PDA

View Full Version : First ever dual core Opteron 866 tested!!!


AthlonXP1800
04-13-05, 06:09 AM
http://www.theinquirer.net/?article=22501

The article of the tests were in Japanese, need to use Babel Fish (http://babelfish.altavista.com/) or other translation to read it.

Dual core Opteron 866 looked brilliant, pity only one benchmark software Sandra 2005 can measure dual core performances, maybe all softwares will issue new patches to take advantage of dual core for next week launch and also Windows XP x64 Edition too. :p

Viral
04-13-05, 10:36 AM
More interestingly, even a dual Opteron 248 were not able to beat a single Opteron 866. AMD has done a fantastic job here.

Simply amazing, for those who don't know a 248 is basically the same as one of those cores, except it has it's own dedicated memory controller and not a shared one. I guess the core enhancements and the fact that the data transferred between the two cores doesn't need to transfer through a physical HT bus means the peformance is better. :)

So there it is. AMD's Dual core implementation is better than Dual Processors.

superklye
04-13-05, 10:41 AM
I love AMD. I really do.

Sazar
04-13-05, 10:44 AM
Gonna wait for the actual numbers to drop before I start jizzin but what with montecito and the 866, turning out to be a good year for proc fans :D

Superfly
04-13-05, 11:03 AM
sounds wicked - im waiting for 2006/7 though before buying a new chip.

I want 2nd/3rd Generation dual core :-)

Viral
04-13-05, 10:11 PM
sounds wicked - im waiting for 2006/7 though before buying a new chip.

I want 2nd/3rd Generation dual core :-)

Yeah, I'm in the same boat.

By that time most new games should be taking advantage of dual core.

superklye
04-13-05, 10:19 PM
Yeah, I'm in the same boat.

By that time most new games should be taking advantage of dual core.
And that delicious new PhysX card.

JoKeRr
04-13-05, 11:00 PM
sounds wicked - im waiting for 2006/7 though before buying a new chip.

I want 2nd/3rd Generation dual core :-)

same here, I expected my P4 to last me a while, and it should---I'll just get a 30 capper 3.0C and OC it more.

Strahd
04-14-05, 10:40 AM
*dreams of dual core AMD's and dual GPU's on a single card*

ynnek
04-14-05, 02:48 PM
you better also dream up software to take advantage of it all.. or it could be a nitemare.. ;)

*dreams of dual core AMD's and dual GPU's on a single card*

moshpit
04-15-05, 08:01 PM
I'll say the same thing here I said in the other thread, I'm unconvinced of the truth of those benchmarks. Before saying anything else tho, I use AMD and NOTHING else in my rigs and am no Intel fanboi, so don't try to shoot me down with that accusation, because it simply isn't true.

The fact is, I've never heard of single threaded applications actually benefiting from a SMP enviorment over the same chip used in a single proc config. The end of those benchmarks shows 2x 866 outperforming 1x 866 on 3Dmark03 (this one shows some minor SMT improvements via hyperthreading on P4's, but never on SMP setups) and Aquamark. That baffles me considering every other SMP rig I've ever seen run those 2 particular benchmarks runs them a tad slower then the same CPU running in single chip config can. Differing platforms show varying degrees of this, Intel SMP rigs suffering the worst from SMP overhead, but Athlon MPs suffer it too, and even the mighty Opterons suffer from it too, tho they suffer the least of all x86 SMP rigs when forced to run a single threaded benchmark. This has been a rule of the SMP world for the entire time x86 has even HAD SMP available to it.

Because of this, I find those benchmarks hard to swallow. I WISH they were true, and hope that maybe the 866 has some trick up it's sleeve to FINALLY break that rule, but I've heard NOTHING that the 866 has up it's sleeves that would account for this, and being dual core would only compound the issue, not ease it.

Any theories or explanaitions anybody? Please?