Fix elapsed time computation from issue #3649 (#3650)

Fixes #3649 

## Description
This PR fixes the timing helper used in the `torch.compile` tutorial by
dividing the CUDA event elapsed time by 1000 (ms -> s) instead of 1024.

## Checklist
<!--- Make sure to add `x` to all items in the following checklist: -->
- [x] The issue that is being fixed is referred in the description (see
above "Fixes #ISSUE_NUMBER")
- [x] Only one issue is addressed in this pull request
- [x] Labels from the issue that this PR is fixing are added to this
pull request
- [x] No unnecessary issues are included into this pull request.


cc @williamwen42 @msaroufim @anijain2305

Co-authored-by: sekyondaMeta <127536312+sekyondaMeta@users.noreply.github.com>

Author: starpiens

intermediate_source/torch_compile_tutorial.py

@@ -163,7 +163,7 @@ def timed(fn):
     result = fn()
     end.record()
     torch.cuda.synchronize()
-    return result, start.elapsed_time(end) / 1024
+    return result, start.elapsed_time(end) / 1000
 
 
 inp = torch.randn(4096, 4096).cuda()