AMD "Piledriver" refresh of Zambezi - info, speculations, test, fans

[FlanK3r]

if Crosshair VI will be better than Crosshair V, then maybe will upgrade of the motherboard

[imamage]

Since they have upgrade the SB1050 to have Native USB 3.0 support
I probably go from Crosshair IV Formula to Crosshair VI Formula (If ASUS going to make one ...lol )

But until Piledriver have improve over BD , I don't see upgrading my 1090T to 8150 as necessary

[FlanK3r]

I think, these are iGPUs for Bobcat refresh and for Trinity

[Smartidiot89]

I think, these are iGPUs for Bobcat refresh and for Trinity

Correct!

[TESKATLIPOKA]

Here is something but without Cinebench
http://www.hardware.fr/articles/842-...acite-cmt.html
and here is another performance chart
http://www.hardware.fr/articles/842-...s-3-2-ghz.html

If the info about ~3.5b in Cinebench using Trinity 3.8Ghz is true then its not bad.
FX4100(4.62Ghz) has 3.78b
FX4100(3.6Ghz) has 2.97b
http://www.legitreviews.com/article/1766/9/
1Ghz added 0.8b so you can say Trinity at 3.8Ghz is as powerful as BD at ~4.2-4.3Ghz with 8MB L3 cache.
I would like to know how much is L3 affecting the scores.

[informal]

Here is something but without Cinebench
http://www.hardware.fr/articles/842-...acite-cmt.html
and here is another performance chart
http://www.hardware.fr/articles/842-...s-3-2-ghz.html

If the info about ~3.5b in Cinebench using Trinity 3.8Ghz is true then its not bad. Cinebench is really one of those client SIMD workloads that doesn't run that well on new FlexFP and if PD(with no L3) improves the score by ~10% vs BD with L3 on same clock then good job AMD. L3 in Vishera may bring few more percent. But before we get excited over some IPC improvements we should wait for real data. For now we have a vague statement from one Chinese guy,that's all.
FX4100(4.62Ghz) has 3.78b
FX4100(3.6Ghz) has 2.97b
http://www.legitreviews.com/article/1766/9/
1Ghz added 0.8b so you can say Trinity at 3.8Ghz is as powerful as BD at ~4.2-4.3Ghz with 8MB L3 cache.
I would like to know how much is L3 affecting the scores.

If we are to judge by the previous cores from both AMD and intel,L3 adds another 3-5% worth of performance (approximately of course;somewhere it may add more,somewhere nothing).
If Piledriver without L3 is ~10% faster than BD with L3,then good job AMD. But before we get excited about IPC improvement(however small it is),we should wait for real data. For now all we have is a vague statement from one Chinese guy.

[undone]

Here is something but without Cinebench
http://www.hardware.fr/articles/842-...acite-cmt.html
and here is another performance chart
http://www.hardware.fr/articles/842-...s-3-2-ghz.html

If the info about ~3.5b in Cinebench using Trinity 3.8Ghz is true then its not bad.
FX4100(4.62Ghz) has 3.78b
FX4100(3.6Ghz) has 2.97b
http://www.legitreviews.com/article/1766/9/
1Ghz added 0.8b so you can say Trinity at 3.8Ghz is as powerful as BD at ~4.2-4.3Ghz with 8MB L3 cache.
I would like to know how much is L3 affecting the scores.

Today I have time to calculate something about trinity.
According to the list of hardware.fr, bulldozer CMT brings average about 50% increase.
a stands for single core of Llano, b stands for trinity(an integer core):

a * 4 cores = b * (100+50)% * 2 modules

result:
b = 1.33a

Trinity equals to 4Ghz Llano per core(not module, don't mix up).

and remember that trinity is ES, something may be fixed before launch just like B0(3.6Ghz) and B2(4.2Ghz) Zambezi. I won't calculate anymore just take B0 and B2 result for reference. It's long time ago I couldn't remember clearly the differences between the retail chip and ES ones.

[TESKATLIPOKA]

informal As you said, It can be false so lets wait for more leaks.

Trinity cpu In desktop may not be faster than Llano but in notebooks it should be if it can clock much higher and stay in the same TDP.
desktop(lets say they have the same performance)
Llano 2.9Ghz
Trinity 3.7Ghz +27.5% frequency
notebook
Llano 1.6Ghz
Trinity 2.4Ghz +50% frequency
mobile Trinity will end up considerably faster just because mobile Llano has low clocks.

[TESKATLIPOKA]

Opteron 3260 EE 2module 2.7GHz and TDP is just 45W. I think at least in notebooks Llano won't stand a chance against Trinity.

[TESKATLIPOKA]

double post

[TESKATLIPOKA]

undone What you wrote is utterly wrong if you thought about it a bit you would realize its a nonsense for Trinity or BD to be 33% faster than Deneb(Llano) core to core. It would be true only if the frequency and performance was the same for both of them not to mention you added an incorrect CMT increase which resulted in additional IPC improvement.

average performance from http://www.hardware.fr/articles/842-...s-3-2-ghz.html
Deneb(3.2Ghz) 4C 100
BD(3.2Ghz) 4M/4C 86.7
BD(3.2Ghz) 2M/4C 75.5

How much slower is BD 2M/4C vs BD 4M/4C
75.5/86.7=0.871
Penalty decreases performance by 12.9%
0.871*200=174.2
As you can see 2M/4C is not 2*150 but 2*174.2

If you want to know how much faster is Deneb 4C vs BD 2M/4C or BD 4M/4C at the same speed
100/75.5=32.5%
100/86.7=15.3%
As you can see, clock to clock and core to core Deneb is faster by 15.3%.
Deneb 2.9Ghz equals BD 3.35Ghz in single threaded applications and in multi-thread BD must work at 3.85Ghz to be equal thanks to the sharing penalty.

Llano more or less equals Deneb and from the Chinese leak it looks as if Trinity has ~10% IPC increase over BD.
With 10% IPC increase it should look something like this
1 thread Deneb 2.9Ghz vs Trinity 3.05Ghz
multi-thread Deneb 2.9Ghz vs Trinity 3.55Ghz.

[TESKATLIPOKA]

Trinity default 3.8Ghz will be faster on average than Llano, the only question is how much faster, the unknown factors are the IPC improvement vs BD and the frequency they will launch at, the minimum looks like 3.8Ghz.

I don't really care about the desktop, but am really interested in notebooks. The best would be if 2M/4C worked at >=2.4Ghz default, but based on the fact Opteron 3260 with TDP 45W works at 2.7Ghz I don't think its out of the question.

P.S. I wouldn't mind if AMD released Trinity with TDP 55W and higher clocks.

[informal]

To me it looks like PD was close to what AMD originally expected out of BDver1 but never managed to pull it off. Around 5-10% better IPC on average and 5-8% higher clock within similar TDP.

[TESKATLIPOKA]

informal I wouldn't say that, It's an improvement over BD but still under Deneb IPC and we all know originally It should have been better than K10.

[undone]

I wouldn't say that, It's an improvement over BD but still under Deneb IPC and we all know originally It should have been better than K10.

I never argue about ipc even before Zambezi being announced.
When you go over the history you'll find something similar to nowaday situation. K6 have 10% higher IPC than K7, netburst is terrible, and now bulldozer is the same. Reanson why they developed a lower IPC model is because the frequency is bottlenecked by architecture. These design always not only need tweak but more important is the process node, lately CPU bottlenecked below 4Ghz and now bulldozer makes a breakthrough.
Bulldozer = K7, and it would be another K8 when everything is OK, since then don't be surprise to see a CPU that stock at 6Ghz+.

[-Boris-]

I never argue about ipc even before Zambezi being announced.
When you go over the history you'll find something similar to nowaday situation. K6 have 10% higher IPC than K7, netburst is terrible, and now bulldozer is the same. Reanson why they developed a lower IPC model is because the frequency is bottlenecked by architecture. These design always not only need tweak but more important is the process node, lately CPU bottlenecked below 4Ghz and now bulldozer makes a breakthrough.
Bulldozer = K7, and it would be another K8 when everything is OK, since then don't be surprise to see a CPU that stock at 6Ghz+.

Wrong, K7 had much higher IPC than K6, K6-III was far behind P2 overall (not integer), and P3 was behind K7. And the frequency is not bottlenecked by architecture, all architectures have a hard time over 4GHz. So it's stupid to sacrifice lots of IPC and die size to squeeze a few hundred MHz more from the chip at these frequencies, every 100Mhz over 4GHz has a high price, AMD decided to pay up. I can't imagine Intels factories being capable of producing BD at competitive frequencies (5-7GHz with less power consumption). The problem is in the design, they decided to pay a high price to get some extra frequency, just the thing that killed Prescott.

No architecture with lower IPC than its predecessor has been successful. All the really successful architectures has had large gains in IPC, like Core i7, Core 2, Athlon 64, K7 and Pentium Pro.

[sergiojr]

And the frequency is not bottlenecked by architecture, all architectures have a hard time over 4GHz.

Completely wrong. Frequencies are bottlenecked by both the process (AKA transistor switching speed) and architecture (AKA number of sequentially placed transistors on critical path). And resulting processor frequency is a result of division of the first by the second. Bulldozer significantly shortens critical path and so it's frequency is much higher than 32nm Llano within the same power budget. Or you should expect ~4,1 Gz base clock for imaginery 6-core 45nm Bulldozer within the power budget of 1100T.

[-Boris-]

What I think would be perfect PD scenario for AMD is: ~5-10% higher IPC and 4-4.2Ghz base clock for top end model. This would put it in 16-28% range over 8150 ,a pretty good spot (over 2600/2700K on average and very close to 980/990x).

That's flawed logic. Being 30% behind in a bench is not the same thing as being 30% slower. BD needs much more than 28% more performance to match SB.

Completely wrong. Frequencies are bottlenecked by both the process (AKA transistor switching speed) and architecture (AKA number of sequentially placed transistors on critical path). And resulting processor frequency is a result of division of the first by the second. Bulldozer significantly shortens critical path and so it's frequency is much higher than 32nm Llano within the same power budget. Or you should expect ~4,1 Gz base clock for imaginery 6-core 45nm Bulldozer within the power budget of 1100T.

Of course architecture matters, but what you don't take in to consideration is that frequencygains isn't linear. And over 4GHz the sacfrifices you have to do to gain each MHz isn't worth it at this point. You can't say that Bulldozer is more efficient than K10 or Llano, Bulldozer is less power efficient tha K10 on 45nm! Your comparision to Llano doesn't work since Llano has an integrated GPU, you don't know how much power the cores in Llano consumes and you don't know how the GPU affects the cores power consumption. If llano is made on a different kind of silicon to make the GPU work good enough then that could cripple energy efficieny in the cores.

[informal]

informal I wouldn't say that, It's an improvement over BD but still under Deneb IPC and we all know originally It should have been better than K10.

No,I meant 5-10% better IPC over what we have with this Bulldozer.Not what we have with Deneb.

[TESKATLIPOKA]

undone and what does it have in common with my comment, I never said BD can't be tweaked .
I don't know why you think they planned for BD to have worse IPC than K10
If that was true then why was JF always saying the IPC will increase over K10 and he meant BDv1(bulldozer) not BDv2(piledriver) or BDv3(steamroller) because that was what he heard from the engineers but sadly It wasn't true, simply put BD didn't turn out like they planned.

[undone]

undone and what does it have in common with my comment, I never said BD can't be tweaked .

To me it looks like you'r saying IPC lower than deneb is a failure. But processor design is complicate than we thought.
Even if Zambezi doesn't have those problems I still doubt IPC would increase.

I don't know why you think they planned for BD to have worse IPC than K10
If that was true then why was JF always saying the IPC will increase over K10 and he meant BDv1(bulldozer) not BDv2(piledriver) or BDv3(steamroller) because that was what he heard from the engineers but sadly It wasn't true, simply put BD wasn't what they wanted.

IIRC, 'IPC increase' is due to some simply statements like '33% more cores but 50% increase compare to Thuban'. But they never mentioned anything about the balance between frequency and ipc.

Ok, I wanna end up discussion about BD, if my comment lead to off-topic then sorry.

[TESKATLIPOKA]

informal

No,I meant 5-10% better IPC over what we have with this Bulldozer.Not what we have with Deneb.

I know you meant IPC increase over BD, but I was referring to JF's old comments how BD's IPC will be better than K10's.

[TESKATLIPOKA]

undone

But they never mentioned anything about the balance between frequency and ipc.

JF said many times IPC will increase even if it wasn't the official statement from AMD.

To me it looks like you'r saying IPC lower than deneb is a failture. But processor design is complicate than we thought.

I don't care if the performance is gained from IPC increase or frequency, but if JF said IPC will increase and It turns out It decreased by >10% I think its a failure compared to what it should have been.

[TESKATLIPOKA]

Smartidiot89 Maybe we will see some increase even in integer performance. Probably some still remember Charlie's claim about 20% increase in ALU performance in the next stepping.

[Smartidiot89]

Smartidiot89 Maybe we will see some increase even in integer performance. Probably some still remember Charlie's claim about 20% increase in ALU performance in the next stepping.

I find it hard to believe with a stepping, even with a new revision I'd raise an eyebrow. I will take 10-15% increase with Trinity as that is what I've heard from AMD and that this guy from Chiphell shows in hes results. 20% i don't think will happen with Piledriver, not IPC-wise - but IPC+clocks i find it very possible, even very likely.

AMD have already gone public that they aren't happy with GlobalFoundries performance with 32nm, and the power consumption of Bulldozer (especially overclocked) points the finger towards high leakage. I think there's alot that can/will/has be done, both on the design and manufacturing side.