mirror of
https://github.com/tile-ai/tilelang.git
synced 2026-10-23 07:00:29 +08:00
Gongen-Ali
4e53b50750
* Fix some bugs on ci and open rocm ci test * fix pre-commit * fix build issue on hip * refactor on hip * pre-commit fix * fix test with pdl * fix pdl test
58 lines
1.5 KiB
Python
58 lines
1.5 KiB
Python
import torch
|
|
import tilelang
|
|
import tilelang.testing
|
|
from tilelang import language as T
|
|
|
|
|
|
@tilelang.jit(
|
|
pass_configs={
|
|
tilelang.PassConfigKey.TL_DISABLE_WARP_SPECIALIZED: True,
|
|
tilelang.PassConfigKey.TL_DISABLE_TMA_LOWER: True,
|
|
},
|
|
)
|
|
def _fill_with_static_region_kernel():
|
|
num_tokens = T.symbolic("num_tokens")
|
|
|
|
@T.prim_func
|
|
def buggy_kernel(x: T.Tensor[(num_tokens,), "int64"]): # noqa: F821
|
|
with T.Kernel(num_tokens, threads=128) as _:
|
|
T.fill(x[0:128], 0)
|
|
|
|
return buggy_kernel
|
|
|
|
|
|
@tilelang.jit(
|
|
pass_configs={
|
|
tilelang.PassConfigKey.TL_DISABLE_WARP_SPECIALIZED: True,
|
|
tilelang.PassConfigKey.TL_DISABLE_TMA_LOWER: True,
|
|
},
|
|
)
|
|
def _fill_with_dynamic_region_kernel():
|
|
num_tokens = T.symbolic("num_tokens")
|
|
|
|
@T.prim_func
|
|
def buggy_kernel(x: T.Tensor[(num_tokens,), "int64"]): # noqa: F821
|
|
with T.Kernel(num_tokens, threads=128) as _:
|
|
a, b = T.alloc_var(T.int), T.alloc_var(T.int)
|
|
T.fill(x[a:b], 0)
|
|
|
|
return buggy_kernel
|
|
|
|
|
|
@tilelang.testing.requires_cuda
|
|
def test_fill_with_static_region_kernel():
|
|
kernel = _fill_with_static_region_kernel()
|
|
x = torch.zeros((256,), dtype=torch.int64, device="cuda")
|
|
kernel(x)
|
|
|
|
|
|
@tilelang.testing.requires_cuda
|
|
def test_fill_with_dynamic_region_kernel():
|
|
kernel = _fill_with_dynamic_region_kernel()
|
|
x = torch.zeros((256,), dtype=torch.int64, device="cuda")
|
|
kernel(x)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
tilelang.testing.main()
|