MCPcopy Create free account

hub / github.com/KnowingNothing/MatmulTutorial / functions

Functions541 in github.com/KnowingNothing/MatmulTutorial

⨍ Functions 541◇ Types & classes 125
↓ 2 callersFunctiontest
Test with arbitrary M, N dimensions (K aligned to BK, C stride aligned to 64)
examples/matmul/this-sm90/test_v10.py:68
↓ 2 callersFunctiontriton_perf_fn
(a_ptrs, b_ptrs, c_ptrs, sizes, lds, group_size)
examples/moe_group_gemm/triton/grouped_gemm.py:208
↓ 2 callersFunctiontriton_perf_fn
(A, B, C, scatter_idx, expert_idx, M, N, K, E, num_valid_tokens, BLOCK_M)
examples/moe_group_gemm/triton/blocked_grouped_gemm.py:104
↓ 2 callersMethodtry_wait
include/barrier.h:44
↓ 1 callersMethod__eq__
(self, other: Any)
util/simulator/data_math.py:150
↓ 1 callersMethod_get_grid_shape
( problem_blocks: dim3, cluster_shape: HyperCube, hw_info: KernelHardwareInfo,
cutlass.py/tile_scheduler.py:114
↓ 1 callersMethod_get_tiled_cta_shape_mnl
(problem_shape, cluster_shape, cta_m, cta_n)
cutlass.py/tile_scheduler.py:217
↓ 1 callersMethod_init_rope
(self)
examples/attention/triton/model.py:113
↓ 1 callersMethod_initialize
( self, problem_blocks: dim3, cluster_shape: HyperCube, hw_info: KernelHardwar
cutlass.py/tile_scheduler.py:49
↓ 1 callersFunction_max
include/common.h:77
↓ 1 callersFunction_min
include/common.h:82
↓ 1 callersMethodadvance
(self, number=1)
cutlass.py/tile_scheduler.py:414
↓ 1 callersMethodadvance_to_next_work
(self, advance_count=1)
cutlass.py/tile_scheduler.py:315
↓ 1 callersMethodapply
include/common.h:484
↓ 1 callersMethodapply
(self, offset)
cutlass.py/swizzle.py:20
↓ 1 callersMethodarrive_and_expect_tx
Performs an arrive operation + expected transaction bytes increment
include/barrier.h:313
↓ 1 callersFunctionbench
(M, N, K, warmup=10, iters=30)
examples/matmul/this-sm100/fp8_level3/bench_compare.py:25
↓ 1 callersFunctionbench
(lib, M, N, K, bn, warmup=10, iters=30)
examples/matmul/this-sm100/fp8_level3/bench_tile_sweep.py:31
↓ 1 callersFunctionbench
Benchmark for a given size
examples/matmul/this-sm90/test_v9.py:79
↓ 1 callersFunctionbench
Benchmark for a given size
examples/matmul/this-sm90/test_v8.py:84
↓ 1 callersFunctionbench_cublas
(M, N, K, warmup=10, iters=100)
examples/matmul/this-sm100/level8/bench_vs_deepgemm.py:37
↓ 1 callersFunctionbench_deepgemm
(M, N, K, warmup=10, iters=100)
examples/matmul/this-sm100/level8/bench_vs_deepgemm.py:52
↓ 1 callersFunctionbench_ours
(M, N, K, warmup=10, iters=100)
examples/matmul/this-sm100/level8/bench_vs_deepgemm.py:26
↓ 1 callersFunctionbenchmark
(lib, M, N, K, warmup=5, iters=20)
examples/matmul/this-sm100/fp8_level3/test.py:81
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level6/test.py:65
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level7/test.py:65
↓ 1 callersFunctionbenchmark
(lib, M, N, K, warmup=5, iters=20)
examples/matmul/this-sm100/fp8_level2/test.py:81
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level5/test.py:65
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level4/test.py:65
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level2/test.py:65
↓ 1 callersFunctionbenchmark
(lib, M, N, K, warmup=5, iters=20)
examples/matmul/this-sm100/fp8_level1/test.py:147
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level3/test.py:65
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level1/test.py:73
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm100/level8/test.py:68
↓ 1 callersFunctionbenchmark
(M=BENCH_SIZE, N=BENCH_SIZE, K=BENCH_SIZE, warmup=10, iters=100)
examples/matmul/this-sm90/test_v1.py:53
↓ 1 callersMethodblock_id_in_cluster
(self)
cutlass.py/hw_info.py:28
↓ 1 callersFunctionblock_rank_in_cluster
include/common.h:419
↓ 1 callersFunctionbuild
()
examples/matmul/cutlass/test_cutlass_perf.py:12
↓ 1 callersFunctionbuild
()
examples/matmul/cutlass-sm90/test_cutlass_perf.py:12
↓ 1 callersFunctionbuild
(M, N, K)
examples/matmul/mlir-gen/test_mlir_perf.py:12
↓ 1 callersFunctionbuild
(sm)
examples/matmul/cublas/test_cublas_perf.py:12
↓ 1 callersFunctionbuild
(version: int)
examples/matmul/this-sm90/test_this_perf.py:12
↓ 1 callersFunctionbuild
()
examples/matmul/this-sm80/test_this_perf.py:12
↓ 1 callersFunctioncalc_diff
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/test_this_perf.py:167
↓ 1 callersFunctioncanonical_warp_idx
Returns a warp index in the CTA. The threads in warp may not be convergent As it doesn't sync the warp, it faster and allows forward progress
include/common.h:496
↓ 1 callersFunctioncheck_correctness
(lib, M, N, K)
examples/matmul/this-sm100/fp8_level3/test.py:65
↓ 1 callersFunctioncheck_correctness
(lib, M=1024, N=1024, K=1024)
examples/matmul/this-sm100/fp8_level2/test.py:65
↓ 1 callersFunctioncheck_correctness
(lib, M=1024, N=1024, K=1024)
examples/matmul/this-sm100/fp8_level1/test.py:109
↓ 1 callersMethodcheck_error
(self, err, name)
examples/matmul/this-sm90/test_v2.py:52
↓ 1 callersMethodcheck_error
(self, err, name)
examples/matmul/this-sm90/test_v3.py:46
↓ 1 callersFunctioncompile
()
examples/matmul/this-sm90/test_v6.py:16
↓ 1 callersFunctioncompile
()
examples/matmul/this-sm90/test_v9.py:15
↓ 1 callersFunctioncompile
()
examples/matmul/this-sm90/test_v8.py:15
↓ 1 callersFunctioncompile
()
examples/matmul/this-sm90/test_v10.py:19
↓ 1 callersFunctioncompile
()
examples/matmul/this-sm90/test_v7.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level3/test.py:23
↓ 1 callersFunctioncompile_kernel
(block_n)
examples/matmul/this-sm100/fp8_level3/bench_tile_sweep.py:9
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level6/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level7/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level2/test.py:23
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level9/test.py:15
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level5/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level4/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level2/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level1/test.py:25
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level1/debug_test.py:9
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level1/debug_layout.py:9
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/fp8_level1/test_umma.py:9
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level3/test.py:18
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level1/test.py:24
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level8/test.py:20
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm100/level8/bench_vs_deepgemm.py:14
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm90/test_v5.py:10
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm90/test_v1.py:17
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm90/test_v2.py:93
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm90/test_v3.py:77
↓ 1 callersFunctioncompile_kernel
()
examples/matmul/this-sm90/test_v4.py:87
↓ 1 callersFunctioncompile_matmul_kernel
()
examples/matmul/this-sm100/test_this_perf.py:202
↓ 1 callersFunctioncomplex_elementwise_mul
( x_real: torch.Tensor, x_imag: torch.Tensor, y_real: torch.Tensor, y_imag: torch.Tensor, )
examples/attention/triton/model.py:23
↓ 1 callersFunctioncomplex_matmul_bf16
(ar, ai, br, bi)
examples/matmul/triton/complex_matmul.py:88
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level6/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level7/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level9/test.py:28
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level5/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level4/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level2/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level3/test.py:31
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level1/test.py:37
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm100/level8/test.py:34
↓ 1 callersFunctioncosine_similarity
(x: torch.Tensor, y: torch.Tensor)
examples/matmul/this-sm90/test_v1.py:25
↓ 1 callersFunctioncosine_similarity
(x, y)
examples/matmul/this-sm90/test_v2.py:125
↓ 1 callersFunctioncosine_similarity
(x, y)
examples/matmul/this-sm90/test_v3.py:115
↓ 1 callersFunctioncosine_similarity
(x, y)
examples/matmul/this-sm90/test_v4.py:119
↓ 1 callersFunctioncount_bytes
(*tensors)
examples/matmul/this-sm100/test_this_perf.py:174
↓ 1 callersFunctioncreate_data
(M, N, K)
examples/matmul/this-sm100/fp8_level3/bench_compare.py:16
↓ 1 callersFunctioncreate_data
(M, N, K, bn)
examples/matmul/this-sm100/fp8_level3/bench_tile_sweep.py:22
↓ 1 callersMethoddim_is_static
(self, key: int)
cutlass.py/tiling.py:122
↓ 1 callersMethoddivide
(self, dividend)
cutlass.py/fast_math.py:19
↓ 1 callersMethoddivmod
(self, dividend)
cutlass.py/fast_math.py:25
↓ 1 callersMethoddivmod
(self, x)
cutlass.py/fast_math.py:38
← previousnext →101–200 of 541, ranked by callers