Anndrey24 commented on code in PR #16899:
URL: https://github.com/apache/tvm/pull/16899#discussion_r1574907415
##########
python/tvm/relay/op/strategy/arm_cpu.py:
##########
@@ -252,6 +252,17 @@ def conv2d_strategy_arm_cpu(attrs, inputs, out_type,
target):
)
# Non-quantized cases
if is_aarch64 and data.dtype in ["float32", "float16"]:
+ if target.features.has_sve:
+ # This strategy is currently suboptimal because of
LLVM's limited support
+ # for scalable vector alias analysis, which causes
redundant loads / stores
Review Comment:
Done!
--
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: commits-unsubscr...@tvm.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
