- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I can only find a reference to em64t/libmkl_lapack.a
and nothing that would be ilp64.
I need to use routine dgesdd, for m=n=65536. However, the appropriate lwork value for this is approx
1.7e10, so too large to fit into a 32 bit integer.
If there is no ilp64 version, does this mean lapack dgesdd can't be used for large matrices?
Thanks...
I can only find a reference to em64t/libmkl_lapack.a
and nothing that would be ilp64.
I need to use routine dgesdd, for m=n=65536. However, the appropriate lwork value for this is approx
1.7e10, so too large to fit into a 32 bit integer.
If there is no ilp64 version, does this mean lapack dgesdd can't be used for large matrices?
Thanks...
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
With MKL 10 or 11, rather than using the backwards compatibility script (which specified lp64, core, thread), you must specify the ilp64 library, core library, and your choice of thread or sequential. For MKL 9, there is a separate download for ilp64.

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