- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi
It has dawned upon me that Cluster OpenMP was let go off sometime last year. Andwith it the easy and perhaps only access to using relatively simple OpenMP on distributed memory is gone as well. Am I right on that? What are the current alternatives, if any? If possible, can someone shed light upon why cluster OpenMP was retired and what alternatives Intel suggested/had in mind?
Abhi
It has dawned upon me that Cluster OpenMP was let go off sometime last year. Andwith it the easy and perhaps only access to using relatively simple OpenMP on distributed memory is gone as well. Am I right on that? What are the current alternatives, if any? If possible, can someone shed light upon why cluster OpenMP was retired and what alternatives Intel suggested/had in mind?
Abhi
Link Copied
3 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
You're more entitled to speculate than I, but coarrays come "to mind" as a more portable and possible more scalable alternative to clomp.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
General lack of interest among customers and fewer cases than expected where it showed a benefit. Like Tim, I'd suggest coarrays at this point.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks Tim and Steve.
I agree that since CO-Arrays is somewhat "more guaranteed" being part of language standard, I wonder the obvious: (a) how quickly robust and well-optimized compilers become available, and (b) how quickly expertise in using CO-Arrays is developed. On the former, IVF is clearly making progress but some key issues as outlined here (http://software.intel.com/en-us/forums/showthread.php?t=82439&o=a&s=lr) mean that, as always,it is going to take time for a mature solution.
Abhi
I agree that since CO-Arrays is somewhat "more guaranteed" being part of language standard, I wonder the obvious: (a) how quickly robust and well-optimized compilers become available, and (b) how quickly expertise in using CO-Arrays is developed. On the former, IVF is clearly making progress but some key issues as outlined here (http://software.intel.com/en-us/forums/showthread.php?t=82439&o=a&s=lr) mean that, as always,it is going to take time for a mature solution.
Abhi

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