Hi everybody,
Yesterday I noticed the download of a new FAHCore_a7 version 0.0.17. Is there a changelog somewhere?
Looks like there was an upgrade of gcc from version 4.9.x to 6.3.x.
Does an FAHCore_a7 changelog exist?
Moderators: Site Moderators, FAHC Science Team
-
- Posts: 30
- Joined: Tue Nov 08, 2016 7:52 pm
- Hardware configuration: Xeon 1230v3 + Geforce RTX 2080
- Location: Germany
-
- Site Moderator
- Posts: 6359
- Joined: Sun Dec 02, 2007 10:38 am
- Location: Bordeaux, France
- Contact:
Re: Does an FAHCore_a7 changelog exist?
It's a minor bugfix in scientific code to allow new studies ...
-
- Posts: 30
- Joined: Tue Nov 08, 2016 7:52 pm
- Hardware configuration: Xeon 1230v3 + Geforce RTX 2080
- Location: Germany
Re: Does an FAHCore_a7 changelog exist?
The major change in the compiler version indicates a distribution upgrade on the part of the developer so I was wondering. Usually newer compilers provide for better support of optimization - yes, even if GROMACS is already highly optimized.
Re: Does an FAHCore_a7 changelog exist?
Gromacs has a long history of using hand-optimized-Assemply-Language code rather than depending on the (much less effective) compiler optimization for the inner loops. I have not looked at the recent code, but I expect they've updated the methology that they originally developed during the early days of SSE to use similar methods to pack the maximum number of parallel operations into the registers associated with SSE(m) and/or AVX(n). If there are still a few rarely used segments of code that can be optimized by the compiler, it's probably doesn't amount to much.
Posting FAH's log:
How to provide enough info to get helpful support.
How to provide enough info to get helpful support.