- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
i am trying to use mpiifx instead of mpiifort to run my code.
In both cases, i am able to compile but when i try to run them, the version compiled with mpiifx is able to work only if i use -check stack option.
Since this option deactivate any kind of optimization, is there a different way to solve this problem?
Thanks in advance
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I need to see all the compiler options you give to mpiifx
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks for your reply, i have some updates.
I was able to compile it with optimizations but with some changes.
1) I had to comment the directive "!DIR$ VECTOR ALIGNED" that was present before two different loops.
2) I need to add one of these options: -fstack-security-check, -fstack-protector-strong, -fstack-protector-all
Both changes are not needed if i use mpiifort.
These are my compiler options for both codes:
mpiifort
-O3 -xHost -mcmodel medium -shared-intel -warn all -warn noexternals
mpiifx
-O3 -xHost -mcmodel medium -shared-intel -warn all -warn noexternals + (see point 2)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
the option
-align array64bytes
could be added. it sometimes helps performance. this aligns all arrays on 64byte boundaries.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thank you for the suggestion.
It helps the performance but still those two fixes remain necessary to be able to use “mpiifx” instead of “mpiifort”

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page