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 ...

 

[noob2222]

Hence why you NEED a reliable CPU test for gaming, which is my entire POINT.

so you feel that lowering the resolution is a reliable test on multiplayer gaming? MP gaming doesn't stress the gpu more, it stresses the cpu in a different way than single player.

In the case of BF3, the game engine runs primarily on 2 cores, hence why the Phenom II x2 560 is capable of putting out respectable framerates. Loading multiplayer puts that part of the game on separate cores, you still have the 2 primary game engine cores, but all addidional cores handle mp.

http://www.techspot.com/review/458-battlefield-3-performance/page7.html

check the usage charts, you will see in every case, there are 2 cores used more than others, and the 560 pushing 81 average fps in the single player test.

look at the PII x2 560 now. How is that reliable by lowering the resolution? it can't even handle medium settings 4x af. but on single player high settings 16af it pushed 81.

funny, even the phenom X4 thats soo much better than BD at low resolutions in single player testing ... wait ... what?!

 

[Chad Boga]

And in other news, AMD expected to post a loss this quarter:

http://www.amd.com/us/press-releases/Pages/press-release-2012jul9.aspx

AMD (NYSE:AMD) today announced that revenue for the second quarter ended June 30, 2012 is expected to decrease approximately 11 percent sequentially. The company previously forecasted second quarter 2012 revenue to increase 3 percent, plus or minus 3 percent sequentially.

I wonder how Triny is doing?

 

[gamerk316]

so you feel that lowering the resolution is a reliable test on multiplayer gaming? MP gaming doesn't stress the gpu more, it stresses the cpu in a different way than single player.

In the case of BF3, the game engine runs primarily on 2 cores, hence why the Phenom II x2 560 is capable of putting out respectable framerates. Loading multiplayer puts that part of the game on separate cores, you still have the 2 primary game engine cores, but all addidional cores handle mp.

http://www.techspot.com/review/458-battlefield-3-performance/page7.html

check the usage charts, you will see in every case, there are 2 cores used more than others, and the 560 pushing 81 average fps in the single player test.

http://www.sweclockers.com/image/diagram/2507?k=ce91bc9f6c6266a1792efe936c92f2d4
http://www.sweclockers.com/image/diagram/2506?k=142b45af179c625ccd8f53fea7385155

look at the PII x2 560 now. How is that reliable by lowering the resolution? it can't even handle medium settings 4x af. but on single player high settings 16af it pushed 81.

funny, even the phenom X4 thats soo much better than BD at low resolutions in single player testing ... wait ... what?!

Noob, you have no idea what you are talking about at this point.

The entire point of doing low-res testing is to compare how strong CPU's are in relation to eachother in gaming, without GPU settings affecting the results. THAT IS ALL. Is that REALLY such a hard concept for you to grasp?

 

[earl45]

Noob, you have no idea what you are talking about at this point.

The entire point of doing low-res testing is to compare how strong CPU's are in relation to eachother in gaming, without GPU settings affecting the results. THAT IS ALL. Is that REALLY such a hard concept for you to grasp?

I can't find the article on the Quad CF set up showing Intel performs better then AMD
pretty much what the low-res benching is saying as well. (2 different ways same result).

 

[noob2222]

Noob, you have no idea what you are talking about at this point.

The entire point of doing low-res testing is to compare how strong CPU's are in relation to eachother in gaming, without GPU settings affecting the results. THAT IS ALL. Is that REALLY such a hard concept for you to grasp?

obviously since i disagree with you, im clueless.

How many variables affect the results between 640x480 and 1900x1200? How much headroom is there at 640x480? Do you really think cpu/gpu and resolution scaling is linear? How many tests out there provide linear results? How much does memory bandwidth affect resolution scaling? Sure running low resolution puts more stress on the cpu, but it also stresses every other component, including the motherboard itself. So now you have to factor in ALL of the above when testing.

I have done testing, and I can tell you that the system as a whole is more sensitive at low resolution, memory speed/timings, bus speed, overclocking, ect, not just what cpu it is. Do all of those factors come into play at high resolution? not nearly as much.

The only test that should be done is adding gpus to the system to see exactly how much money you need to spend on GPUs to justify moving from one cpu to the next. For high end AMD thats somewhere between 3 and 4 6970/gtx 570 level gpus, depending on the game's coding.

Is there any single gpu close to that point?

Hmm where have we seen that idea before ... ahh here it is with low end system testing http://www.tomshardware.co.uk/fx-4100-core-i3-2100-gaming-benchmark,review-32384-7.html
Only when paired with a 6950 does the 2100 manage to start to pull away from the 4100 (not 4170)in some of the games .

Sure makes it easier to justify if you only push low res gaming doens't it, then the 2100 looks better all the time. But people don't play with the lowest settings do they?

answer one question: why do you think everyone quit doing low resolution testing in games? NO ONE CARES, well almost no one.

 

[-Fran-]

Why do you think everyone quit doing low resolution testing in games? NO ONE CARES, well almost no one.

Although I agree with the low res testing just being a Lab test, it does have its merits. Just like every God Damned metric out there. Just one thing to take into account, but not a game changer for a CPU purchase decision. At least, you guys have to agree on that...

And the CPU, like you say, doesn't come alone in this. I've been ranting about Hyper Transport for quite a while now and it's really really showing its age and using a FX8150 or a Athlon II X4 won't change the fact that Hyper Transport will be a bottleneck in some scenarios. Having a beefier CPU per sé it's not the whole story either; chipset should be taken into account as well. Hell, the whole MoBo should, hahaha.

^
You're also forgetting those BF3 benchies are of Single Player, which unlike the Multiplayer is not CPU heavy

Well, the FrostBite engine is supposed to use lower graphical quality in MP as well, even though it's in "ultra settings" like SP, AFAIK. Don't remember where I read that though 😛

Cheers!

 

[bobbybamf12]

YAYYYYYY FOR PILEDRIVER.....

 

[palladin9479]

Hilarious move on the part of the blue painted ones.

They of course kindly forgot to mention.

Although AMD saw clock power drop by a quarter, the potential power savings could be much greater, says Matthew Guthaus, an associate professor of computer engineering at the University of California, Santa Cruz. His team has shown that by customizing the size of inductors and their location on the chip, it may be possible to reduce clock power by as much as 90 percent.


Which is exactly what I said before. RCM can in theory save a ridiculous amount of power, due to how modern CPU's are made it's not feasible to use RCM to it's greatest potential. Thus instead of a 30~50%+ power savings we're going to be looking at 15~25%. Which is exactly what we predicted.

Of course the blue painted ones don't want anyone to know that, their too busy trying to spread the message of "PD iz FAILZOR 101000100!!!??~~~".

 

[-Fran-]

Hilarious move on the part of the blue painted ones.

They of course kindly forgot to mention.

Although AMD saw clock power drop by a quarter, the potential power savings could be much greater, says Matthew Guthaus, an associate professor of computer engineering at the University of California, Santa Cruz. His team has shown that by customizing the size of inductors and their location on the chip, it may be possible to reduce clock power by as much as 90 percent.


Which is exactly what I said before. RCM can in theory save a ridiculous amount of power, due to how modern CPU's are made it's not feasible to use RCM to it's greatest potential. Thus instead of a 30~50%+ power savings we're going to be looking at 15~25%. Which is exactly what we predicted.

Of course the blue painted ones don't want anyone to know that, their too busy trying to spread the message of "PD iz FAILZOR 101000100!!!??~~~".

Uhm... Where did you get that quote? XD

Cheers!

 

[r0ck3tm@n]

I wonder how Triny is doing?

I wonder if you can post something constructive and not inflammatory?

 

[palladin9479]

Uhm... Where did you get that quote? XD

Cheers!

From the same article they got theirs from.

The article stated both pro's and con's about RCM, it's potential limits and possible ways to improve.

They used wordsmithing to lift out only the negatives to persuade people that RCM is a failed technology and that PD would be "failz". It's a continuation of the BD bashing bandwagon that's been going on for months now.

 

[noob2222]

^
You're also forgetting those BF3 benchies are of Single Player, which unlike the Multiplayer is not CPU heavy

Actually the sweclockers was multiplayer, wich is why the x2 560 fell flat on its face even with the 6990 graphics card, pushing minimum fps to 8.

I do tend to agree with you yuka. There is more than just the internal cpu clock holding amd back as far as certain benchmarks go. Remember when intel put their pci-e controller on chip, they gained quite a bit of graphic speed when compared with the same cpu gen. That's why intel will always be slightly faster, not necessarily just because the cpu (alu/fpu) is capped out. We won't see amd take that step with high end till they are off of the AM3 socket. Trinity is there, but the cpu itself is low end, and we haven't seen any reviews exploring the pci_e controller.

 

[-Fran-]

I wonder if you can post something constructive and not inflammatory?

Ah, I missed his question... Since you already took it as a half empty question, I'll take it as a half full, haha.

I don't think Trinity is doing well in mobile. There are still a lot of Llano notebooks out there that need to be sold before the rest of the Trinity line up get cheaper (hence, sells more). Notice that's a feeling I have, no hard numbers to back me up.

And DT, well, a no show still over here at least.

From the same article they got theirs from.

The article stated both pro's and con's about RCM, it's potential limits and possible ways to improve.

They used wordsmithing to lift out only the negatives to persuade people that RCM is a failed technology and that PD would be "failz". It's a continuation of the BD bashing bandwagon that's been going on for months now.

Thanks for the clarification 😛

But I want to keep my hopes low, so I just want to believe RCM won't be the saving grace for PD's power consumption, even though it seems to be quite good. BD's numbers were quite ugly, so a 15-20% efficiency improvement might fall short (always in the context of overclocking, mind you all).

Actually the sweclockers was multiplayer, wich is why the x2 560 fell flat on its face even with the 6990 graphics card, pushing minimum fps to 8.

I do tend to agree with you yuka. There is more than just the internal cpu clock holding amd back as far as certain benchmarks go. Remember when intel put their pci-e controller on chip, they gained quite a bit of graphic speed when compared with the same cpu gen. That's why intel will always be slightly faster, not necessarily just because the cpu (alu/fpu) is capped out. We won't see amd take that step with high end till they are off of the AM3 socket. Trinity is there, but the cpu itself is low end, and we haven't seen any reviews exploring the pci_e controller.

Well, from the 790FX times we haven't seen any major departure in north bridge tech from AMD (aside from RAM and HT changes). It's just refresh after refresh to add minor tweaks or features (in the form of South Bridge) that's getting old in the tooth. Really annoying to be honest. Not that I'm ranting about it as a bad chipset; not at all, but it's showing its age and AMD is doing nothing at all IMO. The north bridges that accompany Llano are also quite lack luster (A70M and A75); only thing they have going for them is native USB3.

Cheers!

 

[Chad Boga]

Ah, I missed his question... Since you already took it as a half empty question, I'll take it as a half full, haha.

I don't think Trinity is doing well in mobile. There are still a lot of Llano notebooks out there that need to be sold before the rest of the Trinity line up get cheaper (hence, sells more). Notice that's a feeling I have, no hard numbers to back me up.

And DT, well, a no show still over here at least.

I was expressing concern for the poster Triny, not enquiring about the product Trinity.

 

[earl45]

I was expressing concern for the poster Triny, not enquiring about the product Trinity.

Your talking about the stock he purchased in AMD?

 

[Chad Boga]

Your talking about the stock he purchased in AMD?

That forms part of my concerns.

 

[skitz9417]

i hope when the pd come out are good because i have a amd phemon x 840 and i might upgrade

 

[gamerk316]

you that the system as a whole is more sensitive at low resolution, memory speed/timings, bus speed, overclocking, ect, not just what cpu it is. Do all of those factors come into play at high resolution? not nearly as much.

I'm laughing at this; the only reason that memory and the like don't show up in high res benchmarking, again, is because the GPU is bottlenecking the entire system.

I can't find the article on the Quad CF set up showing Intel performs better then AMD
pretty much what the low-res benching is saying as well. (2 different ways same result).

HardOCP did one a year or so ago. I dislike Quad CF/SLI benchmarking because FPS can vary so wildly in such configurations. Its a "better" benchmark for the CPU, but given how many games either don't scale well in SLI/CF and how many have SLI/CF driver issues...

 

[dogman_1234]

http://www.google.com/finance?client=ob&q=NYSE:AMD

Anybody interested in penny stock?

EDIT: AMD gets Federal Research Grant.

 

[noob2222]

I'm laughing at this; the only reason that memory and the like don't show up in high res benchmarking, again, is because the GPU is bottlenecking the entire system.

So your agree that there is benefits to running faster memory when your not gpu bottlenecked. 😱

And now back to memory speed discussion.

 

[-Fran-]

Man... AMD is really struggling this quarter. Just like the one before... And the one before that... And the one before... And... You get the point, lol.

I just hope PD puts a little more useful life to my Crosshair V before donating it to someone, lol.

Cheers!

 

[dogman_1234]

I wouldn't mind a little help on the market. If NYSE: (AMD) goes a little more south, I may consider. That is if they offer an alternative in their marketing division, or if this CEO can get this whole circus act together. Other than that, it is fun to watch what happens.

 

[gamerk316]

So your agree that there is benefits to running faster memory when your not gpu bottlenecked. 😱

And now back to memory speed discussion.

Assuming that your running a game that demands lots of memory transfer, yes. But these days, most games preload much of their data into RAM, so even at lower resolutions, you don't see any significant impact. And even then, when you consider the worst case overhead of paging in general, I'd argue the HDD itself would be a bottleneck before RAM is.

My point being is that these are the things you can actually benchmark at lower resolutions, AND NO ONE DOES.

 

[Cazalan]

Man... AMD is really struggling this quarter. Just like the one before... And the one before that... And the one before... And... You get the point, lol.

I just hope PD puts a little more useful life to my Crosshair V before donating it to someone, lol.

Cheers!

They have fierce competition no doubt. Their strongest talent is in the GPU space and that's constantly pressured by NVidia and Intel. With greater pressure by Intel at the high end coming next year with their 22nm MIC cards.

I thought this year would be better for AMD with Trinity having a clear advantage. Come next year Haswell. Investors get skittish if it looks like the oil well is going to dry up in 12 months.

 

[bawchicawawa]

Look at what i've found.
http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&DEPA=0&Order=BESTMATCH&N=100019095&isNodeId=1&Description=a10&x=0&y=0