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 ...

 

[jimmysmitty]

Intel doesn't need to beat trinity's igp if they reach lano they may escape loosing big in the mobile market
same with AMD all they need to do is be competitive with Intel's SB cpu it doesn't have to beat it.

I am willing to bet it does reach Llano. Or at least keeps up with it.

Lets also not forget the power advantage Intel will have due to 22nm. Their top end DT CPU will only be 77W TDP, even the lowest end FX CPU is still 95W TDP and doesn't have four "full cores". Its also going to have a higher stock clock than Llanos top end 100W TDP, and probably overclock higher before hitting that same 100W TDP.

That matter a lot more in the mobile sector than anything else.

Hmm, there are other links posted in the IB/Haswell thread stating that Haswell will be (and I hate to use Baron's hyperbole 😛) a "graphics monster". Something along the lines of 2X or more of IB's GPU. Last fall I saw an Intel slide promising "7X" performance over SB's GPU, although that could have been the HD2K. If Intel does stick low-power dedicated DDR on top of the die with silicon interposer tech, then they could get massive improvements with that feature alone.

I guess we'll all find out in about 14 months from now.

Personally, being a gamer, I'll probably be happy with IB and a high-end 28nm discrete GPU, either AMD or NV depending on which gives the most bang for the buck.

As I said before, its supposed to be 40 EUs vs 16. That means it should see a nice jump alone.

 

[triny]

Trinity should be quite a hit
discrete igp with a big power advantage over Lan0
but next year with Kaveri steamroller should be a big improvement
by excavator they best be thinking of 20nm to remain competitive

 

[esrever]

http://www.forbes.com/sites/ericsavitz/2012/02/24/amd-ubs-pounds-the-table-boosts-target-to-11/

 

[de5_Roy]

Lessee now, in 2006 AMD had something like 30% marketshare in server - the most high-margin and profitable CPU segment by a long shot. Today they are what - 5%?

In 2006 they also had a much larger share in desktop - again, more expensive and thus profitable CPU segment.

Today they are competitive in low-end mobile, which is pretty low-margin.

By investing that $5.4BN in buying ATI, they had to cut back on R&D, sell off various parts of themselves including their fabs, and now appear to have pretty much given up regaining the lead according to the statements Reid made during the past few months, including the analysts day meeting last month.

Thanks to GloFlo, they apparently lost out on selling lots of Llanos to Apple because they couldn't deliver enough.

In short, it seems ATI has reaped the benefit of the merger far more than AMD has..

what!? i remember reading about that '5% share' in servers in an article... but that's a steep drop imo. aren't servers the biggest revenue generating sector like office pcs?
glofo's early difficulty in producing sufficient amount of llano apus held back amd from gaining an even bigger share in notebook market. i'd rather credit tsmc for the majority of apus' success.
this is one of the factors amd-favoring people here have been avoiding... trinity's biggest weakness isn't performance at all, it is wide availability. apus can earn huge profits in markets like china, india but they'll have to be available. if the yields are not enough.....
take any assurance about trinity yields from amd with a grain of salt.

 

[esrever]

since glofo had 10 months to get 32nm down, trinity shouldn't suffer from poor yields like llano.

 

[triny]

Here check this A6 Lano playing BF3 in ultra off the igp no discrete card
http://www.youtube.com/watch?v=olcz7DVXTkM

I seriously doubt IB will have this level of performance

 

[jimmysmitty]

since glofo had 10 months to get 32nm down, trinity shouldn't suffer from poor yields like llano.

Probalem is SOI takes longer to get better yields. And thats also if there are no other isues to work around. Doubt they will have the same yields as Intel will have on 22nm.

Here check this A6 Lano playing BF3 in ultra off the igp no discrete card
http://www.youtube.com/watch?v=olcz7DVXTkM

I seriously doubt IB will have this level of performance

I do wonder what resolution. The guy forgot to add that in.

And you can doubt all you want but we have about a month and we will see facts instead of rumors and thin air facts.

 

[triny]

http://www.forbes.com/sites/ericsavitz/2012/02/24/amd-ubs-pounds-the-table-boosts-target-to-11/

if they pull off trinity they will be back in the game

 

[triny]

Probalem is SOI takes longer to get better yields. And thats also if there are no other isues to work around. Doubt they will have the same yields as Intel will have on 22nm.



I do wonder what resolution. The guy forgot to add that in.

And you can doubt all you want but we have about a month and we will see facts instead of rumors and thin air facts.

it will be interesting to see how close both companies get intel igp amd cpu
price is another factor as well
that the a6 can even do that is astounding for the price point
easily lano is the best gaming cpu for the money

 

[triny]

since glofo had 10 months to get 32nm down, trinity shouldn't suffer from poor yields like llano.

Unless glfo is being paid by Intel ,some may laugh but it fits with Intel's mo.

 

[Chad Boga]

Unless glfo is being paid by Intel ,some may laugh but it fits with Intel's mo.

It also fits with stupid people's ideas.

 

[triny]

It also fits with stupid people's ideas.

stop calling Intel stupid Chad

 

[jdwii]

stop calling Intel stupid Chad

Ha what will he come up with next?

 

[fazers_on_stun]

http://www.forbes.com/sites/ericsavitz/2012/02/24/amd-ubs-pounds-the-table-boosts-target-to-11/

From the article:

The analyst sees improving margins driven by a ramp in higher end application processors, in particular the company’s Trinity processor for ultrathin notebooks, as well as a decline in sales of lower margin chipsets, server processor share gain and graphics margin gains.

IIRC Reid mentioned some improvement in server during the Q4 conference call, but I haven't seen any numbers from IDC or elsewhere yet. I suspect a lot of those were upgrades for HPC as well as server, but frankly Interlagos is fairly mediocre according to Johan de Gelas' review on AT. About the same as BD on desktop - loses out too frequently to the older AMD Magny Cours as well as Xeons.

But the article implies higher prices for Trinity parts, which means higher margins than what AMD gets with Llano. At least AMD seems like they will be overcharging Triny if he gets Trinity 😀..

 

[fazers_on_stun]

what!? i remember reading about that '5% share' in servers in an article... but that's a steep drop imo. aren't servers the biggest revenue generating sector like office pcs?

Yeah, the margins are huge in comparison to any other segment. IIRC AMD was getting over $1K for the higher-end Magny Cours, and of course for dual or quad-sockets that's 2 or 4 CPUs sold for each one...

What's funny is that right before the 12-core multi-chip module MC was released about 2 years ago, AMD had over 12% of the server market, but then started dropping when the 6-core/12-thread Westmere Xeons appeared about the same time. MC was delayed too IIRC, so basically AMD squandered yet another opportunity to take away marketshare from Intel by having the top end to itself. As Reid says, execution is pretty important..

glofo's early difficulty in producing sufficient amount of llano apus held back amd from gaining an even bigger share in notebook market. i'd rather credit tsmc for the majority of apus' success.
this is one of the factors amd-favoring people here have been avoiding... trinity's biggest weakness isn't performance at all, it is wide availability. apus can earn huge profits in markets like china, india but they'll have to be available. if the yields are not enough.....
take any assurance about trinity yields from amd with a grain of salt.

Supposedly the Llano & BD problems, delays and initial low yields were due to moving the GPU from strained silicon to SOI, requiring a redesign of the layout. I'd bet it was more complicated than that - GF had never done HKMG before, and they went with IBM's gate-first version which requires the Hafnium compound to be able to withstand the annealing temperatures (around 1000 degrees). In contrast, Intel & TSMC's gate-last approach use dummy gates for self-alignment, then replaces them with the HKMG gate after the annealing process, so the parameters can be 'tuned' for performance rather than surviving high temps. This would somewhat explain why BD uses too much power as well, since HKMG is supposed to reduce leakage by 30% or more.

Supposedly gate-first has up to 20% more density than gate last, but you can't tell that from BD's huge die size (or the missing 600 million transistors 😛).

So new designs on a major new process node = 2X potential problems. Which is why Intel does the tick-tock, to minimize risks.

Llano was delayed by at least 9 months and BD probably by an equal amount if not more. I don't believe TSMC does SOI, so AMD is pretty much stuck with GloFlo I guess. If Trinity doesn't appear until Q4, then I'd say GF's problems continue to some extent. They are switching to gate-last at 22nm, but who knows when they'll get to that node..

 

[fazers_on_stun]

It also fits with stupid people's ideas.

Apparently Triny didn't get the memo - AMD and Intel settled the lawsuit over 2 years ago 😛. Heck, even NY State has given up, to the chagrin of presidential wanna-be, Andrew Cuomo..

I suppose that when you live in the deep Canadian woods, and are busy trying to avoid being molested by Bigfoot and/or Sasquatch 😀, then such news takes more than 2 years to penetrate the ol' consciousness.. 😗

 

[de5_Roy]

i didn't know that llano was delayed for 9 months. that's understandable considering amd's track record with delays and the amount of effort gone into bringing the cpu and gpu together. trinity's recent rumored launch time - june 2012 seems to be in line with amd's current roadmap.
...... i have started to suspect that someone calling himself/herself llanore/ellanore/l24n0/noall/lanlo/lolan might have joined the forums around this time last year.... 😗

 

[dogman_1234]

So, explain to me again why the discussion of Intel vs. AMD has been prosperous, and how did it start. I see no value,( except for the few guys here who provide us info on how a CPU works and performs,) to it regarding PD.

 

[esrever]

So, explain to me again why the discussion of Intel vs. AMD has been prosperous, and how did it start. I see no value,( except for the few guys here who provide us info on how a CPU works and performs,) to it regarding PD.

some people can't deal with AMD love in an AMD thread.

 

[triny]

From the article:

The analyst sees improving margins driven by a ramp in higher end application processors, in particular the company’s Trinity processor for ultrathin notebooks, as well as a decline in sales of lower margin chipsets, server processor share gain and graphics margin gains.

IIRC Reid mentioned some improvement in server during the Q4 conference call, but I haven't seen any numbers from IDC or elsewhere yet. I suspect a lot of those were upgrades for HPC as well as server, but frankly Interlagos is fairly mediocre according to Johan de Gelas' review on AT. About the same as BD on desktop - loses out too frequently to the older AMD Magny Cours as well as Xeons.

But the article implies higher prices for Trinity parts, which means higher margins than what AMD gets with Llano. At least AMD seems like they will be overcharging Triny if he gets Trinity 😀..

I suspect AMD will fit that resonant clock as well as other changes to it's Opteron line up Bull doozer is about to be eclipsed
I don't mind paying for good designs it's when I buy an Intel laptop for north of 1000$ and a friends AMD laptop at half the price
blows it into the weeds.
I always try to buy good quality Intel doesn't know how to make good laptops
The good thing about Canada is we live with nature and are not over populated what you mock is what
we cherish.

 

[palladin9479]

not when the AMD love gets too unrealistic with ridiculous assumptions that are just not logical to a certain point.
then some (guilty parties) postings might be.

it's kind of strange like bipolarity..
if your offended then your included.

Seeing as BM hasn't posted in a very long time, you making a straw man argument here. Triny is just posting what is possible, seeing as Trinity / PD hasn't been launched yet.

What I find funny is the anti-AMD (BD or whatever) folks bashing anything remotely related to AMD. It's kinda obvious when people are comparing Intel, a company with almost no GPU experience to ATI, a company that's consistently released great GPU's, and somehow mentally justifying victory for the Intel part.

We're here talking bare silicon, but there's another component to GPU's that isn't silicon. Drivers are just as important if not more important then base HW. Both Nvidia and ATI spend large amounts of time fine tuning their drivers for various games / applications. Intel is a good 3~5 years behind both NVidia / ATI. It's going to take awhile for them to create a uArch that can compete. Intel has the money to do the R&D, but it's not something that you can pound out over the course of a few months. It takes years to build up the work and acquire all the tech required. Even if Intel poured tons of their money it would still take several years to catch up, and that's assuming NVidia / ATI are sitting around doing absolutely nothing.

 

[triny]

By looks of things nvidia is having a hard time AMD appears to be capturing market share without resistance .If Trinity can capture mobile sales ,they will be in a much better position to execute their road map.

Personally I think amd trinity is very likely to ransack Intel's bottom line in mobile,opteron server and hondo ULV products
at the same time they appear to have little resistance from Nidia and have already started nvidia hemorrhaging market share

could be quite a bloodletting

 

[nforce4max]

By looks of things nvidia is having a hard time AMD appears to be capturing market share without resistance .If Trinity can capture mobile sales ,they will be in a much better position to execute their road map.

Personally I think amd trinity is very likely to ransack Intel's bottom line in mobile,server and ULV products
at the same time they appear to have little resistance from Nidia and have already started nvidia hemorrhaging market share

could be quite a bloodletting

Trinity will capture mobile sails without hardly much competition due to reasonable cost and ability to game on the go. I am wondering when AMD is going to do APU's with FireGL certification.

 

[triny]

Trinity will capture mobile sails without hardly much competition due to reasonable cost and ability to game on the go. I am wondering when AMD is going to do APU's with FireGL certification.

I'd be surprised to see that anytime soon maybe by the time cell phones replace computers
workstation solutions need bigger guns and is lucrative sales for gpu market
But I can envision the APU growing next year with Kaveri 28nm process will allow for a bigger APU ie: 8860D
by excavator who knows .Ie: 9970D
etc where it stops only AMD knows for sure

 

[viridiancrystal]

I think most anyone has had the mindset that AMD is having trouble meeting the gains they say they get, but that is basically because bulldozer was so much worse than claims had said.

How many of those claims were from AMD themselves, not as many as you might think.

However, aside from BD they had done some amazing things recently. Think about this: Llano was launched last year, and now there is a lot of talk about igp and apu's.

Apu's have forced Intel to scramble to catch up, and that is now a major focus of theirs, much more than improving cpu power.

Along with their situation in the graphics market, AMD right now has a good chance to gain market share and improve the value of the company. While they are gain market share in the graphics sector, gpu sales mean much less than apu sales will/do.

Trinity is going to mean big things, and it is being overlooked by many.