethodget_active_symbols
(self)
python/CuTeDSL/cutlass/base_dsl/ast_preprocessor.py:164
↓ 4 callersMethodget_block_idx
Obtains calling threadblock's linear threadblock index
include/cutlass/gemm/threadblock/threadblock_swizzle_streamk.h:703
↓ 4 callersMethodget_cluster_size
include/cutlass/gemm/kernel/tile_scheduler_params.h:512
↓ 4 callersMethodget_device_workspace_size
(self, arguments)
python/cutlass_cppgen/backend/gemm_operation.py:912
↓ 4 callersFunctionget_epilogue_info
(layer_info)
examples/44_multi_gemm_ir_and_codegen/ir_gen/helper.py:86
↓ 4 callersMethodget_first_block_idx
Obtains calling linear threadblock index of the first block to work on the given tile
include/cutlass/gemm/threadblock/threadblock_swizzle_streamk.h:782
↓ 4 callersMethodget_index
Get current pipeline stage index.
python/CuTeDSL/cutlass/cute/experimental/pipeline.py:923
↓ 4 callersMethodget_ir_location
Get python location information and generate MLIR location
python/CuTeDSL/cutlass/base_dsl/dsl.py:1342
↓ 4 callersMethodget_iter_extents
Obtains iteration extends for the given SK block index
include/cutlass/gemm/threadblock/threadblock_swizzle_streamk.h:755
↓ 4 callersMethodget_layoutC_TV
include/cute/atom/mma_atom.hpp:397
↓ 4 callersMethodget_mask
Gets the mask
include/cutlass/transform/threadblock/ell_predicated_tile_iterator.h:859
↓ 4 callersMethodget_masked_leading_count
Calculate the number of masked trips for the leading mask. This is used for blocks that need special handling due to masking.
examples/python/CuTeDSL/utils/fmha_helpers.py:704
↓ 4 callersMethodget_masked_trailing_count
Calculate the number of masked trips for the trailing mask. This is used for blocks that need special handling due to masking.
examples/python/CuTeDSL/utils/fmha_helpers.py:758
↓ 4 callersFunctionget_max_tmem_alloc_cols
Get the tensor memory capacity in columns for a given compute capability. Returns the maximum TMEM capacity in columns available for the specifie
python/CuTeDSL/cutlass/cute/arch/tmem.py:43
↓ 4 callersMethodget_object_cell_ptr
Get the cell from the tvm_ffi_object struct. Parameters ---------- obj : ir.Value The object ptr Returns
python/CuTeDSL/cutlass/base_dsl/tvm_ffi_builder/tvm_ffi_builder.py:276
↓ 4 callersMethodget_producer_load_callbacks
include/cutlass/epilogue/fusion/sm90_visitor_load_tma_warpspecialized.hpp:352
↓ 4 callersFunctionget_pruning_level_from_global_level
(global_level: int)
python/cutlass_library/sm100_utils.py:86
↓ 4 callersMethodget_sk_tile_idx
Obtains tile index for the given sk iteration
include/cutlass/gemm/threadblock/threadblock_swizzle_streamk.h:645
↓ 4 callersMethodget_smem_size
Get the shared memory size and alignment of current node
python/cutlass_cppgen/backend/evt/ir/node.py:142
↓ 4 callersMethodget_tensor_ref
( extent: "tuple[int]", device_ptr: cuda.CUdeviceptr, layout: LayoutType, )
python/cutlass_cppgen/backend/reduction_operation.py:125
↓ 4 callersMethodget_torch_tensor
(self, shape, dtype=None, fill=None)
test/python/cutlass/evt/utils/evt_testbed.py:128
↓ 4 callersMethodget_trip_count
examples/88_hopper_fmha/collective/fmha_fusion.hpp:135
↓ 4 callersMethodget_visitor_name
Get the visitor name
python/cutlass_cppgen/backend/evt/backend/emitter_base.py:65
↓ 4 callersMethodgflops_per_sec
Math throughput in units of GFLOP/s
tools/profiler/include/cutlass/profiler/performance_result.h:120
↓ 4 callersFunctiongmem_wait
Wait until we have at least one completed global fetch stage
include/cutlass/gemm/threadblock/mma_pipelined.h:274
↓ 4 callersMethodhelper_test_typing
(self, _cls, _obj, cls, expected: bool)
test/python/pycute/test_typing.py:45
↓ 4 callersMethodi16
Create an i16 constant with the given value.
python/CuTeDSL/cutlass/base_dsl/tvm_ffi_builder/mlir_builder.py:165
↓ 4 callersMethodinitialize
Initializes data structures
test/unit/gemm/device/gemm_testbed_3x.hpp:2750
↓ 4 callersMethodintersections
Compute the intersection of this set with multiple other sets. :param others: A list of sets to compute intersections with :type othe
python/CuTeDSL/cutlass/base_dsl/ast_preprocessor.py:86
↓ 4 callersMethodinverse
include/cutlass/layout/tensor_op_multiplicand_sm80.h:142
↓ 4 callersFunctionis_C_load_needed
include/cutlass/epilogue/fusion/sm90_visitor_load_tma_warpspecialized.hpp:1053
↓ 4 callersMethodis_complex
(self)
python/cutlass_library/rank_k_operation.py:77
↓ 4 callersMethodis_complex
(self)
python/cutlass_library/trmm_operation.py:76
↓ 4 callersMethodis_complex
(self)
python/cutlass_library/symm_operation.py:79
↓ 4 callersMethodis_complex
(self)
python/cutlass_library/rank_2k_operation.py:79
↓ 4 callersMethodis_complex
(self)
python/cutlass_cppgen/backend/gemm_operation.py:1580
↓ 4 callersMethodis_contributing
examples/88_hopper_fmha/collective/fmha_fusion.hpp:231
↓ 4 callersFunctionis_cute_algebra_type
(arg_spec: object)
python/CuTeDSL/cutlass/cutlass_dsl/cutlass.py:196
↓ 4 callersFunctionis_float_type
(ty: ir.Type)
python/CuTeDSL/cutlass/base_dsl/_mlir_helpers/arith.py:90
↓ 4 callersFunctionis_int_tuple
(a: object)
python/CuTeDSL/cutlass/cute/typing.py:706
↓ 4 callersFunctionis_nvf4
(kernel_schedule)
python/cutlass_library/generator.py:11391
↓ 4 callersMethodis_same_kind
(cls, other: Type)
python/CuTeDSL/cutlass/base_dsl/typing.py:417
↓ 4 callersMethodis_source_needed
include/cutlass/epilogue/collective/sm70_epilogue_vectorized_array.hpp:167
↓ 4 callersMethodis_source_needed
include/cutlass/epilogue/collective/default_epilogue_array.hpp:150
↓ 4 callersFunctionis_torch_available
()
python/cutlass_cppgen/utils/datatypes.py:147
↓ 4 callersMethodis_valid
include/cutlass/gemm/kernel/sm90_tile_scheduler_stream_k.hpp:107
↓ 4 callersMethodis_valid_mma_tiler_and_cluster_shape
Check if the mma tiler and cluster shape are valid. :return: True if the mma tiler and cluster shape are valid, False otherwise :rtyp
examples/python/CuTeDSL/cute/blackwell/kernel/dense_gemm/dense_gemm.py:1348
↓ 4 callersFunctionisfinite
include/cutlass/half.h:510
↓ 4 callersFunctionkernel
( tiled_mma: cute.TiledMma, tma_atom_a: cute.CopyAtom, mA_mkl: cute.Tensor, tma_atom_b: cute.C
examples/python/CuTeDSL/cute/blackwell/tutorial/tutorial_gemm/fp16_gemm_5.py:91
↓ 4 callersFunctionl
(t)
python/docs/_static/clipboard.min.js:7
↓ 4 callersFunctionl
(t,e)
python/docs/_static/scripts/furo.js:2
↓ 4 callersFunctionl2_flush
Best-effort L2 flush by touching a large temporary tensor.
examples/python/CuTeDSL/cute/blackwell/kernel/moe/torch_grouped_mm.py:1466
↓ 4 callersFunctionld_without_cache
include/cutlass/experimental/distributed/kernel/detail.hpp:48
↓ 4 callersFunctionlibrary_type
(inp)
python/cutlass_cppgen/utils/datatypes.py:239
↓ 4 callersFunctionload
Loads a fragment from memory at the location pointed to by the iterator.
examples/41_fused_multi_head_attention/iterators/warp_iterator_from_smem.h:252
↓ 4 callersFunctionload
Loads a fragment from memory at the location pointed to by the iterator.
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:378
↓ 4 callersMethodload
Loads a fragment from the referenced part of the accumulator tile
examples/44_multi_gemm_ir_and_codegen/fixed_impl/epilogue/warp/fused_bias_act_fragment_iterator_tensor_op.h:146
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
include/cutlass/gemm/warp/mma_complex_tensor_op_tile_iterator_sm80.h:965
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm70.h:1306
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:3348
↓ 4 callersMethodload_auxiliary
include/cutlass/gemm/collective/sm90_mma_array_tma_gmma_ss_warpspecialized_fp8_blockwise_scaling.hpp:587
↓ 4 callersMethodload_sf_init
include/cutlass/gemm/collective/sm100_mma_warpspecialized_blockwise_scaling.hpp:707
↓ 4 callersFunctionload_with_byte_offset
Loads a fragment from memory with additional logical offset
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sparse.h:283
↓ 4 callersFunctionload_with_pointer_offset
Loads a fragment from memory
include/cutlass/epilogue/threadblock/shared_load_iterator.h:163
↓ 4 callersFunctionload_with_pointer_offset
Load
include/cutlass/epilogue/warp/tile_iterator_simt.h:227
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
include/cutlass/gemm/warp/mma_complex_tensor_op_tile_iterator_sm80.h:971
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm70.h:1312
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:3354
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory
include/cutlass/transform/threadblock/regular_tile_iterator_tensor_op.h:315
↓ 4 callersFunctionlogical_divide
( target: Layout, tiler: Tiler, *, loc: Optional[ir.Location] = None, ip: Optional[ir.Inse
python/CuTeDSL/cutlass/cute/core.py:4748
↓ 4 callersFunctionmake_dims_and_alignments_triple
(dim: int, bit_per_element_A: int, bit_per_element_B: int, bit_per_element_C: int)
python/cutlass_library/generator.py:10875
↓ 4 callersMethodmake_fragment
( self, tensor: Tensor, *, loc: Optional[ir.Location] = None, ip: Opti
python/CuTeDSL/cutlass/cute/nvgpu/common.py:741
↓ 4 callersFunctionmake_gather_tensor
examples/common/gather_tensor.hpp:155
↓ 4 callersFunctionmake_math_instruction_w_output
(data_types: Tuple[DataType, DataType, DataType, DataType], instruction_shape: Tuple
python/cutlass_library/generator.py:10881
↓ 4 callersFunctionmake_smem_layout_atom
Makes a SMEM layout Atom. This function creates a composed layout in unit of elements consistent with the requested layout Atom kind and
python/CuTeDSL/cutlass/cute/nvgpu/tcgen05/helpers.py:63
↓ 4 callersFunctionmake_sparse_math_instructions
(math_instructions)
python/cutlass_library/sm90_utils.py:267
↓ 4 callersFunctionmake_tensors
(mn, k, dtype)
examples/python/CuTeDSL/cute/blackwell/tutorial/tutorial_gemm/fp16_gemm_6.py:906
↓ 4 callersFunctionmake_tiled_copy_C_atom
include/cute/atom/copy_atom.hpp:451
↓ 4 callersFunctionmake_tiled_mma
include/cute/atom/mma_atom.hpp:529
↓ 4 callersFunctionmake_tma_copy_C_sm90
include/cute/atom/copy_traits_sm90_tma.hpp:1596
↓ 4 callersFunctionmake_tma_copy_im2col
include/cute/atom/copy_traits_sm90_im2col.hpp:720
↓ 4 callersFunctionmake_tmem_warp_partitioner
include/cute/atom/copy_traits_sm100.hpp:326
↓ 4 callersFunctionmake_transpose_operand_b
include/cutlass/transform/collective/sm90_wgmma_transpose.hpp:713
↓ 4 callersFunctionmax_alignment
include/cute/container/array_subbyte.hpp:356
↓ 4 callersFunctionmin
include/cute/numeric/math.hpp:57
↓ 4 callersFunctionmma
examples/112_blackwell_ssd/reference/reference_ssd.hpp:51
↓ 4 callersMethodmul
Create a multiplication operation between two values.
python/CuTeDSL/cutlass/base_dsl/tvm_ffi_builder/mlir_builder.py:173
↓ 4 callersMethodnode_metas_topological_order
Get the node metas in topological order :return: list[NodeBase]
python/cutlass_cppgen/backend/evt/ir/dag_ir.py:216
↓ 4 callersFunctionnormalize_skip_wait_token
Normalizes a skip wait token to an ir.Value.
python/CuTeDSL/cutlass/cute/experimental/core.py:558
↓ 4 callersMethodpack_arguments
Pack the arguments to pass them through a @cute.jit/@cute.kernel call. If some arguments are provided, pack them, otherwise just use
examples/python/CuTeDSL/cute/blackwell/efc/common_efc.py:202
↓ 4 callersFunctionparams_in_context
(func)
test/utils/test_sharding.py:229
↓ 4 callersMethodpipeline_state_clone_and_advance
Clones the pipeline state and advances it.
examples/python/CuTeDSL/cute/blackwell/kernel/mixed_input_gemm/grouped_mixed_input_gemm_acc_scale.py:1852
↓ 4 callersMethodpointerize_if_grouped
(operation, layout)
python/cutlass_library/gemm_operation.py:912
↓ 4 callersFunctionprint
include/cute/container/tuple.hpp:669
↓ 4 callersMethodprint_warning
(self, message: str)
python/CuTeDSL/cutlass/base_dsl/dsl.py:716
↓ 4 callersMethodprocedural_name
The full procedural name indicates architecture, extended name, tile size, and layout.
python/cutlass_cppgen/backend/gemm_operation.py:1705
↓ 4 callersFunctionproducer_acquire
include/cutlass/pipeline/sm90_pipeline.hpp:530
↓ 4 callersMethodproducer_acquire
include/cutlass/pipeline/sm100_pipeline.hpp:1164
↓ 4 callersMethodproducer_commit
include/cutlass/pipeline/sm100_pipeline.hpp:1168
↓ 4 callersMethodproduct
Returns the product of all elements
include/cutlass/coord.h:319
← previousnext →1,501–1,600 of 19,553, ranked by callers