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 ...

 

[triny]

Its not up to amd for this to happen they can make the hardware but they can't force software to change or expect it to change so quick its going to take time.

If your referring to HSA I imagine software companies will be on it like a tick on a dog as it simplifies their job and streamlines their business.
it's going to happen because we are nearing the end of the multi core era and it's the logical next step.
not just AMD Intel will do it too. It's where the PC becomes a super computer.

 

[de5_Roy]

desktop llano is not a bad product. it performs very well in it's class. things that hold it back from being a 'best' are - poor yields from glofo, top apus being power hungry and hot, top apus' high price compared to competition (sb pentium+ 6450/6570), (rumored) dead end platform. imo poor yields is what led to apple dumping llano and llano's less-than-widespread availability.
i am really hopeful about trinity. amd's demo at ces looked very promising. trinity's bad points are all on cpu side - cpu based on bulldozer arch. from the slide posted earlier, i see that trinity has high clockrate like zambezi and top apus are 4 core i.e. 2 modules. from what i've learned so far about l3 cache (thanks palladin and the guys) i can't help thinking that trinity could benefit from some l3 cache. but that's up to the amd engineers to decide. hopefully they'll make it work. trinity could also use igpu turbo like intel's - it'd make the apu much more power efficient. amd already has power saving functions like zcp in the 28 nm gcn cards. i hope this, along with vce which is somewhat confirmed to be present in trinity - a gcn-specific function, make their way into the apus despite them being based on older arch.
according to rumors and leaked slides: even amd isn't claiming more than 10-15% improved cpu performance. amd claimed around 50-60% gpu improvement. anything higher would be made up. the improvements are task-specific so they will not show up in every single test/benchmark. so the maximum and minimum improvement will fluctuate more. speculations and predictions are different. i hope people who 'speculate' make it clear that they're not claiming 'actual' performance figures. when the apus come out, we'll find out who was right and who was wrong.
i noticed that every time desktop apu is brought up and it's performance is criticized, an argument starts and the topic gets turned to mobile apus and how great they are. i noticed this when i said tom's was right to test a4, a8 apus for the sub $200 gaming cpu round up. same in the sub $200 apu roundup article's comments and again later in this thread. almost no one argues mobile apus' performance. amd's current igpus are the best in the market. but their poor performance as desktop gaming cpus is also true. i only put in the bad points in because i was criticizing, not badmouthing them. i thought that the good points of the apus were understood and people were discussing knowing that. intel igpus have a !@#$ load of bad points, but this is not strictly an 'amd vs intel' topic.
i am also thinking about adding multiple disclaimers with my posts such as - 'i lack deep knowledge about cpu and gpu architectures', 'i have to know more about mobile products', 'still learning!', 'discussing based on rumors only' etc.
thank you for reading.

 

[esrever]

AMD claims 25% cpu and 50% gpu, which was upped from 15% cpu and 30% gpu. the trinity laptops are going to have turbo 3.0 which will have higher turbo speeds, more responsive turbo and gpu turbo. The chip is suppose to run with less power than intel's under the same kind of loads with the same tdp. That is what AMD claims. The gpu is said to be VLIW4 but some rumours say GCN which doesn't add up because of the 28n -> 32n switch they would have to do. The piledriver core is said to have IPC improvements.

Most of the line with be dual module/ qaud core chips except for the A4s which will have a single module. The gpu is supposed to be 7660D in the desktop and 7760G in the laptops so the performance should be somewhat close to that level, maybe 6750 level performance.

Heres hoping AMD makes all that work. Im expecting less than they claim but I wouldn't be surprise if they hit it either.

 

[viridiancrystal]

After i saw the Radeon HD 7770(Heck the 7750 isn't to bad but its a tad slower then the 6770 while costing the same) and BD and how they priced it i'm real disappointed in Amd, Their asking to much money and basing prices on best performance instead of overall performance.



To me their is only 1 major product Amd is producing and that's Llano and i'm probably just going to get a A8 for a Laptop since i want to play games(on the go) but i'm happy with medium settings.


Besides that i'm pretty sure Nvidia's new cards will smoke Amd's on price and even overall performance! But i do think it will take more power and chad asked someone(don't remember who) why he thought that and i can tell you why i think Nvidia's cards will use more power and that's because its supposed to be bigger in size.
I was not so good at making predictions for BD but that's because i was being blinded by marketing this time its not happen again! Or at least for now

660 will probably equal the 7970 this time around but for a much lower price i mean think about it the 7970 is only 20% faster then the 580 and do you guys really think Nvidia can only improve their design 20%!

I'm also pretty sure Piledriver will only be around 10-15% faster per core over the BD and that's from better clock speeds and maybe 5% better IPC.

I'm also pretty sure Trinity will be a replacement and that's pretty much it for Llano on the cpu side of things i'm expecting 5-10% more performance per core and around 10% less performance on multithreading programs such as Handbrake why because of BD lower scaling vs stars. I'm strictly talking about the laptop market and not the desktop market even though its probably going to look the same.

And like i said before the way Amd is pricing their newer products i'm guessing Trinity will be over priced as well.

I have no intrest in starting a war of gpus in the cpu section, so all i will tell you is you are wrong. Very wrong. I'll pm you if you really want to know why.

malmental: while the Radeon 7k series loosely fits into the picture, i agree this is no place for keplar talk. Sorry.

As for trinity talk, i really feel that having it clocked so high is not going to end well in one way or another. Power use is my main concern, and real cpu improvement is another. I will agree with what has been said before, if trinity is only a small improvement over llano it will still sell very well. People dont read benchmarks and know that llano cpu is slower than an i3 or i5 mobile. If one friend comes to anothers' house with his laptop and starts playing a nice looking game like bf3 in 720p, it makes that laptop look pretty nice, and word gets around.

 

[viridiancrystal]

/now, come on. How is that productive 😛 XD

 

[viridiancrystal]

I just got back and read all the 1.5 pages since my last posting.....
I need some popcorn, it got hilarious..

Love it when that happens.

 

[g4114rd0]

AMD's upcoming Pitcairn..."

http://www.fudzilla.com/home/item/25998-amds-upcoming-pitcairn-gpu-detailed

 

[gamerk316]

I never insulted anyone. You called me a fanboi. And from what I know, a fanboi would never buy anything but the product of the company they are a fanboi of, much like BM who has not bought a single personal Intel machine in who knows how long. And as I said, I buy AMD. I just tend to see logic.

I will asmit to being a Corsair fanboi. Have their case, RAM and PSU in most of my builds. Also Asus. But then again neitehr companies have let me down and thats why I continue to buy their products.

And most of my BD was based on the little info we got from third parties and the info we got from AMD themselves about the arch. Sad thing is I was right. I didn't get blinded by AMDs marketing for BD, which is one thing a lot of people tend to do on both sides.

Still I am afraid for Trinity in the CPU space. Having a powerful IGP does not make up for parts that the IGP cannot do.

And again as I said many a times, we shall see when it comes out. We will of course get "leaked" info before release, most will be fake but some will be true just like was with BD.

Don't feel bad; I was the first person who speculated BD would see a performance decrease compared to PII. Despite the fact I was proven right (and for the same exact reasons I speculated in the first place), I'm now forever called an Intel fanboy, despite the fact that I was also one of the few to call Larabee a horrid design, or call their IGP's utter crap [which they still are, just better crap then they used to be].

And yes, I can't see the CPU side of Trinity being a significant upgrade over llano. Fact is, BD is NOT a suitable architecture for low-power usage if you need to ramp up clocks to remain competitive. I look at Trinity as an early indicator of the improvements PD will give [which I suspect won't be much, as there hasn't been enough time to radically re-design large portions of the architecture].

 

[noob2222]

Just for shnitz and giggles:

http://techiser.com/amd-trinity-a8-vs-llano-a8-3850-3dmark-11-benchmark-147387.html

This shows a light. Of course it shows only one part, the CPU. The GPU is just rubbish as Trinity got to use a HD7950, while Llano got to use a HD6550D IGP. So thats nothing to look at. But look at the Physics score, which is mainly CPU bound.

The scores are interesting. Llano has a 13% lead when both are clocked at 3.2GHz. As I said, rumors are showing that Trinity on[strike] a per core [/strike]and per clock level is slower than Llano, showing the same thing as BD which is slower than Stars on a per core and per clock level. Thats why Trinity will have top end CPUs reaching nearly 4GHz, so they can make it look better than it really is.

If we also ad 13% in clock speed to trinity, that means it will take about 3.616GHz Trinity to match a 3.2GHz Llano in the CPU end. In order to get the 10% some are expecting, that means a 3.936GHz Trinity CPU to look better than llano in the CPU and not just the IGP.

Of course as I said its all rumors and guesses. No way to say what I posted is true or if what I calculated is true. It may not be.

you do realize that 3Dmark physics is very multithreaded right? That means with LLano vs PD, your testing CMT, that is not a per-core test, AMD already stated CMT = 80% of a dual core, and in most tests, thats pretty dead on. Take away the 13% disadvantage and that gives PD a ~7-14% ipc over llano on a single core test.

Looks like pd should be an improvement over stars in single threaded tests. If you really thought PD would improve over stars in multithreaded testing, then CMT would either have to be 100% scaling, or ipc 20% higher than stars.

 

[triny]

Don't feel bad; I was the first person who speculated BD would see a performance decrease compared to PII. Despite the fact I was proven right (and for the same exact reasons I speculated in the first place), I'm now forever called an Intel fanboy, despite the fact that I was also one of the few to call Larabee a horrid design, or call their IGP's utter crap [which they still are, just better crap then they used to be].

And yes, I can't see the CPU side of Trinity being a significant upgrade over llano. Fact is, BD is NOT a suitable architecture for low-power usage if you need to ramp up clocks to remain competitive. I look at Trinity as an early indicator of the improvements PD will give [which I suspect won't be much, as there hasn't been enough time to radically re-design large portions of the architecture].

I never bought anything from AMD and they accused me of being a fanboy because I was posting on this thread

 

[king smp]

just read three pages of this thread
good stuff in there
but this was the best thing said

quote palladin7479
"Anyhow, lets all agree to stop insulting each other. We're all intelligent enough to analyze the information presented and come up with our own conclusions. Don't want the thread to devolved into haters fighting each other"

it is fun to talk Intel vs AMD
good thing we have two companies to compare instead of a monopoly
but no reason to get insulting
having a smart tech discussion of Intel vs AMD and their good points and bad points
is great
but we dont need hating

 

[king smp]

that's not the reason.
it's your tone about it and tunnel vision you have when you speak on all assumptions.
you act like it's a definite when you speak on the technology but it's not.

you are like:
it's going to be this and going to be that, don't refute me because I know.....

well, how do you know.?
because the roadmap says so.?
there will be charts and articles that will say one thing and there will be articles and charts that will say another.
WE DON'T KNOW
it's all speculation.

nothing wrong with expressing your thoughts, ideas, and theories but to combat every posting with repetitive statements
based on assumptions and wants that you feel is the truth gets old..
maybe if you just ease your manner about approaching ideologies that do not concur with yours
things wouldn't be like the way they are.

no offense, my opinion to you.

+1 nicely stated
this is just for fun
everyone is entitled to their opinion

 

[Cazalan]

Heh, a "jumbled mess" is what my VLSI design project looked like. My lab partner & I spent quite a lot of time designing just a basic clocked inverter cell, with a fanout of 2, in a delay chain with taps so that you could get 2^n different delays out of it. Extremely simple circuit but we did all the step & place grunt work ourselves once we had the basic cell design, then of course we had to design the output drivers around each output bonding pad and then account for their delay, the clock drivers, etc etc. Just routing Vdd & ground took a few days..

Did that in college too. That's the beauty of hand placed designs. You can put your initials in the unused space and see it with a microscope.

 

[jdwii]

I have no intrest in starting a war of gpus in the cpu section, so all i will tell you is you are wrong. Very wrong. I'll pm you if you really want to know why.

malmental: while the Radeon 7k series loosely fits into the picture, i agree this is no place for keplar talk. Sorry.

As for trinity talk, i really feel that having it clocked so high is not going to end well in one way or another. Power use is my main concern, and real cpu improvement is another. I will agree with what has been said before, if trinity is only a small improvement over llano it will still sell very well. People dont read benchmarks and know that llano cpu is slower than an i3 or i5 mobile. If one friend comes to anothers' house with his laptop and starts playing a nice looking game like bf3 in 720p, it makes that laptop look pretty nice, and word gets around.

PM me then, because i feel Amd is charging way to much for the performance, And the 7950/7970 is not a price/performance king its just priced accordingly. And the reason why i brought it up is to prove Amd is overpricing their products and they will probably do the same with PD and trinity and a 20% improvement over a 580 is nothing to brag about! Sorry but it isn't and i think Nvidia is going to prove that soon! And for once in like 3years beat Amd on price/performance and also have the fastest video card, And i'm not no fanboy of nvidia i only usually buy Amd products! Please though PM me(nicely though).

 

[triny]

that's not the reason.
it's your tone about it and tunnel vision you have when you speak on all assumptions.
you act like it's a definite when you speak on the technology but it's not.

you are like:
it's going to be this and going to be that, don't refute me because I know.....

well, how do you know.?
because the roadmap says so.?
there will be charts and articles that will say one thing and there will be articles and charts that will say another.
WE DON'T KNOW
it's all speculation.

nothing wrong with expressing your thoughts, ideas, and theories but to combat every posting with repetitive statements
based on assumptions and wants that you feel is the truth gets old..
maybe if you just ease your manner about approaching ideologies that do not concur with yours
things wouldn't be like the way they are.

no offense, my opinion to you.

I call it how I see it Currently AMD has been putting the best ideas forth

 

[nforce4max]

I want a vision laptop for m wife
intel laptops don't measure up as far as video is concerned

I personally own two and AMD did very well in the laptop arena compared to intel. One the cpu side things are very much mixed but the rest is where the money is. Upgraded from a crummy a4 3300m to a a8 3530mx. The other laptop was a freebie that being a cheap e-350 based machine. Gaming goes very well and multibox wow time to time on them and hopefully adobe premier will add support for faster encoding.

 

[dogman_1234]

http://i1130.photobucket.com/albums/m529/malmental/male29-male-theater-cinema-smiley-e.gif

+1

 

[triny]

PM me then, because i feel Amd is charging way to much for the performance, And the 7950/7970 is not a price/performance king its just priced accordingly. And the reason why i brought it up is to prove Amd is overpricing their products and they will probably do the same with PD and trinity and a 20% improvement over a 580 is nothing to brag about! Sorry but it isn't and i think Nvidia is going to prove that soon! And for once in like 3years beat Amd on price/performance and also have the fastest video card, And i'm not no fanboy of nvidia i only usually buy Amd products! Please though PM me(nicely though).

Overpriced is what the market will not bear ,should that be the case sales should suffer. Everyone has a limit as to what they are prepared to pay .Since there are still lots of 6870-6850 in the pipeline amd can afford to keep prices higher for now.
Prices are not written in stone
As far as PD and Trinity are concerned prices will be commensurate with improvements over BD 25% improved should equal 25% price hike. Again if the prices are too high sales will suffer and then they will go on sale.
Like 1000$ intel parts they will languish on the shelf .
Trinity will be over priced as lano parts will still be available .
Nvidia is in trouble I'd be surprised if they can come back with low power improved parts this year

 

[Cazalan]

The only way I see that 20+ for CPU is due to higher clocks and turbo. Not IPC. Its the only reason why the top end Llano is 3GHz while the top end Trinity will be 3.8GHz.

The clock speed doesn't surprise me too much. It's basically an FX-4120 with a 100mhz bump.
Minor cache improvements + clock speed should get the 20% over llano without much issue.

While not totally new there is mention of a Bulldozer B3 stepping as early as October '11. That would be the 6th incarnation of Bulldozer.

I think Trinity will do well this year but Haswell will be a problem for them in 2013. 22nm tri-gate is a game changer.

 

[triny]

I personally own two and AMD did very well in the laptop arena compared to intel. One the cpu side things are very much mixed but the rest is where the money is. Upgraded from a crummy a4 3300m to a a8 3530mx. The other laptop was a freebie that being a cheap e-350 based machine. Gaming goes very well and multibox wow time to time on them and hopefully adobe premier will add support for faster encoding.

how does 3530 do in games which quality setting?
My wife wanted a Lano laptop after we saw our friends HP laptop
the video in the HP made me weep I had just bought her a nice sony laptop
anyway we sold the sony and I'm waiting for trinity and IB to release to decide
but I can't see IB beating lano even

 

[triny]

The clock speed doesn't surprise me too much. It's basically an FX-4120 with a 100mhz bump.
Minor cache improvements + clock speed should get the 20% over llano without much issue.

While not totally new there is mention of a Bulldozer B3 stepping as early as October '11. That would be the 6th incarnation of Bulldozer.

I think Trinity will do well this year but Haswell will be a problem for them in 2013. 22nm tri-gate is a game changer.

exactly how I see it
Haswell will be the game changer until then I see the mobile market tilting to AMD
20% over lano is an easy target i'm expecting more 25-30% but 20% is suffiient

 

[triny]

nevermind, totally missed the point.
troll on.

I like seeing your funny faces

 

[nforce4max]

how does 3530 do in games which quality setting?
My wife wanted a Lano laptop after we saw our friends HP laptop
the video in the HP made me weep I had just bought her a nice sony laptop
anyway we sold the sony and I'm waiting for trinity and IB to release to decide
but I can't see IB beating lano even

It does surprisingly well provided you are not trying to play at 1080p and were not cpu limited. Besides the gpu the other main benifit of mobile Llano for gaming to go is the unlocked cpu multiplier. Unlike the desktop version of Llano the multiplier is unlocked and allows for every easy overclocking. Some have gotten a little over 3ghz stable. In most games the gpu is what pulls it's weight. Around 2.4ghz mobile Llano makes up for it's weakness compared to Intel very well.

 

[nforce4max]

will Intel graphics ever be as good, I don't think so unless they have some kind of breakthrough or 'acquisition'.
but the CPU side of things mobile included IB will be a gap widener..
but where AMD is going with APU graphics is far ahead.

Most Intel binned gpus are not even Intel at all but licensed from a forgotten company known as PowerVR. Yes the same company that was behind the Kyro and the Kyro 2 later moved onto the mobile market.

 

[Cazalan]

Most Intel binned gpus are not even Intel at all but licensed from a forgotten company known as PowerVR. Yes the same company that was behind the Kyro and the Kyro 2 later moved onto the mobile market.

PowerVR does scale though. It's just a matter of how many cores/tiles they want to dedicate for the GPU.
It's used in the iPad 2, TI and other SoC using ARM cores.