8 lines
211 B
Python
8 lines
211 B
Python
import triton
|
|
import triton.language as tl
|
|
|
|
@triton.jit
|
|
def kernel(X, stride_xm, stride_xn, BLOCK: tl.constexpr):
|
|
pass
|
|
|
|
ret = triton.compile(kernel, "*fp32,i32,i32", constants={"BLOCK": 256}, output="ttgir") |