- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
When calling DGELSS multiple times with the same arguments, I find that the results of the FIRST call differ ever so slightly from subsequent calls which are all indentical. Typically, the differences are in the last couple of digits.
I have not found such problems with any other MKL functions.
I am using MKL7.0 on Windows XP Pro, Pentium IV 3Ghz. Visual C++ 7.1
Perhaps MKL is altering the floating point precision via the _controlfp / control87 functions?
Thanks.
David.
I have not found such problems with any other MKL functions.
I am using MKL7.0 on Windows XP Pro, Pentium IV 3Ghz. Visual C++ 7.1
Perhaps MKL is altering the floating point precision via the _controlfp / control87 functions?
Thanks.
David.
Link Copied
3 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Assuming that you are using the P4 library, are you passing data with identical alignments? If not, supposing that you are declaring your data in C, try __declspec(align 16) declarations.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi.
Yes I'm using the P4 kernel. I'm also developing in C++.
The data and workspace are in new'ed arrays - which are not always 16 bit aligned. When I replaced the new with an _aligned_malloc the problem disappeared.
Thanks for your help.
David.
Yes I'm using the P4 kernel. I'm also developing in C++.
The data and workspace are in new'ed arrays - which are not always 16 bit aligned. When I replaced the new with an _aligned_malloc the problem disappeared.
Thanks for your help.
David.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
David,
Yes, I think the MKL documents mention the possibility of better performance with aligned arguments, but perhaps not the possibility of slightly different round-off where mis-alignment requires more serial operations. Glad to hear that you found a reasonable solution.
Tim
Yes, I think the MKL documents mention the possibility of better performance with aligned arguments, but perhaps not the possibility of slightly different round-off where mis-alignment requires more serial operations. Glad to hear that you found a reasonable solution.
Tim
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