-
bfbagain: Congratulations Viktor
On such an excellent choice of Operating Systems. CentOS is the big secret among innovative IT shops. It is by far, and I do mean by far, the finest enterprise operating system out there, and no one, can complain about the cost.
What's more is the unbelievable ability to completely transfer whichever hard drive infrastructure that is currently in a system, to a completely different system and not skip a beat. (try that with a Windoze solution, HA!)
I was actually going to make a post about this sometime back but got too busy and didn't. FWIW, there is another adaption of Red Hat Enterprise Edition, and that's Whitebox Linux, however, it doesn't incorporate the same level of timely updates that CentOS does. And of course, Red Hat does deserve recognition for assisting in the distribution to the open-source community.
As to your choice of Opteron, again, it's the best chips out there. Believe it or not, even Windows XP (and server 2003) 64 bit versions fly under the Opterons.
With the technological advances that IDE drives have had, i.e., SATA II, many shops are confident and secure that the use of SATA II drives in mid-level servers are the preferred hard drive solution over SCSI drives.
Again, nice choice Viktor.
cheers
bfb
-
Norm Wattenberger: Re: Congratulations Viktor
> As to your choice of Opteron, again, it's the best
> chips out there.
Untill Monday
-
bfbagain: Opteron vs Xeon
Toms Hardware has a nice report.
-
Norm Wattenberger: Good summary of summaries
But the final words seemed a bit odd. After ten pages talking about Woodcrest the conclusion is that it won't knok out the Opteron in 4P systems and that in future quad-procressors the memory bus architecture will overload. But, on the first point, Woodcrest is not Intel's 4P answer. They have Tulsa and dual-core Itanium's. On the second point, I don't think quad-processor systems will generally be bandwidth limited. If you need the processor power in a quad-processor system, by definition you aren't running an application that is memory bound. Four FB-DIMM channels into two 1333ghz on-board memory busses with large caches should be more than adequate.
Of course the proof of consept, one way or another, will be in real life application. And my application is simulation. The Conroe's architecture is perfect for simulation
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
Bookmarks