yzh119 edited a comment on pull request #10638:
URL: https://github.com/apache/tvm/pull/10638#issuecomment-1068874472
We failed to address the case of outer product:
```python
@T.prim_func
def outer_product_desc(a: T.handle, b: T.handle, c: T.handle) -> None:
A = T.match_buffer(a, (16, 1), offset_factor=1)
B = T.match_buffer(b, (16, 1), offset_factor=1)
C = T.match_buffer(c, (16, 16), offset_factor=1)
with T.block("root"):
T.reads(
A[0 : 16, 0 : 1],
B[0 : 16, 0 : 1],
)
T.writes(C[0 : 16, 0 : 16])
for i, j in T.grid(16, 16):
with T.block("update"):
vii, vjj = T.axis.remap("SS", [i, j])
C[vii, vjj] = C[vii, vjj] + A[vii, 0] * B[vjj, 0]
```
Note that inside block `update`, there is no sign that it's a reduction
block (no `R`, no `init`), but it's actually performing reduction.
@spectrometerHBH WDYT?
--
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]
