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 ...

 

[truegenius]

lol
1gb of cache can do this but will need a socket of itx or matx size

ddr4 can also do that and it will also increase performance of igpu of apu and also intel's igpu.

(it is only a thought
Or maybe amd can introduce some kind of new memory which will use new interface or pcie or hypertransport or a new slot (like of ram) purely dedicated to vram (just like old style cache) or something else by modifing chipset or something else, this will provide +point to amd (as amd is in ram and gpu business and can introduce vram as like ram, removable and upgradable))

 

[triny]

What current discrete card would give the same performance as this 7760x?

the question is which chip has a hd7760 igp I fail to see what discrete cards have to do with that.
If the trinity is that good AMD will dig a deep hole in Intel's mobile pocket

 

[g4114rd0]

-...AMD should never use a Hope's candle."

http://www.donanimhaber.com/islemci/haberleri/DH-Ozel-intelin-ote-nesil-Haswell-islemcilerinin-cikis-tarihi.htm

 

[triny]

using a fast gpu (even if using 1k spu if tdp is still under control by some magic) will not make the trinity a good choice over dedicated card (even 6770)
because of slow ram, and using high speed ram is not cheap which is making apu a flop product for medium gaming.

Amd needs to overcome this memory problem.


(you can try this at home
use 300mhz (to simulate slow vram for igpu)for video memory of your card and first stock and then overclocked gpu to test improvements in performance/benchmark.

Of course alone it won't be ,those who buy one will most likely add the appropriate video card and the 2133 mhz ram
if they are building a gaming htpc

 

[triny]

-...AMD should never use a Hope's candle."

http://www.donanimhaber.com/islemci/haberleri/DH-Ozel-intelin-ote-nesil-Haswell-islemcilerinin-cikis-tarihi.htm

What's that got to do with Piledriver?
post English links

 

[triny]

same thing as I just stated on the Ivy thread.
people talking when not really knowing what they are talking about.
just blah blah blah, wasting your time and mine.

OK, morning rant is over, my apologies..

Looking at your Post I totally agree with you

 

[wh3resmycar]

AMD APUs only make sense on a notebook (nope not even on an HTPC).

but darn it, i'd buy a $800 A-series "ultrabook", because it makes more gaming sense than what's intel is offering right now pre-ivybridge.

platform innovation is just not on AMDs vocabulary. they're constantly trying to create new markets in a market they already own (the budget segment). sometimes you wonder if AMD's business division is ruled by monkeys.

 

[triny]

AMD APUs only make sense on a notebook (nope not even on an HTPC).

but darn it, i'd buy a $800 A-series "ultrabook", because it makes more gaming sense than what's intel is offering right now pre-ivybridge.

platform innovation is just not on AMDs vocabulary. they're constantly trying to create new markets in a market they already own (the budget segment). sometimes you wonder if AMD's business division is ruled by monkeys.

excellent for htpc as well imho
Something like the Arctic mc101 ought sell well if priced right.

 

[king smp]

if it can stream Netflix and Youtube and also play light duty games like Bejeweled or online flash games then more than half the market is happy
if it can handle MMO games like WOW then you really got something
while the computer gaming crowd spends alot of money they are the minority
so a decent CPU with strong IGP/onboard GPU will handle most of what the "common" user is doing
even notice AMD is not putting out a lower line of discrete cards like the 5450 type cards
wont be a need for cards like that especially if they start making onboard GPUs able to handle multiple monitors

 

[triny]

if it can stream Netflix and Youtube and also play light duty games like Bejeweled or online flash games then more than half the market is happy
if it can handle MMO games like WOW then you really got something
while the computer gaming crowd spends alot of money they are the minority
so a decent CPU with strong IGP/onboard GPU will handle most of what the "common" user is doing
even notice AMD is not putting out a lower line of discrete cards like the 5450 type cards
wont be a need for cards like that especially if they start making onboard GPUs able to handle multiple monitors

So true

 

[triny]

IBM Fab Club to Reveal Details on 20nm, 14nm and Beyond Process Technologies Next Month.

http://www.xbitlabs.com/news/other/display/20120208232412_IBM_Fab_Club_to_Reveal_Details_on_20nm_14nm_and_Beyond_Process_Technologies_Next_Month.html

 

[king smp]

IBM Fab Club to Reveal Details on 20nm, 14nm and Beyond Process Technologies Next Month.

http://www.xbitlabs.com/news/other/display/20120208232412_IBM_Fab_Club_to_Reveal_Details_on_20nm_14nm_and_Beyond_Process_Technologies_Next_Month.html

and I saw something about AMD working together with IBM on cpu manafacturing
not sure if I saw the link on this thread or somewhere else on Toms

edit- after research I learned (what most already know LOL) that AMD and IBM have a long history together

 

[triny]

http://www.tested.com/news/researchers-develop-method-to-boost-cpugpu-hybrid-performance/3569/

up to 113% performance gain for APU
the future is APU
No wonder Intel is scrambling to improve igp

 

[gamerk316]

up to 113% performance gain for APU

A: Meaningless if the CPU portion can't keep up
B: "Up to 666% faster performance in SLI" - Some random NVIDIA driver update. Be VERY wary once "up to" numbers start comming out. From my experiance, the average improvement is usually 25% of the "up to" number.

 

[esrever]

up to 113% performance gain for APU

A: Meaningless if the CPU portion can't keep up
B: "Up to 666% faster performance in SLI" - Some random NVIDIA driver update. Be VERY wary once "up to" numbers start comming out. From my experiance, the average improvement is usually 25% of the "up to" number.

well up the performance gain is averaging 20% is still good. Up to just gives an idea of its potential.

 

[triny]

well up the performance gain is averaging 20% is still good. Up to just gives an idea of its potential.

Perhaps today but I can easily see a line of 10-20% gains the next few years

 

[triny]

dude tell the entire story and not just edit what you want to make it look better than it is.
it's still good and definitely an improvement, but dang, slow down.
'fanboism shows hard in you it does.'

quote:
by up to 113 percent and 21.4 percent on average.
end quote:

1) lol I use Intel so your claims are bogus
2) this applies to Intel as well again your claims are bogus
3) your imagination is getting the best of you Dude

what part of 20% a year do you not understand about it .

 

[triny]

I'm poking at you, just play nicely.... 😀

but seriously, tell the entire story, both sides.

read the white paper at the link
over time both companies will gain substantial performance
up to 113%

 

[Chad Boga]

the question is which chip has a hd7760 igp I fail to see what discrete cards have to do with that.
If the trinity is that good AMD will dig a deep hole in Intel's mobile pocket

Because I am trying to work out how good or not this supposed 7760 IGP will be?

Just calling it a 7760 IGP means nothing, if we have no idea what its performance is like.

By what percentage will it be faster than the IGP in Llano?

Frankly, I think you have no idea whatsoever about this subject and are just blathering absolute nonsense, which is why you were once again so evasive to a question being put to you.

 

[esrever]

Because I am trying to work out how good or not this supposed 7760 IGP will be?

Just calling it a 7760 IGP means nothing, if we have no idea what its performance is like.

By what percentage will it be faster than the IGP in Llano?

Frankly, I think you have no idea whatsoever about this subject and are just blathering absolute nonsense, which is why you were once again so evasive to a question being put to you.

they named the llano gpu pretty comparable to the discrete parts.

 

[noob2222]

http://www.tested.com/news/researchers-develop-method-to-boost-cpugpu-hybrid-performance/3569/

up to 113% performance gain for APU
the future is APU
No wonder Intel is scrambling to improve igp

Thats been amd's plan all along with separating the FPU. Doing it at the hardware level takes more time and baby steps to ensure its maturing properly.

That's when AMD plans to release a pair of APUs code named Kaberi and Kavini that will sport a CPU-GPU combo that shares a unified memory cache. Those chips will also feature unified address space for the CPU and GPU components, the latter of which will use pageable system memory via CPU pointers.

In 2014, AMD intends to take HSA from the architectural integration stage to the system integration stage. In simple terms, that means computers that will know how to throttle up the CPU portion of the APU that runs them and dial down the graphics component for scalar processing tasks, while doing the opposite for parallel processing work that's more suited to the GPU.

http://www.pcmag.com/article2/0,2817,2399789,00.asp

 

[palladin9479]

Because I am trying to work out how good or not this supposed 7760 IGP will be?

Just calling it a 7760 IGP means nothing, if we have no idea what its performance is like.

By what percentage will it be faster than the IGP in Llano?

Frankly, I think you have no idea whatsoever about this subject and are just blathering absolute nonsense, which is why you were once again so evasive to a question being put to you.

No, your trying to compare a chip designed for platform integration to a gaming rig.

It makes absolutely no sense to put an APU inside any box that you could easily put a mid range discrete GPU in. It belongs on thin mobile platforms, and low profile mini-ITX case's (some HTPCs included). Having built a metric ton of mini-ITX boxs, for either dedicated home servers, network routes / gateway / VPN devices or just set-top internet browsers, I can safely say that their isn't enough room / available power for a discrete GPU. Your talking about a 60~90W pico-PSU, maybe even a 120W if your wanting to push it. Most of my design's were using the Via Nano platforms, but they have absolute sh!t video capabilities which makes them unsuitable for HTPCs. Now if you take an AMD APU on a Mini-ITX board and put it in the same situation, suddenly it all works out. Give it 4GB of memory or 6~8 if it's cheap enough, and some local storage and you got yourself a cheap set-top HTPC / medium gaming box. Hook it up to your HDTV and ensure it's case is color coordinated with the rest of your living room furniture and you really got something. Same scenario but inside a small 15 inch laptop, the Mac Book Air proved that there is a market segment for ultra-thin stylish looking mobile computing devices. I know I absolutely love my HP DV6z-qe, thinking about getting a second one with the additional GPU upgrade (let GF have the first one).

Also I'm solidly in the "gamer / enthusiast" crowd, my gaming rig is using two GTX580 Hydros inside a 800D with a Samsung 830 256GB. My gaming box is rather large and unmobile, I need something to play games on and do work when I'm traveling on business or pleasure. I also have a Sony desktop replacement laptop with a C2D, GTX 9600M (older laptop) and a full size keyboard. It's huge and heavy, low battery time and is a PITA to travel with. Thus I find the DV6z-qe to be a solid upgrade, my back thanks me.

 

[Cazalan]

AMD will unlikely want to introduce a new type of RAM because:

a) they dont have the R&D and manufacturing capacity
b)They're not in a position to guide market trend, besides introducing something just for VRAM of IGP is even less likely to take off than new RAM

They don't need to make a new type of RAM. Several already exist, they just need to implement them.

JEDEC WIDEIO spec- stacked on the DIE for a 3D chip
Low power and high bandwidth

 

[palladin9479]

On-die memory wouldn't be a good idea, too limited a space that you could put other things.

Better idea is next big socket revision, add a back side bus for 128-bit GDDR5 that can be used for dedicated GPU memory. On a board you could introduce a small socket for a small 512MB ~ 2GB GDDR5 memory add on. If no add on is detected it use's UMA, if add on is detected then it use's the much faster add on.

As the GPUs power starts to ramp up there will come a time when the system memory interface is severely handicapping the GPU. It's already acting as a bottleneck right now. Laptops will just have the memory soldiered to the board like they do discrete GPUs.

 

[palladin9479]

http://forum.notebookreview.com/hp-pavilion-notebooks/639355-definitive-dv6z-llano-overclock-optimization-guide.html

This is the link to a enthusiast site about the HP DV6zqe (AMD APU notebook by HP). Demonstrates how to modify and work with it. Also includes benchmarks, although their used to demonstrate the differences between different parts.

This is the integrated 6620G part and how it reacts to memory.
http://forum.notebookreview.com/gaming-software-graphics-cards/641254-amd-llano-6620g-benchmarked-various-ram-configurations.html