On Wed, 11 Apr 2012, Rich wrote:
You neglect to mention your test platform's kernel or userland version
- are you running the latest illumos head, the stock kernel+userland
provided by OpenIndiana/Nexenta of some version, etc, etc?
The OS was installed by someone else and the testing is on a borrowed
remote machine (possibly vaporizing today). I assume that it is stock
kernel+userland from OI 151a2:
$ uname -a
SunOS openindiana 5.11 oi_151a2 i86pc i386 i86pc Solaris
This would pre-date the release of Bulldozer-based CPUs and it is a
wonder that the system works at all.
I know that the Bulldozer family is a strange beast, and a few commits
related to them have made it into illumos recently, though I don't
know of any that would make a difference for your purposes.
Yes, it is a strange beast. I have noticed mention of "lock
contention" in several Bulldozer reviews. With 64-cores, it is
necessary for locking to work quite efficiently but perhaps it is
falling back to a slow fail-safe implementation.
I think that my software should run very well on this hardware, but
AMD benchmarked only on two specific commercial Linux variants and
using the compiler they maintain (Open64) so there is no telling what
is necessary in the OS, run-time libraries, and compiler, to achieve
the sweetness indicated by the SPEC benchmark reports.
It is of course possible that there is an issue with my software or in
GCC's GOMP library.
Bob
--
Bob Friesenhahn
[email protected], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/
-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com
