AMD CPUs, SoC Rumors and Speculations Temp. thread 2

[gamerk316]

AMD releases it's quarterly financial results:

http://www.amd.com/en-us/press-releases/Pages/press-release-2015jul16.aspx

"Lower then expected APU Sales" shows up a LOT in that report. Debt still hovering around $2.3 Billion, or about the companies total worth.

Zen better be a hit, because debt payments start in two years...

 

[juanrga]

http://venturebeat.com/2015/07/16/more-hints-that-amd-is-building-nintendo-nxs-processor/

If true it could help AMD revenue.

It will help of course, but it is not enough. That design barely pays Zen R&D.

 

[juanrga]

AMD releases it's quarterly financial results:

http://www.amd.com/en-us/press-releases/Pages/press-release-2015jul16.aspx

"Lower then expected APU Sales" shows up a LOT in that report. Debt still hovering around $2.3 Billion, or about the companies total worth.

Zen better be a hit, because debt payments start in two years...

It is better to look at numbers from the company split point of view. The Computing and Graphics business revenue was down 54% from this time last year. The division posted an operating loss of $147 million, compared to a $6 million loss in the same quarter last year.

The Enterprise, Embedded, and Semi-Custom business revenue suffered an 8% year-on-year decline. Operating income was $27 million, compared to $97 million in the same quarter last year.

Thus the PC division is a complete mesh, whereas the other division --which is maintaining afloat the company-- has also contracted.

 

[cdrkf]

These numbers are hardly surprising, as is the assertion sue makes regarding growth in graphics.

The simple fact is amd have no new product available in cpus right now so with no new kit the situation won't improve vs competition (carrizo looked nice but still no laptops, hopefully they will come through soon?).

 

[-Fran-]

Yeah, I have to say I was quite surprised the numbers for this Quarter were so bad; I was expecting red, but not bloody red. Intel was impacted as well, but not with this magnitude.

All hopes and dreams of staying alive are deposited in Zen it seems. I will up my stake in saying that they can't make do with a close contender here, they need at least a solid tie. And that is a lot to say, given how far ahead is Intel.

On the graphics department, I wonder if they're still waiting on some sort of change in the market? I mean, CUDA is not losing any track in the GPGPU Pro world and Intel is about to enter it full throttle ahead. HSA has remained silent for too long now and no signs of Pro Market penetration. Maybe they need to scrap GCN and re-focus the Arch to OCL compliance. Yeah, GCN is good at number crunching, but it's far from being optimized for it. It's a jack of all trades Arch.

Cheers!

 

[cdrkf]

Yeah, I have to say I was quite surprised the numbers for this Quarter were so bad; I was expecting red, but not bloody red. Intel was impacted as well, but not with this magnitude.

All hopes and dreams of staying alive are deposited in Zen it seems. I will up my stake in saying that they can't make do with a close contender here, they need at least a solid tie. And that is a lot to say, given how far ahead is Intel.

On the graphics department, I wonder if they're still waiting on some sort of change in the market? I mean, CUDA is not losing any track in the GPGPU Pro world and Intel is about to enter it full throttle ahead. HSA has remained silent for too long now and no signs of Pro Market penetration. Maybe they need to scrap GCN and re-focus the Arch to OCL compliance. Yeah, GCN is good at number crunching, but it's far from being optimized for it. It's a jack of all trades Arch.

Cheers!

What would count as a 'solid tie' though?

I mean there are many ways to measure cpu performance, based on the bits I've heard and being optimistic (given their current position) I think the best we can hope for from zen is:

- Lots of cores with SMT, leading to competitive performance in well threaded software e.g. Cinebench multi threaded benchmark. I expect AMD to offer more cores / threads than Intel does at any given price point (e.g. 6 core / 12 thread zen cpu vs 4 core 8 thread in high end consumer).

- Big increase in perf / w, I expect zen at least in multi thread to offer comparable perf / w to equivalent intel cpus, hopefully both on desktop and laptop solutions. The recent work on processors like Carrizo would suggest an emphasis on power consumption so this isn't out of the realms.

- Less per core single thread performance than Intel- hopefully a closer gap than we have now. I'm thinking they'll probably go for aggressive turbo frequencies when dealing with single thread to compensate a bit so whilst IPC will definitely be in Intel's favor the actual single thread gap might be somewhat less than now. Base clock speeds probably more conservative to maintain overall sensible tdp and efficiency.

- Much stronger FP performance than current AMD solutions due to having a full FP unit per core (rather than per 2 cores), hopefully this will close the gap with Intel in games somewhat in both DX11 and DX12 (the latter may benefit more if AMD go the '6 cores vs 4 cores' route as DX12 scales up to about 6 physical cores from what we've seen so far).

Overall that would (for a given price point) mean AMD offers better multi thread performance, slightly (rather than catastrophically like it is now) weaker single thread performance, and proves to be close or equal to Intel for driving games thanks to significant increase in FP capability.

Now the question is, would the above be enough? I certainly think it would provide some arguments to actually consider an AMD cpu again. As much as I like their kit and even have an AMD system at the moment for my main rig I cannot recommend an AMD cpu for many builds because the number of situations where it's an equal or better option are disappearing fast (I mean the Haswell i3 basically makes any FX cpu, even the high end 8 core parts, pretty much redundant).

 

[jdwii]

I just noticed this if this is right Amd gained share.
http://www.cpubenchmark.net/market_share.html

 

[gamerk316]

Yeah, I have to say I was quite surprised the numbers for this Quarter were so bad; I was expecting red, but not bloody red. Intel was impacted as well, but not with this magnitude.

All hopes and dreams of staying alive are deposited in Zen it seems. I will up my stake in saying that they can't make do with a close contender here, they need at least a solid tie. And that is a lot to say, given how far ahead is Intel.

On the graphics department, I wonder if they're still waiting on some sort of change in the market? I mean, CUDA is not losing any track in the GPGPU Pro world and Intel is about to enter it full throttle ahead. HSA has remained silent for too long now and no signs of Pro Market penetration. Maybe they need to scrap GCN and re-focus the Arch to OCL compliance. Yeah, GCN is good at number crunching, but it's far from being optimized for it. It's a jack of all trades Arch.

Cheers!

A tie isn't good enough though. The general consumer isn't going to go out of their way to purchase an AMD CPU if it's "as good" as Intel. To win in the mass market, AMD has to be significantly BETTER then Intel. Which no one here is predicting is going to happen.

AMDs problem is they are trying to fight Intel at every price point, which is preventing them from making either super high end chips, or super low performance chips. They're stuck being "meh" throughout their entire product line.

If I were running AMD, I'd abandon the HEDT and focus entirely on APUs, with the goal of challenging NVIDIA, Qualcomm, and PowerVR in mobile. Of course, that's what I would have done five years ago, right before mobile became huge.

 

[Ranth]

I just noticed this if this is right Amd gained share.
http://www.cpubenchmark.net/market_share.html

Isn't that just the amount tested on passmark and not really marketshare? 7% seems a bit fishy aswell

 

[cdrkf]

@gamerk, your missing a major point though, amds main customers are the oems, end users usually don't care what the cpu is provided it works. AMD need a compelling range of parts that the oems want to integrate to gain market share, and whilst I agree having an outright winning part would be best, having a comparable part and keeping cost a little cheaper than the direct Intel competition (again talking modest difference, eg Intel chip at $200, amd performance equal chip at $190) could entice oems to use them. The unification of platform could also help keep bom costs down for oems which is another argument for them. Point is given amds current position they don't need to dominate the market, just secure enough additional market share to stabilise things. On the home builder side people aren't anti amd and again if they can offer a broadly competitive platform many would use them (I'd update my rig to Zen if it makes sense to do so, for that to happen though it needs to not be a massive compromise)

 

[de5_Roy]

MSI 990FXA GAMING Motherboard Review
http://www.play3r.net/reviews/motherboards/msi-990fxa-gaming-motherboard-review/

AMD spreads not really good news to analysts
http://semiaccurate.com/2015/07/17/amd-spreads-not-really-good-news-analysts/

The last point of note is about the ramp of the AMD ARM CPUs, specifically Seattle. If anyone thinks the first generation of ARM server CPUs is going to be a sales leader, they simply don’t understand the market, how it works, and why it takes the time it does to ramp.

In short Seattle was never meant to be a volume part, and it is going to live up to those expectations. This is not a bad thing though, it is good and means the CPU will meet it’s intended goals, from what we hear it is doing quite well at the moment.

The node transition explains the delay in Zen and will end up with a lot better product for the short delay. AMD’s ARM servers are doing what they should, just not what people wrongly expect them to.

shelds some light on seattle's situation and debunks quite a few baseless stuff that were being spread around using false context. wrong expectaions is ...wrong.

AMD puts out a few goodies amid vague terms
http://semiaccurate.com/2015/05/06/amd-puts-goodies-amid-vague-terms/
relin k to an old article.

TSMC’s 7nm tech will use 10nm elements, production starts in 2018
http://www.kitguru.net/components/graphic-cards/anton-shilov/tsmcs-7nm-tech-will-use-10nm-elements-on-track-for-production-in-2018/
AMD begins development of its third major semi-custom design
http://www.kitguru.net/components/anton-shilov/amd-begins-development-of-its-third-major-semi-custom-design/
TSMC: 10nm is on-track for volume production start in Q4 2016
http://www.kitguru.net/components/anton-shilov/tsmc-10nm-is-on-track-for-volume-production-start-in-q4-2016/
AMD vows to keep R&D investments at appropriate levels
http://www.kitguru.net/components/graphic-cards/anton-shilov/amd-vows-to-keep-rd-investments-at-appropriate-levels/
TSMC begins shipments of chips made using 16nm FinFET process tech
http://www.kitguru.net/components/graphic-cards/anton-shilov/tsmc-begins-shipments-of-chips-using-16nm-finfet-process-technology/
AMD: We have taped out our first FinFET products
http://www.kitguru.net/components/graphic-cards/anton-shilov/amd-we-have-taped-out-our-first-finfet-products/

edit:
seems like early zen-processors were bound for 20nm. i didn't expect that. and the zen+ ones were for finfet - assuming the foundries didn't get their interconnects in a bunch (which they did at 20nm). re-laying the designs for finfet is costing them moniez. in reality, this act is costing everyone moniez from nvidia to qualcomm - any fabless chip designer that work with tsmc and samsung. amd's is being made to look worse because it's going through a tough phase.

 

[cdrkf]

I still maintain turbo frequencies up to about 4ghz are feasible at least for single thread. Even kaveri hits 3.9 on turbo if I remember correctly. We'd need more info on the process though to be certain.

 

[de5_Roy]

I still maintain turbo frequencies up to about 4ghz are feasible at least for single thread. Even kaveri hits 3.9 on turbo if I remember correctly. We'd need more info on the process though to be certain.

it's possible. only if samsung's die has similar electrical characteristics as intel's. intel's 4 GHz cpu (max turbo) is actually quite hot and high power consuming despite having 84-88w rating. a very similar acting die with different power management tech will act differently.

 

[juanrga]

I still maintain turbo frequencies up to about 4ghz are feasible at least for single thread. Even kaveri hits 3.9 on turbo if I remember correctly. We'd need more info on the process though to be certain.

~3.5GHz base and ~4GHz turbo is possible. But Intel is already on 4GHz base.

 

[Cazalan]

Su said there were 35+ designs using Carrizo. The issue is they are waiting for Win10 to release those new models.

The reason they're expecting to gain ground in professional and high end graphics is:

S9170 with 32GB RAM - highest SP/DP single slot GPU (NVidia has no counter for it (yet))
Fury X/Fury Pro - demand higher than supply (even with all the negative reviews)
Fury Nano - August
Fury X2 - This will obliterate every NVidia card

They've worked out the kinks with HBM and have smooth upgrade path to HBM2. (6+ mo ahead of NVidia)


The reason they're excited for Zen is because their products are currently 2 nodes behind and without finfets. Say 2.5-3 nodes depending how you classify finfets. Next year they'll be at ~14nm and with finfets. Now technically Samsun/GF 14nm is a mix node. The BEOL is 20nm. It does not have the same density of Intels 14nm. So they'll be 0.5=1 node behind still. However that is a huge gain for them in closing the gap.

 

[Cazalan]

If I were running AMD, I'd abandon the HEDT and focus entirely on APUs, with the goal of challenging NVIDIA, Qualcomm, and PowerVR in mobile. Of course, that's what I would have done five years ago, right before mobile became huge.

That's kind of what they did. They shifted focus almost entirely to APUs. Which put them in a precarious position when cheap laptop sales are in decline competing with even cheaper tablets. FX cpu didn't get Steamroller or Excavator updates. It also means when they launch their high end GPUs they demo with Intel CPUs. Not a good thing.

They reported increased sales in desktop CPU/APU meaning the gaming industry is strong. China is not buying consoles. They buy desktop PCs for gaming.

 

[de5_Roy]

fury is a boon for people looking to get a headstart on how to use HBM. i am guessing that devs will look forward to getting one - i think fury has an unusual appeal to developers for this reason. devs' and people's familiarity with HBM will trickle all the way down to new apus and possibly future apus and socs.

since fiji was always a low yield product, it's pretty easy to see why demand would exceed supply. amd's long time work with HBM and agreements with sk hynix is getting them early access.

 

[cdrkf]

Juanrga, are Intel at 4ghz base on 14nm already? I don't remember seeing any Intel cpu clocked much higher than 4 ....

3.5 ghz base with 4 turbo is probably the highest I'd expect from Zen so Intel may well have a clock advantage (if they choose to use it of course). I guess they'll wait and see where Zen lands, then tune a few knobs and release something faster if needed to retain the performance lead. If AMD can muster a design fast enough to prompt that by Intel then I'd still say it's a success for them given the last few years they haven't really posed a credible threat at the high end.

 

[8350rocks]

Su said there were 35+ designs using Carrizo. The issue is they are waiting for Win10 to release those new models.

The reason they're expecting to gain ground in professional and high end graphics is:

S9170 with 32GB RAM - highest SP/DP single slot GPU (NVidia has no counter for it (yet))
Fury X/Fury Pro - demand higher than supply (even with all the negative reviews)
Fury Nano - August
Fury X2 - This will obliterate every NVidia card

They've worked out the kinks with HBM and have smooth upgrade path to HBM2. (6+ mo ahead of NVidia)


The reason they're excited for Zen is because their products are currently 2 nodes behind and without finfets. Say 2.5-3 nodes depending how you classify finfets. Next year they'll be at ~14nm and with finfets. Now technically Samsun/GF 14nm is a mix node. The BEOL is 20nm. It does not have the same density of Intels 14nm. So they'll be 0.5=1 node behind still. However that is a huge gain for them in closing the gap.

Pretty much this.

Also, I would like to add, John Byrne, before he left AMD, told me they were planning to likely end up being on 14nm because of 20nm and the node issues. He was expecting to have 14nm CPUs when Intel was at 10nm. It looks like whatever AMD did, they did not listen if they were preparing to move to 20nm, as he had already told me on a separate occasion that the yields were not going to line up with their hopes.

 

[juanrga]

Juanrga, are Intel at 4ghz base on 14nm already? I don't remember seeing any Intel cpu clocked much higher than 4 ....

3.5 ghz base with 4 turbo is probably the highest I'd expect from Zen so Intel may well have a clock advantage (if they choose to use it of course). I guess they'll wait and see where Zen lands, then tune a few knobs and release something faster if needed to retain the performance lead. If AMD can muster a design fast enough to prompt that by Intel then I'd still say it's a success for them given the last few years they haven't really posed a credible threat at the high end.

i7 6700K has base clock of 4GHz and Skylake refresh will increase that. Therefore Zen will be suffer IPC and frequency disadvantages, and will only tie on core count (max. 8 cores).

 

[juanrga]

If I were running AMD, I'd abandon the HEDT and focus entirely on APUs, with the goal of challenging NVIDIA, Qualcomm, and PowerVR in mobile. Of course, that's what I would have done five years ago, right before mobile became huge.

That's kind of what they did. They shifted focus almost entirely to APUs. Which put them in a precarious position when cheap laptop sales are in decline competing with even cheaper tablets. FX cpu didn't get Steamroller or Excavator updates. It also means when they launch their high end GPUs they demo with Intel CPUs. Not a good thing.

They reported increased sales in desktop CPU/APU meaning the gaming industry is strong. China is not buying consoles. They buy desktop PCs for gaming.

APUs is what maintained AMD afloat thanks to console contracts. AMD problem in laptops is caused by inefficient architectures plus a process node disadvantage, which means their mobile APUs only can sell for cheap to be used on cheap products that will compete with dozen of other cheap products and will not sell in enough volume.

Their late tablet adventure finished fast because Jaguar-based APUs are far from competitive.

Steamroller didn't appear on FX CPUs because the IPC gain over Piledriver wasn't high enough to compensate for the frequencies reduction associated to the 28SHP node. Similar thoughts about Excavator.

Even if AMD managed magically to maintain frequencies on imagined Steamroller/Excavator FX CPUs, Intel CPUs would continue to be much faster and AMD would continue using Intel CPUs in its GPU demos because Nvidia does.

Finally about 90% or so of "desktop gaming PCs" have APUs not CPUs.

 

[cdrkf]

@juan, that last star is
a little misleading in that the majority of those apus have the igpu disabled in favour of a dgpu. That may change in the future, but currently an apu or dcpu has no impact for a gaming machine.

 

[jimmysmitty]

Yeah, I have to say I was quite surprised the numbers for this Quarter were so bad; I was expecting red, but not bloody red. Intel was impacted as well, but not with this magnitude.

All hopes and dreams of staying alive are deposited in Zen it seems. I will up my stake in saying that they can't make do with a close contender here, they need at least a solid tie. And that is a lot to say, given how far ahead is Intel.

On the graphics department, I wonder if they're still waiting on some sort of change in the market? I mean, CUDA is not losing any track in the GPGPU Pro world and Intel is about to enter it full throttle ahead. HSA has remained silent for too long now and no signs of Pro Market penetration. Maybe they need to scrap GCN and re-focus the Arch to OCL compliance. Yeah, GCN is good at number crunching, but it's far from being optimized for it. It's a jack of all trades Arch.

Cheers!

What would count as a 'solid tie' though?

I mean there are many ways to measure cpu performance, based on the bits I've heard and being optimistic (given their current position) I think the best we can hope for from zen is:

- Lots of cores with SMT, leading to competitive performance in well threaded software e.g. Cinebench multi threaded benchmark. I expect AMD to offer more cores / threads than Intel does at any given price point (e.g. 6 core / 12 thread zen cpu vs 4 core 8 thread in high end consumer).

- Big increase in perf / w, I expect zen at least in multi thread to offer comparable perf / w to equivalent intel cpus, hopefully both on desktop and laptop solutions. The recent work on processors like Carrizo would suggest an emphasis on power consumption so this isn't out of the realms.

- Less per core single thread performance than Intel- hopefully a closer gap than we have now. I'm thinking they'll probably go for aggressive turbo frequencies when dealing with single thread to compensate a bit so whilst IPC will definitely be in Intel's favor the actual single thread gap might be somewhat less than now. Base clock speeds probably more conservative to maintain overall sensible tdp and efficiency.

- Much stronger FP performance than current AMD solutions due to having a full FP unit per core (rather than per 2 cores), hopefully this will close the gap with Intel in games somewhat in both DX11 and DX12 (the latter may benefit more if AMD go the '6 cores vs 4 cores' route as DX12 scales up to about 6 physical cores from what we've seen so far).

Overall that would (for a given price point) mean AMD offers better multi thread performance, slightly (rather than catastrophically like it is now) weaker single thread performance, and proves to be close or equal to Intel for driving games thanks to significant increase in FP capability.

Now the question is, would the above be enough? I certainly think it would provide some arguments to actually consider an AMD cpu again. As much as I like their kit and even have an AMD system at the moment for my main rig I cannot recommend an AMD cpu for many builds because the number of situations where it's an equal or better option are disappearing fast (I mean the Haswell i3 basically makes any FX cpu, even the high end 8 core parts, pretty much redundant).

1. We will have to see. I don't think AMD can offer more cores at the same price if Zen is to be a normal CPU. The reason they could with BD/PD was because the extra costs for the secondary cores was not that much. IT cost more than SMT but not more than a actual second core. Remember that the 14nm they will be using will be new and more cores means lower yields so I expect them to match Intel in cores.

2. Will have to wait and see. AMDs 65nm was not a good process node, was one of the big issues with K10 as it had leakage issues. Their 45nm was better but Intels was still better thanks to the HK/MG. We also have to consider this is going to be the first time AMD uses FinFETs while 14nm is Intels second generation so they have a bit of an advantage there.

3. Considering the rumors I see it around Haswell at best. Which will not allow for better pricing but the same as usual, just under Intel probably top end matching top end i5/i7.

4. Anything they do can be better than BD/PD. As for DX12, from what we have also seen games using something like Mantle/DX12 also do not benefit as much on higher end systems. Low end CPUs with top of the line GPUs do as do low end GPUs with low end CPUs. But when it was tested, Mantle offered little to no performance gains for people with i5/i7s and top end AMD GPUs. I tested it myself and it was within 5-10FPS on any Mantle game. Of course we have to wait and see how DX12 will handle it in the end but we have yet to get an official DX12 game worth testing. Maybe in 3-6 months.

Su said there were 35+ designs using Carrizo. The issue is they are waiting for Win10 to release those new models.

The reason they're expecting to gain ground in professional and high end graphics is:

S9170 with 32GB RAM - highest SP/DP single slot GPU (NVidia has no counter for it (yet))
Fury X/Fury Pro - demand higher than supply (even with all the negative reviews)
Fury Nano - August
Fury X2 - This will obliterate every NVidia card

They've worked out the kinks with HBM and have smooth upgrade path to HBM2. (6+ mo ahead of NVidia)


The reason they're excited for Zen is because their products are currently 2 nodes behind and without finfets. Say 2.5-3 nodes depending how you classify finfets. Next year they'll be at ~14nm and with finfets. Now technically Samsun/GF 14nm is a mix node. The BEOL is 20nm. It does not have the same density of Intels 14nm. So they'll be 0.5=1 node behind still. However that is a huge gain for them in closing the gap.

The S9170 is nice but they still haven't countered CUDA and that is major.Having software support means a lot more than having the highest SP/DP out there. If anything AMD has always had higher TFLOPS capabilities than nVidia but I have always said that there is more to it than that and that does not always translate to performance, hence why a R9 290X has equal TFLOPS to a 980Ti yet doesn't perform the same.

How do we know they are in high demand? I have never seen a launch, short of the Bitcoin craze period, where they were sold out or out of stock most of the time. I think yields are just really low.

I think Nano will be interesting.

As for Fury X2, I will wait and see. They say the radiator can handle 500W but dual Fury GPUs is a lot of power and heat. I think it iwll be pushing that radiator close to its limits if they even do it.

 

[8350rocks]

1. We will have to see. I don't think AMD can offer more cores at the same price if Zen is to be a normal CPU. The reason they could with BD/PD was because the extra costs for the secondary cores was not that much. IT cost more than SMT but not more than a actual second core. Remember that the 14nm they will be using will be new and more cores means lower yields so I expect them to match Intel in cores.

2. Will have to wait and see. AMDs 65nm was not a good process node, was one of the big issues with K10 as it had leakage issues. Their 45nm was better but Intels was still better thanks to the HK/MG. We also have to consider this is going to be the first time AMD uses FinFETs while 14nm is Intels second generation so they have a bit of an advantage there.

3. Considering the rumors I see it around Haswell at best. Which will not allow for better pricing but the same as usual, just under Intel probably top end matching top end i5/i7.

4. Anything they do can be better than BD/PD. As for DX12, from what we have also seen games using something like Mantle/DX12 also do not benefit as much on higher end systems. Low end CPUs with top of the line GPUs do as do low end GPUs with low end CPUs. But when it was tested, Mantle offered little to no performance gains for people with i5/i7s and top end AMD GPUs. I tested it myself and it was within 5-10FPS on any Mantle game. Of course we have to wait and see how DX12 will handle it in the end but we have yet to get an official DX12 game worth testing. Maybe in 3-6 months.

---------------------------------
The S9170 is nice but they still haven't countered CUDA and that is major.Having software support means a lot more than having the highest SP/DP out there. If anything AMD has always had higher TFLOPS capabilities than nVidia but I have always said that there is more to it than that and that does not always translate to performance, hence why a R9 290X has equal TFLOPS to a 980Ti yet doesn't perform the same.

How do we know they are in high demand? I have never seen a launch, short of the Bitcoin craze period, where they were sold out or out of stock most of the time. I think yields are just really low.

I think Nano will be interesting.

As for Fury X2, I will wait and see. They say the radiator can handle 500W but dual Fury GPUs is a lot of power and heat. I think it iwll be pushing that radiator close to its limits if they even do it.

Here is the thing regarding (1): AMD is shooting for top SKU to be 16 cores if they can keep the die size manageable...if not 16 then 12, if not 12 then 8. Nothing less than 8 physical cores will be the flagship from my information...but it will be 8C/16T versus 8C/8T now.

Regarding (2): This may be AMDs first attempt with FinFETs, however, I expect this to be less an issue with design, and more an issue with "fine tuning" so to speak. I would imagine they get everything they want to do in, and have no issues there...I just expect that the second generation (Zen+) will likely see some pretty good gains from low hanging fruit, plus optimizing current uarch after observation/R&D.

As far as pricing goes...sure...you may potentially have less IPC...however...if you can get an 8C/16T AMD for the same money as an i7 Quad intel...and AMD catches up to haswell...that is a might powerful processor...especially if power consumption is inline. If they get SMT nailed on the first run, they may even see a large uptick in server penetration and market share.

CUDA is a farce, you can do all that with OCL, and AMD would benefit hugely having someone to write a lot of these programs that are broadly applicable to specific markets with heavy CUDA saturation.

 

[jimmysmitty]

1. We will have to see. I don't think AMD can offer more cores at the same price if Zen is to be a normal CPU. The reason they could with BD/PD was because the extra costs for the secondary cores was not that much. IT cost more than SMT but not more than a actual second core. Remember that the 14nm they will be using will be new and more cores means lower yields so I expect them to match Intel in cores.

2. Will have to wait and see. AMDs 65nm was not a good process node, was one of the big issues with K10 as it had leakage issues. Their 45nm was better but Intels was still better thanks to the HK/MG. We also have to consider this is going to be the first time AMD uses FinFETs while 14nm is Intels second generation so they have a bit of an advantage there.

3. Considering the rumors I see it around Haswell at best. Which will not allow for better pricing but the same as usual, just under Intel probably top end matching top end i5/i7.

4. Anything they do can be better than BD/PD. As for DX12, from what we have also seen games using something like Mantle/DX12 also do not benefit as much on higher end systems. Low end CPUs with top of the line GPUs do as do low end GPUs with low end CPUs. But when it was tested, Mantle offered little to no performance gains for people with i5/i7s and top end AMD GPUs. I tested it myself and it was within 5-10FPS on any Mantle game. Of course we have to wait and see how DX12 will handle it in the end but we have yet to get an official DX12 game worth testing. Maybe in 3-6 months.

---------------------------------
The S9170 is nice but they still haven't countered CUDA and that is major.Having software support means a lot more than having the highest SP/DP out there. If anything AMD has always had higher TFLOPS capabilities than nVidia but I have always said that there is more to it than that and that does not always translate to performance, hence why a R9 290X has equal TFLOPS to a 980Ti yet doesn't perform the same.

How do we know they are in high demand? I have never seen a launch, short of the Bitcoin craze period, where they were sold out or out of stock most of the time. I think yields are just really low.

I think Nano will be interesting.

As for Fury X2, I will wait and see. They say the radiator can handle 500W but dual Fury GPUs is a lot of power and heat. I think it iwll be pushing that radiator close to its limits if they even do it.

Here is the thing regarding (1): AMD is shooting for top SKU to be 16 cores if they can keep the die size manageable...if not 16 then 12, if not 12 then 8. Nothing less than 8 physical cores will be the flagship from my information...but it will be 8C/16T versus 8C/8T now.

Regarding (2): This may be AMDs first attempt with FinFETs, however, I expect this to be less an issue with design, and more an issue with "fine tuning" so to speak. I would imagine they get everything they want to do in, and have no issues there...I just expect that the second generation (Zen+) will likely see some pretty good gains from low hanging fruit, plus optimizing current uarch after observation/R&D.

As far as pricing goes...sure...you may potentially have less IPC...however...if you can get an 8C/16T AMD for the same money as an i7 intel...and AMD catches up to haswell...that is a might powerful processor...especially if power consumption is inline. If they get SMT nailed on the first run, they may even see a large uptick in server penetration and market share.

CUDA is a farce, you can do all that with OCL, and AMD would benefit hugely having someone to write a lot of these programs that are broadly applicable to specific markets with heavy CUDA saturation.

1. Right now it is 4M/8T not 8C/8T. I don't count their CPUs as 8 cores. And depending on the yields of the new 14nm will determine pricing as will more cores. 16 cores would have to be the very top end and again low yields would make it harder to put a lower price if true. Still waiting to see official specs.

2. That is what I am getting at, again Intel has already had time to fine tune the use of FinFETS and by the time Zen is out they will have "tuned" their 14nm for better yields/lower voltages. However without seeing the design we wont know this for sure. Remember, BD looked good on paper but is a leaky bastard.

3. Yes it would be but again more cores does not always mean better. AMDs Phenom II X6 wasn't better and still is not. It all depends on a lot of factors. This is why I hate rumors, they never bring anything good but hype for something to never live up to.

4. CUDAs advantage is that they help in writing the software to take advantage of it and is guaranteed to be updated. Problem with most open source software is it depends on someone having an interest in it to update and write for it and if AMD is not going to put the foot work to do so, no one is going to do it for free. It is the same with Gameworks. While I hate the proprietary nature of things like these, I can't hate that nVidia pushes to get their ideas and tech into games a bit more than AMD.

That said, Zen is nothing but rumors and hope. If we let the hype build up it could just turn into another K10 if it has an issue or falls short. I say we just hope it is decent and not some magical cure all for AMD because it is very certain that it could flop.