frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

Updated Dispatcher Structured Kernels Lab (markdown)

Brian Hirsh
2022-04-19 09:17:52 -04:00
parent a2fa600791
commit 2e3c584d1e
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
Dispatcher---Structured-Kernels-Lab.md

@ -90,6 +90,7 @@ Run the following snippet before/after your change, and compare
(Also, make sure that you built with `REL_WITH_DEB_INFO=1`, so you're using a release builds. Otherwise your times will be unnecessarily slow.)
```
>>> import torch
>>> with torch.profiler.profile() as prof:
... a = torch.ones(2)
@ -97,7 +98,7 @@ Run the following snippet before/after your change, and compare
... c = a + b
...
>>> str(prof.events())
```
You should see that aten::mul() no longer shows up in the profiler! The profiler is no longer aware of aten::mul(), since the function bypasses the dispatcher. It also completely excludes the time spent running at::mul() in the total time, since it’s not aware of the call.