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 ...

 

[sarinaide]

IMHO OBR-Hardware is just shooting from the hip, I don't understand the CPUID screen, Zambezi and Vishera are different families both reflect as Zambezi. Even if PD DT was just a debugged Zambezi it would likely then be a OR-B2/B3 revision then, by simple terms its a OR-C0 which implies rework of basic layout it should then be Piledriver unless PD/DT is just reworked bulldozer cores while Trinity and Abu Dhabi are PD cores, or PD is using BD branch predictions....dunno seems odd, thats without getting into the synthetic side.

I think its photoshopped honestly.

 

[cgner]

I think its photoshopped honestly.

Same.

 

[Class B]

 

[cgner]

^Who did that "testing"? Source? Link?

 

[truegenius]

AMD Dual Richland 28nm has 192 shaders
http://www.fudzilla.com/home/item/28600-amd-dual-richland-28nm-has-192-shaders

^Who did that "testing"? Source? Link?

[strike]excel[/strike] internet 😗

 

[de5_Roy]

Eleven AMD Zambezi processors EOL in Q4 12
http://www.fudzilla.com/home/item/28601-eleven-amd-zambezi-processors-eol-in-q4-12
seems like a rather quick phase-out.
assumptions -
on positive side, seems like amd is confident that they can maintain a good supply of pd cpus and trinity apus.
stock performance and efficiency has improved.
sub $160-170 market will have more quad cores. not something high-performance, but quad core nonetheless.
trinity's overclocked performance might reach 6770 or higher levels.
trinity will not have to worry about competing against ivb core i3. only one sku with hd4k igpu which is less performing than 7660d.
on not-so-positive side,
oc power efficiency seems same as bd,
imc and pcie controller are still same (yawn).
ivy bridge core i3 will put up a strong showing (much rage to c.a.l.f.) in desktop gaming against 8 core pd. 😛

 

[Reynod]

Thanks for all of the links with the updated material guys !!



Keep it coming ...

 

[sarinaide]

Eleven AMD Zambezi processors EOL in Q4 12
http://www.fudzilla.com/home/item/28601-eleven-amd-zambezi-processors-eol-in-q4-12
seems like a rather quick phase-out.
assumptions -
on positive side, seems like amd is confident that they can maintain a good supply of pd cpus and trinity apus.
stock performance and efficiency has improved.
sub $160-170 market will have more quad cores. not something high-performance, but quad core nonetheless.
trinity's overclocked performance might reach 6770 or higher levels.
trinity will not have to worry about competing against ivb core i3. only one sku with hd4k igpu which is less performing than 7660d.
on not-so-positive side,
oc power efficiency seems same as bd,
imc and pcie controller are still same (yawn).
ivy bridge core i3 will put up a strong showing (much rage to c.a.l.f.) in desktop gaming against 8 core pd. 😛

It makes sense to quickly phase out Zambezi on the premise that manufacturing costs, I don't have any issues with them doing that. On Trinity I don't think AMD are actually under any pressure on iGPU performance, it seems at best HD5K will be competitive with Llano but still far behind where AMD will be at by the time Steamroller rolls out, if the talk is hybrid 7700 series integration that is a massive step forward in iGPU performance without factoring in asymetrical crossfire.

On Vishera desktop, from what I know its improved but some aspects are not that different from Zambezi but I think people will be very surprised by the time official stuff comes out.

 

[truegenius]

Thanks for all of the links with the updated material guys !!

Keep it coming ...

pleasure

trinity's overclocked performance might reach 6770 or higher levels.

6770 ???
i can expect for upto 6670 but 6700 is way ahead
at the time of purchase mine 6770 is was and is still expensive than top apu 🙁

 

[esrever]

^Who did that "testing"? Source? Link?

maybe it was OBR again, LOL

 

[esrever]

Thanks for all of the links with the updated material guys !!

Keep it coming ...

pleasure

trinity's overclocked performance might reach 6770 or higher levels.

6770 ???
i can expect for upto 6670 but 6700 is way ahead
at the time of purchase mine 6770 is was and is still expensive than top apu 🙁

OC'ed llano could do close to the 6670 level performance, I'd assume trinity will be somewhat faster. 6770 would be too much tho. Will be hard for it to even reach 6750 performance no matter how much OC imo.

 

[de5_Roy]

6770 ???
i can expect for upto 6670 but 6700 is way ahead
at the time of purchase mine 6770 is was and is still expensive than top apu 🙁

OC'ed llano could do close to the 6670 level performance, I'd assume trinity will be somewhat faster. 6770 would be too much tho. Will be hard for it to even reach 6750 performance no matter how much OC imo.

i know it's a bit much but...
i will dream!! 😍

 

[gamerk316]

I can't believe the only PD benches we have are from OBR. PD isn't too far off, I hope for a legitimate ES leak. OBR definitely doesn't have a PD chip.

Just like with BD, I see that as a VERY worrying sign.

 

[sarinaide]

That or AMD have decided to control those they trust with ES samples not to disclose, its nothing to worry about...

 

[-Fran-]

Even if it is a very early sample of a PD chip, the OBR numbers are not THAT bad. The thing that worries me looking at them is the bump in hertz and not in design changes.

As an OCer, I want to know clock per clock performance, not stock =/

Cheers!

EDIT: Forgot to say... If it is indeed an early sample, that means (and I hope, really) that latter iterations are actually better.

 

[obsama1]

Looks pretty good, but 2GB of RAM to benchmark? LOL, please. IF PD is just as good as SB, I want to see the Microcenter mobo bundles for this. 😀

 

[Blandge]

Please understand that ES samples are a VERY poor indicator of final performance. Early steppings can have debilitating defects, but even the most latest pre-production ES silicon will be using under-optimized early BIOS versions.

 

[fazers_on_stun]

it seems at best HD5K will be competitive with Llano but still far behind where AMD will be at by the time Steamroller rolls out

Links?

 

[sarinaide]

Considering the jump from 3K to 4K the margin was still big, there is no reason to assume that HD5K is going to be anything mindblowing irrespective if it has the name Intel on it.

 

[colinp]

Please understand that ES samples are a VERY poor indicator of final performance. Early steppings can have debilitating defects, but even the most latest pre-production ES silicon will be using under-optimized early BIOS versions.

That's what people said about the early BD benchmarks.

 

[Cazalan]

Sure is a long waiting game for some real numbers.

 

[cgner]

maybe it was OBR again, LOL

lol Maybe he went by a different name this time

Just like with BD, I see that as a VERY worrying sign.

That or AMD have decided to control those they trust with ES samples not to disclose, its nothing to worry about...

I personally don't see it as a good sign. Every time AMD keeps quiet about things till the last hour its because they got nothing good to bring to the market. Just saying.

 

[fazers_on_stun]

Considering the jump from 3K to 4K the margin was still big, there is no reason to assume that HD5K is going to be anything mindblowing irrespective if it has the name Intel on it.

OK, so just your opinion then. Not clear from your statement "it seems at best HD5K will be competitive with Llano but still far behind where AMD will be at by the time Steamroller rolls out" which implies some sort of objective evidence or at least a rumor. Opinion is usually prefaced by "I think" or "IMO" or something equivalent..

So I take it that you are not persuaded even a little by Intel's '7x' performance improvement over SB's iGPU, or S/A's or Anandtech's statements that Haswell will be a "graphics monster"?

Most enthusiasts hope both companies are competitive in both CPU and iGPU since that is what drives progress and innovation..

 

[Blandge]

OK, so just your opinion then. Not clear from your statement "it seems at best HD5K will be competitive with Llano but still far behind where AMD will be at by the time Steamroller rolls out" which implies some sort of objective evidence or at least a rumor. Opinion is usually prefaced by "I think" or "IMO" or something equivalent..

So I take it that you are not persuaded even a little by Intel's '7x' performance improvement over SB's iGPU, or S/A's or Anandtech's statements that Haswell will be a "graphics monster"?

Most enthusiasts hope both companies are competitive in both CPU and iGPU since that is what drives progress and innovation..

A surprising amount of enthusiasts have a vendetta against Intel graphics that originates from some time before Sandy Bridge was released. Even now, the standard answer the Intel's increasingly powerful iGPU is that the Intel's graphics drivers are terrible. This is no longer the case. There was a time when Intel graphics (hardware and driver) were almost unuseable, but since then Intel has produced a great iGPU solution.

Most who complain about Intel graphics have not used them on Ivy Bridge, and quote lackluster high resolution gaming benchmarks as proof that HD graphics are not suitable for even average users. However, the budget Intel CPUs provide a very capable CPU/GPU combination for the average user with enough 3D graphics power for casual gaming, and unmatched media performance for HTPCs.

Intel graphics have already more than doubled in processing power since Nehalem with much more driver stability. There's nothing that indicates that this trend will not continue with Haswell and beyond.

 

[dattimr]

That's what people said about the early BD benchmarks.

In the case of Zambezi, we knew that the ESs were based on Bulldozer. Also, there wasn't just a single leak, but rather dozens of them. (Mind you that OBR himself has stated that "his" results come from a chinese owner, who is probably the same guy from Coolaler.) This time, however, people cannot come to a conclusion as to whether what is being tested is indeed a Vishera ES, which should be based on Piledriver+, or an unreleased Zambezi stepping (which might turn out to be Vishera after all, but this seems highly unlikely).