wrongtest-intellif commented on code in PR #18421:
URL: https://github.com/apache/tvm/pull/18421#discussion_r2533751212
##########
tests/python/codegen/test_target_codegen_cuda.py:
##########
@@ -877,5 +877,37 @@ def main(A: T.Buffer((16, 16), "float32"), B:
T.Buffer((16, 16), "float32")):
assert "return;" in cuda_code
[email protected]_gpu
[email protected]_cuda
+def test_cuda_loop_step():
+ @T.prim_func
+ def cuda_loop_step(
+ A: T.Buffer((1024,), "float32"),
+ B: T.Buffer((1024,), "float32"),
+ C: T.Buffer((1024,), "float32"),
+ ):
+ # Each thread computes a strided subset of the i loop: start = tx*3,
step = 96 (3 * 32 threads)
+ for bx in T.thread_binding(1, "blockIdx.x"):
+ for tx in T.thread_binding(96, "threadIdx.x"):
+ for i in T.serial(tx, 4096, step=96):
Review Comment:
/gemini review
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
