Functionmixed_async_num_stages_one(
x: T.Tensor((M, K), T.float16),
meta: T.Tensor((2 * cp_async_bytes,), T.uint8),
y: T
testing/python/issue/test_tilelang_issue_tma_no_ws.py:252
Functionmixed_gemm_shared_barrier(
A: T.Tensor((M, K), T.float16),
B: T.Tensor((K, N), T.float16),
C: T.Tensor((M, N),
testing/python/issue/test_tilelang_issue_tma_no_ws.py:311
Functionmma_op(
a_ptr: Pointer,
a_offset,
b_ptr: Pointer,
b_offset,
c_ptr: Pointer,
tilelang/contrib/cutedsl/ptx_mma.py:91
Functionmma_sp_op(
a_ptr: Pointer,
a_offset,
b_ptr: Pointer,
b_offset,
c_ptr: Pointer,
tilelang/contrib/cutedsl/ptx_mma.py:180
Functionmoe_forward_tilelang_routed(
input,
routed_expert_gate,
routed_expert_up,
routed_expert_down,
routed_expert_weights,
examples/fusedmoe/example_fusedmoe_tilelang.py:92
Functionmoe_forward_tilelang_shared(
input,
shared_W_gate,
shared_W_up,
shared_W_down,
up_logits,
output,
dtype,
examples/fusedmoe/example_fusedmoe_tilelang.py:12
Functionmxfp8_blockscaled_gemm_2cta(
A,
B,
SFA,
SFB,
block_M,
block_N,
block_K,
in_dtype,
out_dtype,
accu
examples/blockscaled_gemm_sm100/gemm_mxfp8_blockscaled_1d1d.py:170
Functionmxfp8_blockscaled_gemm_2cta_persistent(
A,
B,
SFA,
SFB,
block_M,
block_N,
block_K,
in_dtype,
out_dtype,
accu
examples/blockscaled_gemm_sm100/gemm_mxfp8_blockscaled_1d1d.py:319
Functionnaive_gemv(
A,
B,
BLOCK_N: int,
BLOCK_K: int,
dtype: T.dtype = T.float16,
accum_dtype: T.dtype =
examples/gemv/example_gemv.py:15
Functionnaive_nsa_simpler""" Args: q (torch.Tensor): queries of shape `[B, T, HQ, K]` if `head_first=False` else `[B, HQ, T, K]`. k (torch.Ten
examples/deepseek_nsa/reference.py:152
Functionnaive_splitk_gemv(
A,
B,
BLOCK_N: int,
BLOCK_K: int,
dtype: T.dtype = T.float16,
accum_dtype: T.dtype =
examples/gemv/example_gemv.py:46
Functionnegative_index_before(A: T.Buffer((16,), T.float32), B: T.Buffer((16,), T.float32))
testing/python/language/test_tilelang_language_negative_index.py:8
Functionnegative_index_expected(A: T.Buffer((16,), T.float32), B: T.Buffer((16,), T.float32))
testing/python/language/test_tilelang_language_negative_index.py:14
Functionnegative_index_loop_before(A: T.Buffer((16,), T.float32), B: T.Buffer((4,), T.float32))
testing/python/language/test_tilelang_language_negative_index.py:20
Functionnegative_index_loop_expected(A: T.Buffer((16,), T.float32), B: T.Buffer((4,), T.float32))
testing/python/language/test_tilelang_language_negative_index.py:27
Functionnegative_index_symbolic_before(shift: T.int32, A: T.Buffer((16,), T.float32), B: T.Buffer((16,), T.float32))
testing/python/language/test_tilelang_language_negative_index.py:34
Functionparallel_nsar""" Args: q (torch.Tensor): queries of shape `[B, T, HQ, K]` if `head_first=False` else `[B, HQ, T, K]`. k (torch.Ten
examples/deepseek_nsa/benchmark/benchmark_nsa_fwd.py:229
Functionparallel_nsa_bwd_kernel_dkv(
q,
k,
v,
lse_slc,
lse_swa,
delta_slc,
delta_swa,
do_slc,
do_swa,
dk,
examples/deepseek_nsa/example_triton_nsa_bwd.py:229
Functionparallel_nsa_bwd_kernel_dq(
q,
k,
v,
lse_slc,
delta_slc,
do_slc,
lse_swa,
delta_swa,
do_swa,
dq,
examples/deepseek_nsa/example_triton_nsa_bwd.py:380
Functionparallel_nsa_fwd_kernel(
q,
k,
v,
o_slc,
o_swa,
lse_slc,
lse_swa,
scale,
block_indices,
block
examples/deepseek_nsa/example_triton_nsa_fwd_varlen.py:32
Functionparallel_nsa_fwd_kernel(
q,
k,
v,
o_slc,
o_swa,
lse_slc,
lse_swa,
scale,
block_indices,
block
examples/deepseek_nsa/example_triton_nsa_bwd.py:32
Functionparallel_nsa_fwd_kernel(
q,
k,
v,
o_slc,
o_swa,
lse_slc,
lse_swa,
scale,
block_indices,
block
examples/deepseek_nsa/example_triton_nsa_fwd.py:32
Functionparallel_nsa_fwd_kernel(
q,
k,
v,
o_slc,
o_swa,
lse_slc,
lse_swa,
scale,
block_indices,
block
examples/deepseek_nsa/benchmark/benchmark_nsa_fwd.py:28
Functionparallel_nsa_kernel_mask(
block_indices,
block_counts,
block_mask,
T: tl.constexpr,
H: tl.constexpr,
S: tl.con
examples/deepseek_nsa/example_triton_nsa_bwd.py:345