[Coolaler] AMD Piledriver FX Engineering Sample Benchmarks

[oceanside]

Guys remember that the PD CPU being tested is the lowest tier 8 core. There are actually 2 more steps above this CPU. So compare to a 2500k at most IMO. If they manage to fix the power consumption problem while gaining an overall 10% in singlethreaded performance, I think it will be good, especially for upcoming Windows 8.

If they can do that, it won't be a bad chip at all. Now Windows 8.... It's going to be a monumental task to prevent it from going over like Unity and Gnome Shell 3. Looks like the iPhone generation of GUI developers is upon us for the desktop (Lord help us).

 

[jhu]

Not too bad. Let's see how they price them.

 

[exar333]

So an 8-core PD is roughly equivalent to the 6-core PhII and everyone is singing praises? Am I missing something? That is per the CB screenshot where the PD is slightly below in performance, but also slightly lower in freq. And that's with a 125w TDP.

 

[BallaTheFeared]

In before:

Drivers

Microcode

Bios

Fake!

Windows 8


Did I miss any?

 

[Rickyyy369]

I am extremely skeptical that these scores would be indicative of the final PD chips. Looking at the 3dMark06 scores, those benchmarks are claiming a score of 4986 for a 3.3ghz PD, yet a 8120 @3.1+turbo scores 5452, and an 8150 scores 5900.

Smells pretty fishy to me.

 

[MustangSVT]

poor amd.. is it over?

 

[ShintaiDK]

...lol, it's fake...( or a bug in ES )

PD have worst IPC than BD in that winRAR benchmark

I wouldnt put anything into the Winrar. Its too dependent on outside factors. Specially memory speed.

 

[ShintaiDK]

In before:

Drivers

Microcode

Bios

Fake!

Windows 8


Did I miss any?

Ye, you missed Steamroller. Seeing that as an excuse already lol.

 

[Borealis7]

what's more interesting is how high does it overclock and it's O/Ced performance and power consumption. BD was a massive clocker but also very power hungry.

 

[ShadowVVL]

why does it say 2008 in the lower right corner?

 

[inf64]

From XS' The Stilt. 2nd quote.

Extended model <10h = Orochi (Bulldozer), 10-1Fh = Trinity/Vishera.

The displayed by CPU-Z are raw values read from CPUID register, does not matter if the software supports the chip or not.

That is not a Vishera.
It is just the Bulldozer C0 spin.

There is quite a big chance this is not Vishera at all. The guy who said above is NOT clueless about this stuff.

 

[ShintaiDK]

why does it say 2008 in the lower right corner?

Just BIOS clock not set. The benching software wasnt avaliable by then if thats your thought.

 

[nemesismk2]

I'd like to see a little more than just Cinebench scores.

I don't rate cinebench very much. i want to see scores for games, novabench, geekbench and nuclearus. I am waiting for an official review from Anandtech.

 

[BD231]

Nothin that starts off with the name " Pile " is goin into this rig.

 

[sangyup81]

MOAR benchmarks

Why does the CPU-Z say Zambezi instead of Vishera?

 

[ShadowVVL]

Just BIOS clock not set. The benching software wasnt avaliable by then if thats your thought.

ah.yeah i dont know why i thought that lol.

Why does the CPU-Z say Zambezi instead of Vishera?

might be cpu z error.

 

[inf64]

I am extremely skeptical that these scores would be indicative of the final PD chips. Looking at the 3dMark06 scores, those benchmarks are claiming a score of 4986 for a 3.3ghz PD, yet a 8120 @3.1+turbo scores 5452, and an 8150 scores 5900.

Smells pretty fishy to me.

Fishy indeed. FX8120 scores 5450 in 3dmark06 CPU subtest. If this ES was indeed PD based it is slower than Zambezi which doesn't make any sense. Note that this ES has 3.3/3.7Ghz stock/all cores Turbo which is higher than what FX8120 has (3.1/3.4Ghz stock/all cores Turbo). Even 5800K scores 4300pts in 3dmark06 and it has only 2 modules versus 4 in this ES.
Also superpi 32M score is worse than FX8120's score (both at stock). This ES has higher one core Turbo since both stock and full core Turbo are higher(should be 100-200Mhz higher), so this is another anomaly.

 

[-Slacker-]

Are these pi numbers good or bad?

 

[MisterMac]

Are these pi numbers good or bad?

24 minutes when a 8150 does it at stock @ 19,3 minutes yes.
It's absolutely horrible.


Anyone extremely scared about how many blind fanbois see that coolaler thread as the holy grail for AMD @ overclock.net?

Damn it's scary.


(Granted we have our TSX\AVX2 fanbois too here, but atleast generally we like to use some logical theorectical max\fact as evidence).

 

[MisterMac]

Damnit, what is the clock on this thing?

Is this ES really 1.4 ghz?

 

[boxleitnerb]

No, that is the idle clock after the benchmark.

 

[-Slacker-]

24 minutes when a 8150 does it at stock @ 19,3 minutes yes.
It's absolutely horrible.

Wouldn't that suggest that there is something wrong with these results?

 

[Rvenger]

Ok . Good post . 1 question tho . whats windows 8 going to do for AMD pile ? Also did you notice that this 3rd tear cpu was using 124 watts. Intels are 95 watts . So If intel wanted to they could bring out a 125 watt. IB./ For some reason everyone wants to compare to intels last generation yet again . really a sad lot.

You honestly think CPU-Z is reporting correctly?

 

[Arzachel]

From these benchmarks it seems that AMD lost all the design and technical documents on Piledriver after releasing Trinity and resorted to putting their engineering team in a cage with an electron microscope, a Trinity chip, lifetime supply of Jaegermeister and a parrot. In the end, they produced something similar to a Piledriver module if you squint a bit and the parrot is now the lead engineer due to winning a drunken engineering swearoff.

That or this ES just might not be the final one 😉

 

[ShintaiDK]

That or this ES just might not be the final one 😉

When was the last time that ES samples performed differently from retail? Never?