In June 2003, I wrote an article titled “Why AMD does not want a test?”, referring to AMD European PR team blocking us from reviewing then newly released AMD Opteron platform. The reason for that was an on-the-record comment from AMD EMEA that “we would not be able to properly test the Opteron processors.” That e-mail stuck to my head lik glue and I still remember it sometimes. After all, I was just a former game developer that worked for AAA publishers until Sony bought [and ruined] the F1 license. You know. I certainly wasn’t capable of reviewing such an intelligent piece of equipment such as a processor.
Back in the day, a then 24-year old journalist gave’em a middle finger, purchased two Opteron 1.4 GHz processors out of my own dime and saw those Opty’s obliterating Intel Xeon DP 2.8 GHz. Yes, a 1.4 GHz Opteron demolished double-clocked NetBu[r]st-architecture based Xeons. Needless to say, AMD representatives were not amused with our publication of these findings, even though they worked in their favor.
Fast forward to end of May, 2010 and again, after several months of trying to acquire Magny-Cours processors from AMD to perform an independent review, we found ourselves in the same boat – we asked AMD could they provide us with the CPUs so that we could pit Magny-Cours against Intel Nehalem-EP [Xeon 5500], Westmere-EP [Xeon 5600] and Nehalem-EX [Xeon 7500]. This time around, we got the boot from the Opteron marketing/PR team at AMD US. Given that we enjoy otherwise positive relationship with AMD Consumer CPU and GPU and Commercial GPU team, this behavior seems out of character for AMD.
At BSN*, we have a combined several dozen years of experience in designing systems ranging from a workstation to high-performance computing clusters. In fact, in his previous company – our APAC editor sold the first AMD-based server in the APAC region and was an architect of the first AMD-based cluster in whole of APAC region.
Our upcoming review is not supported by Advanced Micro Devices in any way. Thus, we won’t have support if we encounter some issues, but that is the risk we are willing to take. The company commercial CPU press representatives went on the record and stated that Bright Side of News*will not receive Opteron processors and we had no other choice but to acquire this platform using other means. In order to protect the sources, we blurred out the serial numbers of the processors and motherboard in question.
Now, onto more happier topics – in order to avoid any possible claim of bias, we are going to pair this system with the best possible case scenario. We teamed up with CoolIT to provide us with yet-unreleased liquid cooling kit for Socket G34 platform [thanks to Geoff Lyon], we thank to David L.Szabados for providing us with four Seagate Enterprise-class Hard Drives – 2TB Constellation.ES, and lastly, David Leong from Kingston Technology for providing us with maybe the most important part of our review: 128GB of High-Capacity Registered ECC DDR3-1333 memory.
At well over $9,000 – this 128GB of system memory flat out beats a $1,500 [value of a single CPU] Opteron 6176 SE processor we’ll test. Theoretically, two Opteron 6176 SE processors should have 42.66GB/s of bandwidth towards 64GB of their direct attached memory and 85.31GB/s of combined bandwidth.
Needless to say, we are very excited to see the results of 24 physical cores and 128GB of memory. We’ll also pay close attention to how Windows 7 and Windows Server 2008 R2 behave, given that they caused us serious issues on our Nehalem-EX 7500-based workstation.
If you have any suggestions regarding tests we should run, we highly recommend you post them there. Thanks.
Original Autor: Theo Valich
Webmaster’s note: This news article is part of our extensive Archive on gadget news that have been happening in the past 10 years. For up to date stuff we would recommend a visit to our homepage. Additionally, we take great pride in our user manual section, as well as our VPN Troubleshooting guide, so be sure to check them out as well.
