- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi there,
when comparing a multi-threaded and single-threaded version of our application we can see 15% runtime overhead throughout all functions for the multi-threaded variant.
Might this overhead be caused through the VTunes Profiler (running the Call Graph Module) ?
Or can we rely on the measured overhead to reside completely within our code?
thanks,
jan
when comparing a multi-threaded and single-threaded version of our application we can see 15% runtime overhead throughout all functions for the multi-threaded variant.
Might this overhead be caused through the VTunes Profiler (running the Call Graph Module) ?
Or can we rely on the measured overhead to reside completely within our code?
thanks,
jan
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The documentation tells you explicitly that call graph has a significant overhead, incurred by the instrumentation it adds to your code. For accurate performance evaluation, Event Sampling is much better.
Message Edited by tim18 on 11-24-2005 11:40 AM
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page