Friday, January 20th 2012

AMD Vishera Packs Quad-Channel DDR3 IMC, G34 En Route Desktop?

AMD might be a little sore that its "Zambezi" FX processor family based on its much-hyped "Bulldozer" architecture didn't quite meet the performance expectations of a ground-up new CPU architecture, but it doesn't want to take chances and build hype around the architecture that succeeds it. From various sources, some faintly-reliable, we have been hearing that the next-generation of high-performance desktop processors based on "Piledriver" architecture, codenamed "Vishera", will pack five modules or 10 cores, and will be structured essentially like Zambezi, since Piledriver is basically a refinement of Bulldozer architecture. The latest leak comes from the Software Optimization Guide for AMD 15h family (read here), which was picked up by CPU World while most of us were busy with CES.

CPU World compiled most of the features of what it suspected to be AMD referring to its future processors based on the Piledriver architecture, that's "Vishera" (desktop high-performance), "Terramar" (high-density server), and "Sepang" (small-medium business server) parts. The three are not the first chips to be based on Piledriver, AMD has a new mainstream desktop and notebook APU in the works codenamed "Trinity", which is en route for a little later this year. Trinity basically has an identical CPUID instruction-set as Vishera, Terramar, and Sepang, confirming their common lineage compared to today's "Bulldozer" architecture. The most catchy detail is of Vishera featuring 4 DDR3 channels.

The plot thickens where "HyperTransport Assist feature" is listed as being supported on Vishera. HT Assist is a feature found on AMD's enterprise socket G34 processors, which facilitates better inter-die communication between the two dies of a typical socket G34 Opteron processor. The G34 (LGA1972) package is a multi-chip module of two quad-core, six-core, or four-module dies, which combined have four DDR3 memory channels, and a number of HyperTransport links to communicate with neighbouring sockets and the system's chipset. Could this be the first indication that AMD wants to take on Intel LGA2011 HEDT (high-end desktop) using Vishera chips in the G34 package? It will be a while before we find out.

Apart from using common silicon between client and enterprise platforms, AMD does have a history of colliding the two.Source: CPU World
Add your own comment

229 Comments on AMD Vishera Packs Quad-Channel DDR3 IMC, G34 En Route Desktop?

#1
sergionography
by: xenocide
You should be on AMD's Marketing team with that goofy logic. I can see it now; No, No, it's not a 5% performance gain, it's a 40% gain across the whole chip!
idk how you come up to such a conclusion, I'm atualy being as technical as possible to be in any marketing team lol, maybe if you follow up with the math you would understand, also note I'm speculating based on 30% performance gain per core and how it affects multicore performance in theoretical computation, note that i also ignored turbo core which might be part part of that 30% increase but i did so just to avoid confusion
And I stated this is total speculation and optimism as I hardly expect pd to be this way, mainly because single thread enables turbo core, while multithread will not scale 100%on the cores due to the shared resources
Posted on Reply
#2
allnights
Thanks Aquinus and Sergionography

Well I must say I really did enjoy the posts by you two!! (not sarcasm believe me)

I myself have been around computers for what seems like an eternity and have also acquired a Computing Science Degree.

Now I have always been interested in the architecture of CPUs and GPUs and for a long time read threads regarding this. However and in recent times this has ceased to be both enjoyable and informative. Due to the never ending stone throwing by fanboys defending large and powerfu companies that would not give two hoots about them.

However today I read through this thread and I must say that the posts by Aquinus and Sergionography have been refreshing and a joy to read.

Personally, and I admittedly lean to AMD but not a fanboy, I love the introduction of any new architecture. Sure Bulldozer was not quite what I was hoping it would be but these things happen. These are phenomenally difficult things to design and create and the research into it is mind boggling. What I find amusing though is how many 'experts' come out of the woodwork with their bold claims, obviously they should be placing the majority of their focus into CPU design with intel!

I do agree that the BD design is not as bad as has been reported and made out as it is, to my mind, a prototype and the first iteration on what is to follow.

From what I have read concerning Piledriver is somewhat reassuring and came as a pleasent surprise and against what everyone has predicted.

Now these naysayers are still sticking to the doom and gloom predictions (ridiculous like that of AMD going bust) for the Vishera CPU.

I think that the refinement of the BD core in Trinity is just the beginning now and that with Vishera this will take another small step. Naysayers argue the case against these small steps but are you not getting arguments put across that in many scenarios BD is not that bad?

Secondly there is the case of the clock speeds too, another area of improvement and these are just the ones we can speculate on and in all honesty that is all anyone can do.

As with any upcoming chip all we can do is speculate and whatever the the possibilites are indeed possibilities, apart from the dreamers that speculate a little over the top. In fact in one thread, maybe here, on a previous ocassion soe chap by the name of JF-AMD I kind of reprimanded for something he quoted, cannot remember what it was, in that he could not possibly back his statement up. So not a fanboy as I stated.

Also Ivy Bridge I fully expected to take a giant leap forward making it harder for Piledriver but this has not really happened in the way I expected.

I have no idea why but these intel fanboys need to stop throwing stones at the house of AMD, they will not make a blind bit of difference so why do it? Also AMD are the reason their 'PRECIOUS' Sanby Bridge Chips as it would likely be 2020 before you saw those (little joke their but you catch my drift).

I like the existence of both companies and I hope it continues, though I long for the day I would see AMD on a level playing field, but then I also long for the day I see England win the World Cup. We can but dream and maybe, just maybe things ay get close to what we want.

Personally I now do believe that that maturing of the BD architecture will continue unabated now the design is out. I would also like to see a die-shrink too as I believe with that along with other 'changes' will make a big difference.

But hell I have been wrong and disappointed before, LOL.

Thank you Sergionography and Aquinus for the discussions you put to this thread. It was a real pleasure to see an intelligent discussion for once.
Posted on Reply
#3
lebronjames316


AMD might be a little sore that its "Zambezi" FX processor family based on its much-hyped "Bulldozer" architecture didn't quite meet the performance expectations of a ground-up new CPU architecture, but it doesn't want to take chances and build hype around the architecture that succeeds it. From various sources, some faintly-reliable, we have been hearing that the next-generation of high-performance desktop processors based on "Piledriver" architecture, codenamed "Vishera", will pack five modules or 10 cores, and will be structured essentially like Zambezi, since Piledriver is basically a refinement of Bulldozer architecture. The latest leak comes from the Software Optimization Guide for AMD 15h family (read here), which was picked up by CPU World while most of us were busy with CES.

CPU World compiled most of the features of what it suspected to be AMD referring to its future processors based on the Piledriver architecture, that's "Vishera" (desktop high-performance), "Terramar" (high-density server), and "Sepang" (small-medium business server) parts. The three are not the first chips to be based on Piledriver, AMD has a new mainstream desktop and notebook APU in the works codenamed "Trinity", which is en route for a little later this year. Trinity basically has an identical CPUID instruction-set as Vishera, Terramar, and Sepang, confirming their common lineage compared to today's "Bulldozer" architecture. The most catchy detail is of Vishera featuring 4 DDR3 channels.
Posted on Reply
#4
Inceptor
by: lebronjames316
http://www.uuom.com/cs/images/signature_Amz.jpg

AMD might be a little sore that its "Zambezi" FX processor family based on its much-hyped "Bulldozer" architecture didn't quite meet the performance expectations of a ground-up new CPU architecture, but it doesn't want to take chances and build hype around the architecture that succeeds it. From various sources, some faintly-reliable, we have been hearing that the next-generation of high-performance desktop processors based on "Piledriver" architecture, codenamed "Vishera", will pack five modules or 10 cores, and will be structured essentially like Zambezi, since Piledriver is basically a refinement of Bulldozer architecture. The latest leak comes from the Software Optimization Guide for AMD 15h family (read here), which was picked up by CPU World while most of us were busy with CES.

CPU World compiled most of the features of what it suspected to be AMD referring to its future processors based on the Piledriver architecture, that's "Vishera" (desktop high-performance), "Terramar" (high-density server), and "Sepang" (small-medium business server) parts. The three are not the first chips to be based on Piledriver, AMD has a new mainstream desktop and notebook APU in the works codenamed "Trinity", which is en route for a little later this year. Trinity basically has an identical CPUID instruction-set as Vishera, Terramar, and Sepang, confirming their common lineage compared to today's "Bulldozer" architecture. The most catchy detail is of Vishera featuring 4 DDR3 channels.
I think that's a bit old. Five modules, or Ten cores, is what they were originally planning; they've scaled it back to four modules, as far as I know.
If Trinity is really showing a 15% improvement over Zambezi, then they would have to go to quad memory controllers to increase that bandwidth and give them a few more percentage points in increases in benchmarks, in addition to more 'aggressive' stock clock speeds. I think it's looking to be in the 20-30% total performance increase zone. Which seems to be where they wanted to be, one year ago.
Posted on Reply