- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I am finding that cluster_sparse_solver_64 fails to compute the correct solution when the number of non zeros exceeds 2147483647. Attached is a modification of the the cl_solver_sym_f.f example to demonstrate the issue. The example is modified to use a full symmetric matrix of size n. For n=65000 (nnz=2114576615) the solver succeeds. For n=67500 (nnz=2278158750), it fails. Tests fail/pass regardless of mpisize.
Tests require about 200Gb Ram (if run a single node) and take several hours to run.
Thanks
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Pan,
I saw that you filed a ticket of this same issue in the online service center.
I will communicate with you in the online service center.
This thread will be closed.
Best regards,
Khang
Link Copied
- « Previous
-
- 1
- 2
- Next »
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
As I mentioned before, I would love to get in contact with the developers as I have done extensive testing with both pardiso and cluster_sparse_solver routines and could provide more info that may be relevant. For example for pardiso_64, if you use the two-level factorization algorithm (fortran: iparm(24)=1), it also fails to provide to provide the correct solution for nnz > 2278158750, same way as the cluster_sparse_solver_64.
Is this issue been resolved? I still get a similar problem with pardiso_64 and nnz > 2**31 and iparm(24)=1 (Fortran), with oneAPI 2022.1.0. Setting iparm(24) to 0 solved the issue.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Nothing yet. Khang mentioned offline that a solution may come in the 23.1 release.
The frustrating issue is that although pardiso_64 works with iparm(24)=0, sparse_cluster_solver_64 does not. So, for large systems where you need distributed memory and cores in multiple nodes, the solver fails.
P.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Pan,
Thank you for the confirmation!
This is exactly what I thought. I just want to make sure.
If you would like private support, please register and create an issue ticket at the online service center:
Best regards,
Khang
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Pan,
I saw that you filed a ticket of this same issue in the online service center.
I will communicate with you in the online service center.
This thread will be closed.
Best regards,
Khang
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
I am still waiting for a solution to this issue. Seems that Khang is no longer with Intel and no one responds to the ticket I opened up. Is anyone monitoring this thread or open tickets?
Thanks, P.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- « Previous
-
- 1
- 2
- Next »