- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
We noticed that many of our test cases do not show performance improvement from 1-core, 2-core, 3-core, to 4-core computers. One case with 411 RHS vectors actually ran a bit slower in 4-core than 1-core. For this case, 95% of CPU is on Pardiso back-substitution. In the 4-core run in a 4-core computer, the Task Manager did show 100% CPU. When running the same case with just 1-core in the same computer, it is also confirmed that Task Manager showed 25% CPU. Do we miss anything?
Link Copied
2 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
What is the size of the input problem?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The problem is known about andthey are working on it. See the following,
http://software.intel.com/en-us/forums/showthread.php?t=73240&o=d&s=lr
http://software.intel.com/en-us/forums/showthread.php?t=73240&o=d&s=lr
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