Icemist commented on code in PR #13726:
URL: https://github.com/apache/tvm/pull/13726#discussion_r1064507883
##########
src/runtime/profiling.cc:
##########
@@ -894,15 +891,21 @@ PackedFunc WrapTimeEvaluator(PackedFunc pf, Device dev,
int number, int repeat,
std::max((min_repeat_ms / (duration_ms / number) + 1), number *
golden_ratio));
}
+ int64_t accum_t_nanos = 0;
// start timing
- Timer t = Timer::Start(dev);
for (int j = 0; j < number; ++j) {
+ // call preprocessing function
+ if (f_preproc != nullptr) {
+ f_preproc.CallPacked(args, &temp);
Review Comment:
Maybe warmup has a place, also we already have skip_first=1 similar to
pytorch.
But this would require changes not only in the C++ part, but also in the C
and webasm part.
For ProfileFunction plz look original PR with
https://github.com/apache/tvm/pull/9553 there are a couple of words about the
motivation for adding.
--
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]
