MCPcopy Create free account

hub / github.com/tile-ai/tilelang / functions

Functions7,254 in github.com/tile-ai/tilelang

↓ 1 callersFunctionkernel_all_sync
All lanes always pass predicate 1 → all_sync should return non-zero.
testing/python/language/test_tilelang_language_warp_vote.py:60
↓ 1 callersFunctionkernel_alloc_var_eq
(data)
testing/python/language/test_tilelang_language_alloc_var_cmp.py:10
↓ 1 callersFunctionkernel_alloc_var_ne
(data)
testing/python/language/test_tilelang_language_alloc_var_cmp.py:27
↓ 1 callersFunctionkernel_any_sync
Lane 0 has a non-zero predicate; any_sync should return non-zero for all lanes.
testing/python/language/test_tilelang_language_warp_vote.py:28
↓ 1 callersFunctionkernel_ballot
All lanes pass predicate 1 → lower 32 bits of ballot must all be set.
testing/python/language/test_tilelang_language_warp_vote.py:125
↓ 1 callersFunctionkernel_ballot_sync
Only lane 0 has a non-zero predicate → bit 0 of ballot must be set.
testing/python/language/test_tilelang_language_warp_vote.py:91
↓ 1 callersFunctionkernel_complex
()
testing/python/language/test_tilelang_language_assume.py:57
↓ 1 callersFunctionkernel_ffs_ballot_sync
Find the first lane whose ballot predicate is true.
testing/python/language/test_tilelang_language_warp_vote.py:159
↓ 1 callersFunctionkernel_fn
()
testing/python/issue/test_tilelang_issue_2188.py:17
↓ 1 callersFunctionkernel_func
(block_M, block_N, num_stages, threads)
benchmark/blocksparse_attention/benchmark_tilelang_block_sparse_fmha.py:47
↓ 1 callersFunctionkernel_func
(block_M, block_N, num_stages, threads)
examples/minference/example_vertical_slash_sparse_attn.py:38
↓ 1 callersFunctionkernel_func
(block_M, block_N, num_stages, threads)
examples/blocksparse_attention/example_tilelang_block_sparse_attn.py:48
↓ 1 callersFunctionkernel_func
(block_M, block_N, num_stages, threads)
examples/seer_attention/block_sparse_attn_tilelang.py:49
↓ 1 callersFunctionkernel_func
(block_M, block_N, num_stages, threads)
testing/python/transform/test_tilelang_transform_config_index_bitwidth.py:26
↓ 1 callersFunctionkernel_match_all_sync_false
Lanes disagree → match_all_sync returns 0.
testing/python/language/test_tilelang_language_warp_vote.py:413
↓ 1 callersFunctionkernel_match_all_sync_true
All lanes share value 7 → match_all_sync returns the full mask.
testing/python/language/test_tilelang_language_warp_vote.py:397
↓ 1 callersFunctionkernel_match_any_sync
Lanes 0-15 share value 1; lanes 16-31 share value 2. match_any_sync should return 0x0000FFFF for the first half and 0xFFFF0000 for the second
testing/python/language/test_tilelang_language_warp_vote.py:359
↓ 1 callersFunctionkernel_syncthreads_and_false
Thread 0 passes predicate 0 → syncthreads_and returns 0.
testing/python/language/test_tilelang_language_warp_vote.py:271
↓ 1 callersFunctionkernel_syncthreads_and_true
All threads pass predicate 1 → syncthreads_and returns non-zero.
testing/python/language/test_tilelang_language_warp_vote.py:255
↓ 1 callersFunctionkernel_syncthreads_count
Exactly half the threads (lanes 0–15) pass predicate 1.
testing/python/language/test_tilelang_language_warp_vote.py:224
↓ 1 callersFunctionkernel_syncthreads_or_false
No thread passes predicate 1 → syncthreads_or returns 0.
testing/python/language/test_tilelang_language_warp_vote.py:323
↓ 1 callersFunctionkernel_syncthreads_or_true
At least one thread (lane 0) passes predicate 1 → syncthreads_or != 0.
testing/python/language/test_tilelang_language_warp_vote.py:307
↓ 1 callersFunctionkernel_vectorize
(M)
testing/python/language/test_tilelang_language_assume.py:28
↓ 1 callersFunctionkernel_with_assume
()
testing/python/language/test_tilelang_language_assume.py:8
↓ 1 callersFunctionkernel_with_shfl_sync
()
testing/python/language/test_tilelang_language_warp_sync.py:38
↓ 1 callersFunctionkernel_with_warp_sync
()
testing/python/language/test_tilelang_language_warp_sync.py:9
↓ 1 callersFunctionkernels_with_pdl_overlap_window
Create a two-kernel PDL benchmark with real pre-sync overlap work.
testing/python/language/test_tilelang_language_pdl.py:120
↓ 1 callersFunctionkernels_without_pdl_overlap_window
Create a strict same-stream two-kernel baseline for the PDL overlap benchmark.
testing/python/language/test_tilelang_language_pdl.py:82
↓ 1 callersFunctionldg128_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:71
↓ 1 callersFunctionldg128_pred_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:201
↓ 1 callersFunctionldg256_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:101
↓ 1 callersFunctionldg256_pred_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:238
↓ 1 callersFunctionldg32_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:14
↓ 1 callersFunctionldg32_pred_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:130
↓ 1 callersFunctionldg64_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:42
↓ 1 callersFunctionldg64_pred_kernel
(X, Y)
testing/python/language/test_tilelang_language_ldg.py:165
↓ 1 callersMethodldmatrix_a
(self, A_local_buf, A_shared_buf: Buffer | BufferRegion, ki, rk=0)
tilelang/rocm/intrinsics/mfma_macro_generator.py:374
↓ 1 callersMethodldmatrix_b
(self, B_local_buf, B_shared_buf: Buffer | BufferRegion, ki, rk=0)
tilelang/rocm/intrinsics/mfma_macro_generator.py:418
↓ 1 callersFunctionlet_expr
(v: Var, value: PrimExpr, body: PrimExpr)
tilelang/language/ast/ir.py:939
↓ 1 callersFunctionlet_stmt
(v: Var, value: PrimExpr)
tilelang/language/ast/ir.py:943
↓ 1 callersFunctionlinear_ramp_factor
Computes a linear ramp function used to smooth values between a minimum and maximum range. Args: min (float): Minimum va
examples/deepseek_v32/inference/model.py:390
↓ 1 callersMethodload_from_disk
Load a complete autotune result and its compiled kernel from disk.
tilelang/autotuner/param.py:470
↓ 1 callersFunctionload_from_file
Inverse of ``dump_to_file``. Tolerant of missing / unknown fields so older cache entries keep working when the dataclass evolves.
tilelang/contrib/hip_resource_info.py:130
↓ 1 callersFunctionload_frontend_cached
( frontend_key_data: dict, *, out_idx: list[int] | int | None = None, target: str | Target | N
tilelang/cache/__init__.py:128
↓ 1 callersFunctionlocal_var_scalar
()
testing/python/amd/test_tilelang_hip_codegen.py:490
↓ 1 callersFunctionlogits_fn
()
examples/deepseek_v32/fp8_lighting_indexer.py:262
↓ 1 callersFunctionlower
enable_host_codegen: whether to enable host codegen, default is False, as we have our own host codegen implementation in jit. enable_devi
tilelang/engine/lower.py:310
↓ 1 callersMethodlower
(self, mod: IRModule, target: Target)
tilelang/backend/pass_pipeline/pipeline.py:22
↓ 1 callersMethodlower
(self, target: Target, layout_map: dict, thread_bounds: Range, thread_var: tirx.Var)
tilelang/tileop/gemm_sp/__init__.py:61
↓ 1 callersMethodmacro
(self, name=None, annotations=None)
tilelang/language/eager/builder.py:240
↓ 1 callersFunctionmacro_apply_func
(x, fn)
testing/python/language/test_tilelang_language_frontend_v2.py:238
↓ 1 callersMethodmacro_arg
(self, name, value)
tilelang/language/eager/builder.py:709
↓ 1 callersFunctionmacro_return_constant
()
testing/python/language/test_tilelang_language_frontend_v2.py:225
↓ 1 callersFunctionmacro_return_expr
(x)
testing/python/language/test_tilelang_language_frontend_v2.py:233
↓ 1 callersFunctionmacro_return_frame
(x)
testing/python/language/test_tilelang_language_frontend_v2.py:229
↓ 1 callersFunctionmain
( A: T.Tensor((M, K), T.float4_e2m1fn), B: T.Tensor((N, K), T.float8_e4m3fn), D: T.Ten
maint/gemm/correctness_evaluation_tcgen05_fp4_unpacked.py:28
↓ 1 callersFunctionmain
()
maint/host_checks/08_device_id_mismatch.py:7
↓ 1 callersFunctionmain
()
maint/host_checks/02_pointer_type_error.py:11
↓ 1 callersFunctionmain
()
maint/host_checks/01_num_args_mismatch.py:11
↓ 1 callersFunctionmain
()
maint/host_checks/06_strides_mismatch.py:7
↓ 1 callersFunctionmain
()
maint/host_checks/run_all.py:6
↓ 1 callersFunctionmain
()
maint/host_checks/03_ndim_mismatch.py:7
↓ 1 callersFunctionmain
()
maint/host_checks/09_null_data_pointer.py:15
↓ 1 callersFunctionmain
()
maint/host_checks/07_device_type_mismatch.py:7
↓ 1 callersFunctionmain
()
maint/host_checks/04_dtype_mismatch.py:7
↓ 1 callersFunctionmain
()
maint/host_checks/10_scalar_type_mismatch.py:6
↓ 1 callersFunctionmain
()
maint/host_checks/05_shape_mismatch.py:7
↓ 1 callersFunctionmain
Main execution function.
maint/precision/compare_ops.py:437
↓ 1 callersFunctionmain
(do_bench: bool = True)
examples/gemv/example_gemv.py:330
↓ 1 callersFunctionmain
( A: T.Tensor((M, K), dtype), B: T.Tensor((N, K), dtype), C: T.Tensor((M,
examples/gemm/example_gemm_autotune.py:129
↓ 1 callersFunctionmain
(M=4096, N=4096, K=4096)
examples/gemm/example_gemm_persistent.py:90
↓ 1 callersFunctionmain
()
examples/gemm/example_gemm.py:29
↓ 1 callersFunctionmain
(M=4096, N=4096, K=4096)
examples/gemm/example_gemm_intrinsics.py:161
↓ 1 callersFunctionmain
( A: T.Tensor((M, K), dtype), B: T.Tensor((N, K), dtype), C: T.Tensor((M,
examples/gemm/example_gemm_advanced_autotune.py:184
↓ 1 callersFunctionmain
()
examples/deepseek_nsa/example_tilelang_nsa_fwd.py:121
↓ 1 callersFunctionmain
()
examples/deepseek_nsa/example_tilelang_nsa_decode.py:131
↓ 1 callersFunctionmain
()
examples/blocksparse_gemm/example_blocksparse_gemm.py:93
↓ 1 callersFunctionmain
Main function to load the model and perform interactive or batch text generation. Args: ckpt_path (str): Path to the model checkpoin
examples/deepseek_v32/inference/generate.py:81
↓ 1 callersFunctionmain
Converts and saves model checkpoint files into a specified format. Args: hf_ckpt_path (str): Path to the directory containing the in
examples/deepseek_v32/inference/convert.py:35
↓ 1 callersFunctionmain
()
examples/deepseek_deepgemm/example_deepgemm_fp8_2xAcc.py:169
↓ 1 callersFunctionmain
(M=16384, N=16384, K=16384)
examples/dynamic_shape/example_dynamic.py:97
↓ 1 callersFunctionmain
(B=1, S=1024, H=16, D=128)
examples/linear_attention/example_linear_attn_bwd.py:153
↓ 1 callersFunctionmain
(B=1, S=512, H=16, D=128)
examples/linear_attention/example_linear_attn_fwd.py:114
↓ 1 callersFunctionmain
()
examples/linear_attention/example_retention_fwd.py:85
↓ 1 callersFunctionmain
(A, B)
examples/convolution/example_convolution.py:9
↓ 1 callersFunctionmain
(A, B)
examples/convolution/example_convolution_autotune.py:17
↓ 1 callersFunctionmain
(batch=1, heads=1, seq_len=4096, head_dim=64, vertical_size=1000, slash_size=200)
examples/minference/example_vertical_slash_sparse_attn.py:561
↓ 1 callersFunctionmain
()
examples/gemm_splitk/example_tilelang_gemm_splitk.py:29
↓ 1 callersFunctionmain
()
examples/gemm_splitk/example_tilelang_gemm_splitk_vectorize_atomicadd.py:28
↓ 1 callersFunctionmain
( A: T.Tensor[A_shape, in_dtype], B: T.Tensor[B_shape, storage_dtype], C: T.Tensor[C_s
examples/dequantize_gemm/example_dequant_gemv_fp16xint4.py:78
↓ 1 callersFunctionmain
( A: T.Tensor(A_shape, in_dtype), B: T.Tensor(B_shape, storage_dtype), C: T.Tensor((M,
examples/dequantize_gemm/example_dequant_gemm_fine_grained.py:47
↓ 1 callersFunctionmain
( B: T.Tensor(B_shape, storage_dtype), C: T.Tensor((N, K), in_dtype), )
examples/dequantize_gemm/example_dequant_gemm_fp4_hopper.py:53
↓ 1 callersFunctionmain
( B: T.Tensor(B_shape, storage_dtype), C: T.Tensor((N, K), in_dtype), )
examples/dequantize_gemm/example_dequant_gemm_w4a8.py:44
↓ 1 callersFunctionmain
( Q: T.Tensor(q_shape, dtype), K: T.Tensor(kv_shape, dtype), V: T.Tensor(kv_shape, dty
examples/amd/example_amd_flash_attn_bwd.py:120
↓ 1 callersFunctionmain
( Q: T.Tensor(q_shape, dtype), K: T.Tensor(kv_shape, dtype), V: T.Tensor(kv_shape, dty
examples/amd/example_amd_flash_attn_fwd.py:138
↓ 1 callersFunctionmain
(BATCH=1, H=32, Q_CTX=128, KV_CTX=8192, D_HEAD=128, causal=False)
examples/flash_decoding/example_mha_inference.py:239
↓ 1 callersFunctionmain
(batch: int = 1, heads: int = 32, groups: int = 8, kv_seqlen: int = 8192, dim: int = 128, tune: bool = False)
examples/flash_decoding/example_gqa_decode.py:403
↓ 1 callersFunctionmain
( Q: T.Tensor(shape, dtype), K: T.Tensor(shape, dtype), V: T.Tensor(shape, dtype),
examples/flash_attention/example_mha_fwd_bshd.py:30
↓ 1 callersFunctionmain
( BATCH: int = 8, H: int = 32, N_CTX: int = 1024, D_HEAD: int = 64, causal: bool = False,
examples/flash_attention/example_mha_bwd_bhsd.py:301
↓ 1 callersFunctionmain
( Q: T.Tensor(q_shape, dtype), K: T.Tensor(kv_shape, dtype), V: T.Tensor(kv_shape, dty
examples/flash_attention/example_gqa_fwd_bshd.py:77
← previousnext →2,401–2,500 of 7,254, ranked by callers