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
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.
Apart from using common silicon between client and enterprise platforms, AMD does have a history of colliding the two.
229 Comments on AMD Vishera Packs Quad-Channel DDR3 IMC, G34 En Route Desktop?
Blah blah blah.
I don't think there will be any consumer desktop cpu from AMD after Piledriver. It'll be APUs.
In 2014, the 4th gen "real" fully 'fused' APUs get released, at which point, I think that'll be all that is available for desktop and laptop processors from AMD. APUs will also eliminate the x4xx, x5xx, and x6xx GPUs, and it'll be a three tier system of x7xx/x8xx/x9xx performance level discreet GPUs.
'Steamroller' integer cores will be in the 3rd gen 28nm APUs (Kaveri) and Opterons.
'Excavator'?? integer cores will be in the 4th gen and Opterons.
As far as PD wouldn't you consider the 5800k improvements a good example of what Vishera will be?
A10-5800k clock at 3.8ghz vs A8-3850 clock at 2.9ghz & the A10-5800k gets only 2.264sec better than the A8-3850 in SuperPI with 900mhz advantage i call this a big , big fail :shadedshu
If the A10-5800k did 2.264sec better but was clock at around 2.6-2.9ghz than i would said it is better but no it's WORST!!!
AMD CPU's performance is getting down all the time they should just give up & concentrate on GPU's only :o :rolleyes:
You are thinking way too theoretically that 30% per core is 240% improvement overall added to the CPU, and that is true but not insane, let me explain, if one core is 100% then 8cores is 800% right? Add to that 240 u end up with 1040%, 800/1040=0.76, so bulldozer will have 76% the performance of pd in multithread(assuming 30% percore increase) so the 240% overall computational increase that u mentioned sounds crazy because u are comparing it to one core but when compared the whole chip its consistent with the single core increase
Clock speed is only one part of thread performance... Memory, cache latency, NB and even HDD can affect per thread performance. So you can't expect a 1/1 ratio for improvement from CPU clock speed on any CPU.
My old PII 925 with an 900mhz overclock get a good 5sec better then original speed on Spi
The A10-500k with 900mhz advantage barely get 2sec better the the A8-3850
If both were at same clock i am pretty sure the A10-5800k will fall behind by a good 3-5sec
We're back in the P4 days except now it's AMD turns...
Granted, on the desktop where I would be wanting to overclock, that information about clock speed does become relevant. As an example, if processor A beats processor B by 10% when running at stock clocks, but can only overclock 3%, while B can be overclocked 25%, I'd be more inclined to pick processor B. So basically, if the Piledriver cores are clocked up high, and can't overclock much more, I'd pick the option that had the more headroom for ultimately more performance (Sandy vs Ivy is kind of in this situation right now). And IMHO on the desktop, performance / watt is not nearly as huge a concern.
Bulldozer FX chips have huge overclock potential. I gained 40% on passmark with fx-6100 @4.7ghz stable on cafa70. But I overclock everything that I can. The less cores, the higher clocks people are getting. The FX-4100 hits 5ghz on $30 coolers regularly. Intel locks their chips to prevent higher speeds on cheaper cpu's!
Also, Bulldozer chips hit higher clocks then 1155 chips when paired with the same cooler.
There isn't enough info yet to say that trinity's clocks are as a result of lower IPC, Thus misleading.... Remember Memory, NB and HT I think are all the same on Llano and Trinity. Contrary to what people are posting at 2.2ghz trinity is more efficient but when Trinity hits 4.2 efficiency is the same for FP and Int as Llano. And since Llano doesn't hit 4.2... It's hard to say for sure.
On the FP and Int tests the 5800k is 12% or so behind the 2500k without the benifit of L3 cache.
news.softpedia.com/news/AMD-Trinity-Field-Test-Shows-FPU-Performance-Superior-to-Bulldozer-263649.shtml
SO will Vishera's overclock potential be leveled compared to bulldozer???:eek:
I doubt it, Llano can't be tested at 4.2, If it could be I'm sure Trinity would have better IPC, since it does at 2.2. I think though, that the PD FX-6300 will have to hit 5.2Ghz stable on my same cooler to not be a dissapointment, compared to my fx-6100, if Vishera's improvment matches Trinity. My fx-6100 loads windows @ 5.2 but gets errors and overheats fast. So I'm optimistic with RCM.
I would go 2011 over Ivy for sure, if i needed the power, but even the fx-6100 is too much for me, for the price I paid I figured I would find something to use it for. I'm out to post more Hi-def Baby Video's to Facebook then anyone in history....:)
IPC is the defacto measurement of performance, and low power consumption without performance to meet expectation is the reason we don't use such low power chips in anything.
Don't get all butthurt over AMD failing on the CPU aspect, considering Intel spends more per year on R&D and they started as a second source supplier to Intel the fact they are still around and making chips that perform as well as the (some) do is amazing.
I still like to tweak and play with them, its a hobby, and really unless you are trying to play the benchmark game AMD performs just fine, and no one will tell the difference.
I feel the clock speed is what people still look at today or the higher or lower model numbers and prices. So if People can get an AMD for cheaper than an Intel they will go for it, Average Joe's are the Majority aka almost 99% of the Market the 1 percent is enthusiast, Aslong at it can turn on/off Play audio/video/ run the tasks they do or even play games they dont care about the CPU just its price.