- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I use pardiso to solve a million cells problem, i.e. the matrix size is only 1 million * 1million, the nozero number is only 3.9 million numbers, in my pc with 2G RAM.
Unfortunately, pardiso crashes during LU factoration. Pardiso tells me during reordering in partition metis it produces nearly 800 million nozeros in L+U.
It seems to be a little ridiculous, right? In LU0 factoration, I think the nozero number should not exceed the original matrix's nozero.
I do think some parameters are set to be wrong. But I do not know which parameter should be responsible for such case , can anyone point me out for this problem?
Thanks very much
I use pardiso to solve a million cells problem, i.e. the matrix size is only 1 million * 1million, the nozero number is only 3.9 million numbers, in my pc with 2G RAM.
Unfortunately, pardiso crashes during LU factoration. Pardiso tells me during reordering in partition metis it produces nearly 800 million nozeros in L+U.
It seems to be a little ridiculous, right? In LU0 factoration, I think the nozero number should not exceed the original matrix's nozero.
I do think some parameters are set to be wrong. But I do not know which parameter should be responsible for such case , can anyone point me out for this problem?
Thanks very much
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
What MKL version are you using?
Since version 10.0, MKL support out-of-core version (OOC).
The OOC PARDISO can solve very large problems by holding the matrix factors in files on the disk.
So, please set the parameters as:
iparm(3) = 1
iparm(60)= 2
For FortranOr
Iparm[2] = 1
Iparm[59] =2
for c/c++:
and try to use pardiso ooc. it should work. :)
regards, Gennady
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