AMD Piledriver rumours ... and expert conjecture

[Reynod]

We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...

 

[cgner]

Well sonny if "U" don't like reading it, I suggest "U" stop coming here 😛.

I've been a member here since 2006, and I link news and reviews. What exactly do "U" bring to the table, eh?

Everything you ever say comes down to "AMD sucks." Your news and reviews don't bring anything helpful to the discussion. I guess u just like trolling people. Oh well.

 

[de5_Roy]

Be cool if they did but i certainly wouldn't blame Intel for not allowing it. Also this is the only site i even go to for forums, The people here are a lot better then other sites. Oh yeah and the mods are way cooler, i still can't believe overclock.net block leaks before the NDA was up on bulldozer. And i find AnandTech to be nothing but a whole butch of Intel fanboys including most of the writers. Tomshardware writers/mods have a nice balance to them.

i see a few amd fanboys on at and other sites still pitching bd and spreading amd-glory. they use the same id so it's easy to spot them. amdzone is funnier. 😀

how much advantage will cpa (common platform alliance) bring to manufacturing cpus? cpa will have samsung, glofo and ibm as major partners.
http://semiaccurate.com/2010/10/14/common-platform-alliance-key-foundry-success/
cpa and tsmc might be able to take on intel.

intel will never allow anything on their fabs that does not enhance their own business. that's why they'd never allow apus.
i suppose they could have allowed bulldozer dt and server cpus on their fabs, look how many customers they brought in for intel. 😛

some moar steamroller, other sites talk more about papermaster's speech on steamroller. more or less the same thing.
http://vr-zone.com/articles/amd-pushes-steamroller-and-excavator-forward-bullish-about-performance-increases/17088.html
i lol'ed at this -

"Steamroller is not Bulldozer Enhanced. F*** no. The layout might look the same but our LEGO blocks are completely different. When all is said and done we should get 45% improvement and this goes to show how the Bulldozer was f***** design. This is all what Bulldozer was supposed to be."

This is a direct quote from our source which will remain anonymous, but is important enough to be awake at 3:32AM European time (the source is in the obamaland). The bullish statements coming from engineers are nothing uncommon, as the technology and products are nothing less than their own babies, male or female.

take the 45% figure and the quote with a large grain of salt but.. when amd's own says bd was !@#$ed, you know it's !@#$ed, and how amd suckered a bunch of hopefuls (except the ones who bought it despite knowing everything) into buying [strike]a subpar[/strike] an intermediate product. 😀

http://hexus.net/tech/news/cpu/44429-amd-steamroller-details-hot-chips-symposium/

http://hothardware.com/News/AMDs-Next-Gen-Steamroller-CPU-Could-Deliver-Where-Bulldozer-Fell-Short/

http://www.rage3d.com/articles/amd_hot_chips_2012/

http://techreport.com/articles.x/23485

 

[cgner]

i see a few amd fanboys on at and other sites still pitching bd and spreading amd-glory. they use the same id so it's easy to spot them. amdzone is funnier. 😀

how much advantage will cpa (common platform alliance) bring to manufacturing cpus? cpa will have samsung, glofo and ibm as major partners.
http://semiaccurate.com/2010/10/14/common-platform-alliance-key-foundry-success/
cpa and tsmc might be able to take on intel.

intel will never allow anything on their fabs that does not enhance their own business. that's why they'd never allow apus.
i suppose they could have allowed bulldozer dt and server cpus on their fabs, look how many customers they brought in for intel. 😛

some moar steamroller, other sites talk more about papermaster's speech on steamroller. more or less the same thing.
http://vr-zone.com/articles/amd-pushes-steamroller-and-excavator-forward-bullish-about-performance-increases/17088.html
i lol'ed at this -

"Steamroller is not Bulldozer Enhanced. F*** no. The layout might look the same but our LEGO blocks are completely different. When all is said and done we should get 45% improvement and this goes to show how the Bulldozer was f***** design. This is all what Bulldozer was supposed to be."

This is a direct quote from our source which will remain anonymous, but is important enough to be awake at 3:32AM European time (the source is in the obamaland). The bullish statements coming from engineers are nothing uncommon, as the technology and products are nothing less than their own babies, male or female.

take the 45% figure and the quote with a large grain of salt but.. when amd's own says bd was !@#$ed, you know it's !@#$ed, and how amd suckered a bunch of hopefuls (except the ones who bought it despite knowing everything) into buying [strike]a subpar[/strike] an intermediate product. 😀

http://hexus.net/tech/news/cpu/44429-amd-steamroller-details-hot-chips-symposium/

http://hothardware.com/News/AMDs-Next-Gen-Steamroller-CPU-Could-Deliver-Where-Bulldozer-Fell-Short/

http://www.rage3d.com/articles/amd_hot_chips_2012/

http://techreport.com/articles.x/23485

I would not be surprised if AMD's top was deeply disappointed with BD. So long in development and its weaker than previous archie. Really now? AMD has kicked azz in the past. If they get their things together somehow for SR, it will be an interesting showdown.

 

[mayankleoboy1]

AMD has started a rapid-release-cycle. But the people who buy a BD or PD are screwed.

 

[newb88]

anyone here buy processors in bulk? If yes, is it possible for me to join in?

 

[cgner]

AMD has started a rapid-release-cycle. But the people who buy a BD or PD are screwed.

Rapid release could mean rapid progress, or as in case with EA games, halfassery. Judging by dozer, i would say its the latter. We will see. Hopefully they will catch up.

 

[esrever]

AMD has started a rapid-release-cycle. But the people who buy a BD or PD are screwed.

an incremental improvement every year isn't rapid at all...

 

[truegenius]

AMD Vishera six cores detailed
http://www.fudzilla.com/home/item/28569-amd-vishera-six-cores-detailed

 

[fazers_on_stun]

well, even with the power throughput of PCIe 3 up to 150 watts, I'm still willing to bet it will take 2 8pins, maybe more. It does have 2x as many transistors of a 680, which draws nearly 200 watts.

I'm really not sure about this chip for desktop use.

Well most of us would like to see it benched .

I just hope that it doesn't exhaust the hot air inside the case, like NV and AMD do with the top dual GPU cards nowadays. That's just plain silly from an engineering viewpoint - blows right on the hard drive bay and forces the case fans to then get rid of the excess heat..

 

[fazers_on_stun]

And regarding the manufacturing process chat: not a chance in hell Intel would give AMD access to their Fabs. At least, not without a court order or something, lol. Even more when the strongest part of their advantage is process.

Cheers!

Well it wouldda been something if AMD had stuck it in the settlement agreement anyway 😀. Somehow I doubt the subject even came up though - AMD probably had way too much confidence in GF back in 2009.

 

[fazers_on_stun]

how much advantage will cpa (common platform alliance) bring to manufacturing cpus? cpa will have samsung, glofo and ibm as major partners.
http://semiaccurate.com/2010/10/14/common-platform-alliance-key-foundry-success/
cpa and tsmc might be able to take on intel.

Hmm, that article dates from about 2 years ago - the CPA doesn't appear to have done what it was supposed to do, since GF and other fabs have abandoned gate-first HGMG, in favor of gate-last. Also dunno about those "200mm" wafers mentioned throughout the article - 300mm have been the norm for some time now .

some moar steamroller, other sites talk more about papermaster's speech on steamroller. more or less the same thing.

Heh, well we've seen the same mantra from AMD for the last 5-6 years now: "OK our current CPUs can't compete with Intel, but just wait until ______ a year or so from now!!" Hopefully even the fanboys have learned by now and will take a "prove it or shaddap" attitude. And interesting that the one engineer implied PD is not going to be that wonderful. What does he think that'll do for sales in the next 15 months, considering that's what funds his paycheck?

 

[mayankleoboy1]

it would have made more sense for AMD to stop PD and directly roll with SR. Trinity could have been PII+GCN.

 

[obsama1]

And interesting that the one engineer implied PD is not going to be that wonderful. What does he think that'll do for sales in the next 15 months, considering that's what funds his paycheck?

Source? I don't doubt you, but I'd like to know more about this.

 

[fazers_on_stun]

^ From De5roy's link on the previous page: http://vr-zone.com/articles/amd-pushes-steamroller-and-excavator-forward-bullish-about-performance-increases/17088.html

 

[davemaster84]

it would have made more sense for AMD to stop PD and directly roll with SR. Trinity could have been PII+GCN.

I don't think that would have made sense at all, why a company such as AMD would have throw away a whole architecture gen and bring up the next generation with all the investment they could have to use? Piledriver seems to be a very good architecture and for sure it'll be finally a worth upgrade after Phenom II , we don't have any benchmarks yet to think this is true or not, but what we do know is that AMD improved power consumption which will result in higher OCs as well.

 

[davemaster84]

According to Overclock.net PD will be performing in the same way SB does, or at least that is their conjectures:

http://www.overclock.net/t/1291114/coolaler-amd-piledriver-fx-vishera-engineering-sample-benchmarks

 

[fazers_on_stun]

it would have made more sense for AMD to stop PD and directly roll with SR. Trinity could have been PII+GCN.

Think AMD's too far along with PD to just abandon it, plus 15% improvement is nothing to turn one's nose up at - IB was only ~5% improvement over SB. Plus BD's almost a year old by now, and AMD needs some sort of product refresh or else risk losing more marketshare, as Steamy won't be out until Q4 of next year IIRC.

And it may take AMD another 15 months to fix all the problems - seems like a pretty significant redesign effort. Plus it took them 5 years to come up with BD to begin with (counting from when BD was first mentioned on AMD roadmaps - 2004 if memory serves).

Anyway, AMD has scrapped any further development of K10.5 (PII) and is rolling the dice with BD derivatives from now on. They don't have Intel's R&D budget to pursue multiple technologies aimed at the same segment I guess.

 

[cgner]

we've seen the same mantra from AMD for the last 5-6 years now: "OK our current CPUs can't compete with Intel, but just wait until ______ a year or so from now!!"

I can certainly agree with that. They need to stop BSing and start developing.

I don't think that would have made sense at all, why a company such as AMD would have throw away a whole architecture gen and bring up the next generation with all the investment they could have to use? Piledriver seems to be a very good architecture and for sure it'll be finally a worth upgrade after Phenom II , we don't have any benchmarks yet to think this is true or not, but what we do know is that AMD improved power consumption which will result in higher OCs as well.

It will hardly sell well though. They will be losing time and money by keeping it on the market, imo. Also, a minor upgrade from PII after 4 years in development is pretty weak. I mean, last time I kept Pentium D840 for 5 years, the performance quadrupled. Either both companies are hitting plateau of performance, or AMD is falling behind and Intel sees no reason to sweat and work faster/harder.

According to Overclock.net PD will be performing in the same way SB does, or at least that is their conjectures:

http://www.overclock.net/t/1291114/coolaler-amd-piledriver-fx-vishera-engineering-sample-benchmarks

The question is; is it going to be faster?

 

[viridiancrystal]

it would have made more sense for AMD to stop PD and directly roll with SR. Trinity could have been PII+GCN.

That is half of what Llano was. PII cores minus L3 cache. Trinity uses PD, and it beats it across the board iirc.

 

[truegenius]

pd in trinity is with rcm
and if they can use rcm in k10 then it can perform better than pd in performance per watt comparisions

 

[esrever]

pd in trinity is with rcm
and if they can use rcm in k10 then it can perform better than pd in performance per watt comparisions

very unlikely that AMD would be able to clock k10 cores up to a frequency where RCM actually does enough. The target frequency for RCM is 3-4ghz.

 

[fazers_on_stun]

Still haven't heard anything about whether resonant clock mesh is effective at frequencies other than stock (where the mesh is tuned). Even with a low "Q" a tuned circuit would be ineffective too far outside the center frequency. So I'd think AMD would have to switch out segments of the inductor loops according to the frequency if they wanted to overclock (turbo) or underclock (idle).

 

[obsama1]

IMHO, AMD shouldn't have launched the 4130, when the 4170 is $10 more.

 

[king smp]

AMD Rumored to Discontinue Performance CPU’s after Piledriver – Will Focus on APU development
http://wccftech.com/amd-rumored-di [...] velopment/

Rumor: AMD CPU canceled their Vishera?
http://technewspedia.com/rumor-amd [...] r-vishera/

Phil Hughes ‏@AMDphil
https://twitter.com/AMDphil/status/238044301588848643
Quote :


we don't comment on specific rumor/speculation but in terms of Vishera and other 2012 products,
there have been no changes

 

[esrever]

I highly doubt AMD will cancel AM3+ after vishera. It would be easy for them to sell a few cut down, lower binned, higher clocked server chips and make them AM3+. I mean they aren't changing server sockets until DDR4 and thats still 2-3 years away for AMD.