AMD CPUs, SoC Rumors and Speculations Temp. thread 2

[juanrga]

AMD Confident Zen Will Deliver on Performance
http://www.eteknix.com/amd-confident-zen-will-deliver-performance/
excuse the 40% ipc boost bait.

The more they insist, the more doubts they raise.

 

[juanrga]

This recent comment doesn't prove my hypothesis of Keller leaving due to strong discrepancies with Lisa Su due to K12 status, but it seems I was close, very close!

The idea of designing a different arch cpu was there before Jim Keller rejoin. But the arch/model design stage was done early 2013. Although Rory Read hired him for K12 more than ZEN, he's still quite influential on ZEN. Lisa Su sort of made him and his Sunnyvale folks a little bit disappointed because of de-prioritizing K12. But ZEN, so far it's really "on track".

😀

 

[de5_Roy]

This recent comment doesn't prove my hypothesis of Keller leaving due to strong discrepancies with Lisa Su due to K12 status, but it seems I was close, very close!

The idea of designing a different arch cpu was there before Jim Keller rejoin. But the arch/model design stage was done early 2013. Although Rory Read hired him for K12 more than ZEN, he's still quite influential on ZEN. Lisa Su sort of made him and his Sunnyvale folks a little bit disappointed because of de-prioritizing K12. But ZEN, so far it's really "on track".

😀

doesn't prove or disprove anything, really. it's an analyst's opinion.... if that's what you're basing your claim upon. not facts.
any official statements/logical proof on read hiring keller mainly for k12?

 

[jdwii]

I think i could see it being more true then false since keller already worked heavily on Arm designs before hand at Apple obviously that had a major part of his work at Amd.

Also i find it sad that so many people think Jim keller leaving and Amd breaking up Radeon into its own company is such great news to me it seems like a horrible idea.

Be nice if Jim keller worked at Amd forever.

As for the Zen article if its 40% more powerful that would still put it a little behind haswell in IPC, however on 14nm and they keep claiming it will be more efficient, heard something previously about a 3.5ghz clock range.

Wonder how Zen will be in performance per watt compared to intel? If they are extremely close they might have a win in the server market.

 

[jimmysmitty]

I think i could see it being more true then false since keller already worked heavily on Arm designs before hand at Apple obviously that had a major part of his work at Amd.

Also i find it sad that so many people think Jim keller leaving and Amd breaking up Radeon into its own company is such great news to me it seems like a horrible idea.

Be nice if Jim keller worked at Amd forever.

As for the Zen article if its 40% more powerful that would still put it a little behind haswell in IPC, however on 14nm and they keep claiming it will be more efficient, heard something previously about a 3.5ghz clock range.

Wonder how Zen will be in performance per watt compared to intel? If they are extremely close they might have a win in the server market.

Hard to say it will win in the server with the Skylake Purley announcement it is hard to even say if AMD has much of a chance in the server market.

 

[Cazalan]

From my understanding, K12 and Zen were been develop in conjunction.
It is probably waiting. AMD did delay it to 2017, as of the last I have heard.

The ARM server ecosystem is still a mess from what I've seen so it's no surprise K12 was delayed. They have enough ARMv8 platforms to develop the ecosystem so releasing additional product doesn't really help the situation. It would just be more product sitting on the shelf as you can only sell so many development systems. When the ecosystem is more mature and companies are willing to buy, they will make the product available. That is just economics 101, supply n demand.

 

[de5_Roy]

The ARM server ecosystem is still a mess from what I've seen so it's no surprise K12 was delayed.

it seems to me that amd finished k12 development quite earlier than it did with zen and is keeping it internal to amd itself. x86 and arm cores are developed very differently. arm's development model can reduce the development time significantly. this way, amd gets time to hype zen and gloss over keller's departure. i also wouldn't be surprised of any leak about arm trying to screw over amd with it's a72 cores and those cores having looots of functional similarities with k12... 😗

 

[Cazalan]

Anyone can make mistakes, especially when your tools have issues. I don't know what you mean by "daunted", but from what I can interpret, Zen is coming along just fine and the Team behind it has confidence.

In any case, if you make such a blunder, of course that is going to resonate for a good while in your head. You're only unworthy if you can't get your act together and keep on moving.

In regards to the shake this announcement has made, a lot of people in comments that remembers/knows AMDs history say Jim Keller did the same with K8's dev cycle. He left for Apple before the actual CPU was in production. I think it's worthy or not such a long shot to think that Zen will be good. How good, it's about to be seen, but at least not a bust.

Cheers!

One must also remember his age. He's 57 with a lot of money. His contract ended and he is ready to either retire or do something different with his life.

I don't think it affects Zen/K12 one way or the other. The specifications have been made long long ago.

 

[Cazalan]

The ARM server ecosystem is still a mess from what I've seen so it's no surprise K12 was delayed.

it seems to me that amd finished k12 development quite earlier than it did with zen and is keeping it internal to amd itself. x86 and arm cores are developed very differently. arm's development model can reduce the development time significantly. this way, amd gets time to hype zen and gloss over keller's departure. i also wouldn't be surprised of any leak about arm trying to screw over amd with it's a72 cores and those cores having looots of functional similarities with k12... 😗

It wouldn't surprise me if they are done with K12 and holding it for other reasons. Likewise with Fiji it appears to have been done 6-12mo before SKHynix was ready. Thus the lack of HDMI 2.0. That's the love/hate relationship of having to deal with partners and not controlling it 100%. GF cancelling nodes on them. The small guy has little control of those things.

 

[Cazalan]

Hard to say it will win in the server with the Skylake Purley announcement it is hard to even say if AMD has much of a chance in the server market.

At the highest end they are unlikely to compete with Purley. There is still a huge range of server CPUs though and Intel has to keep their margins high to feed the machine.

 

[Clariska]

The day that Microsoft buys amd are the day i move over to intel......even with there high prices as AMD prices will go through the roof with ms sitting behind them.

 

[juanrga]

I think i could see it being more true then false since keller already worked heavily on Arm designs before hand at Apple obviously that had a major part of his work at Amd.

Also i find it sad that so many people think Jim keller leaving and Amd breaking up Radeon into its own company is such great news to me it seems like a horrible idea.

Be nice if Jim keller worked at Amd forever.

As for the Zen article if its 40% more powerful that would still put it a little behind haswell in IPC, however on 14nm and they keep claiming it will be more efficient, heard something previously about a 3.5ghz clock range.

Wonder how Zen will be in performance per watt compared to intel? If they are extremely close they might have a win in the server market.

Yes, some people always spin bad news as if were good news.

Now I start to believe that Feldman, Rory, and Keller left AMD by the same reason. I only have a doubt now. Has Lisa Su only delayed K12 or canceled K12 at Keller's departure?

If K12 was delayed, AMD still has an infinitesimal opportunity to do a comeback. If Lisa Su canceled K12, then AMD is in my opinion death.

About Zen, the problem is not the IPC. The problem will be the clocks. And the performance per watt depends seriously of the node used and Samsung/GF 14nm is clearly inferior to Intel 14nm.

 

[-Fran-]

Yes, some people always spin bad news as if were good news.

Now I start to believe that Feldman, Rory, and Keller left AMD by the same reason. I only have a doubt now. Has Lisa Su only delayed K12 or canceled K12 at Keller's departure?

If K12 was delayed, AMD still has an infinitesimal opportunity to do a comeback. If Lisa Su canceled K12, then AMD is in my opinion death.

About Zen, the problem is not the IPC. The problem will be the clocks. And the performance per watt depends seriously of the node used and Samsung/GF 14nm is clearly inferior to Intel 14nm.

And some people always spin *any* news as bad news. It always comes full circle, isn't it?

In any case, You're again shooting blanks here, Juan. Why canceling K12 would mean "doom" for AMD? Care to elaborate on that? Plus, what is it with Lisa Su? Do you think everyone hates her or something?

Also, efficiency needs to be on par for important workloads. I do have my doubts with GF/Sammys node like you do, but I don't have any chips showing how big is the difference. I remember you putting some information about each process'es density and some slight thermal data.

Cheers!

 

[gamerk316]

The day that Microsoft buys amd are the day i move over to intel......even with there high prices as AMD prices will go through the roof with ms sitting behind them.

MSFT has been trying to get out of the HW business since Balmer left; they're a services company now. There is literally ZERO reason MSFT would want AMD, and the MSFT purchase is more smoke from people who honestly don't have a clue.

 

[jimmysmitty]

Hard to say it will win in the server with the Skylake Purley announcement it is hard to even say if AMD has much of a chance in the server market.

At the highest end they are unlikely to compete with Purley. There is still a huge range of server CPUs though and Intel has to keep their margins high to feed the machine.

Both companies had high margins in the server market which is why I am saying AMD needs to be highly competitive in that market, not just the low end. They used to focus on it. K8 and K10 were all server designed chips first, then desktop came second.

Point is, Purley looks to be a monster of a server chip. AMD needs that too.

 

[de5_Roy]

MSFT has been trying to get out of the HW business since Balmer left; they're a services company now. There is literally ZERO reason MSFT would want AMD, and the MSFT purchase is more smoke from people who honestly don't have a clue.

isn't the failure of the HW biz the reason ballmer [strike]got fired[/strike] left M$? however, i do see why m$ would want amd, though. it's teh patents and a deeper foothold up arm's !@##. it sounds gross but makes sense from business standpoint. also prevents p-trolls like nvidia off m$.


@JS: amd is simply not capable of building something like skylake-EX/EP. but they sure can try to undercut it and get through using TCO advantages. there's a slight chance of that... as long as zen's 'EP counterpart is up to snuff. i am cautiously optimistic that it will be for a small amount of time until intel releases broadwell-DE (xeon-d)'s successor.

 

[Vogner16]

MSFT has been trying to get out of the HW business since Balmer left; they're a services company now. There is literally ZERO reason MSFT would want AMD, and the MSFT purchase is more smoke from people who honestly don't have a clue.

isn't the failure of the HW biz the reason ballmer [strike]got fired[/strike] left M$? however, i do see why m$ would want amd, though. it's teh patents and a deeper foothold up arm's !@##. it sounds gross but makes sense from business standpoint. also prevents p-trolls like nvidia off m$.


@JS: amd is simply not capable of building something like skylake-EX/EP. but they sure can try to undercut it and get through using TCO advantages. there's a slight chance of that... as long as zen's 'EP counterpart is up to snuff. i am cautiously optimistic that it will be for a small amount of time until intel releases broadwell-DE (xeon-d)'s successor.

also remember the key point to the hardware game. there is no such thing as a bad product, only bad prices.
balmer must have priced things wrong. that's it.

even if zen is ipc of ivy and single thread perf becomes close to intel, amd has ALWAYS had much better multiplier ratios than Intel for example ivy was around 4.9 when piledriver was 6. if intel core does 1.3 work and amd does .9 work the equate to close to each other in multithread perf. (4.9 x 1.3~=~6 x .9) (don't check that math I made numbers up just proves a point)

point is even if IPC of zen is poor (to ivy spec), and amd continues to hold their great multiplier ratios, they will have a very viable competitor to skylake top end i7 cpu's maybe even haswell EP

 

[jimmysmitty]

MSFT has been trying to get out of the HW business since Balmer left; they're a services company now. There is literally ZERO reason MSFT would want AMD, and the MSFT purchase is more smoke from people who honestly don't have a clue.

isn't the failure of the HW biz the reason ballmer [strike]got fired[/strike] left M$? however, i do see why m$ would want amd, though. it's teh patents and a deeper foothold up arm's !@##. it sounds gross but makes sense from business standpoint. also prevents p-trolls like nvidia off m$.


@JS: amd is simply not capable of building something like skylake-EX/EP. but they sure can try to undercut it and get through using TCO advantages. there's a slight chance of that... as long as zen's 'EP counterpart is up to snuff. i am cautiously optimistic that it will be for a small amount of time until intel releases broadwell-DE (xeon-d)'s successor.

also remember the key point to the hardware game. there is no such thing as a bad product, only bad prices.
balmer must have priced things wrong. that's it.

even if zen is ipc of ivy and single thread perf becomes close to intel, amd has ALWAYS had much better multiplier ratios than Intel for example ivy was around 4.9 when piledriver was 6. if intel core does 1.3 work and amd does .9 work the equate to close to each other in multithread perf. (4.9 x 1.3~=~6 x .9) (don't check that math I made numbers up just proves a point)

point is even if IPC of zen is poor (to ivy spec), and amd continues to hold their great multiplier ratios, they will have a very viable competitor to skylake top end i7 cpu's maybe even haswell EP

What do you mean multiplier ratios?

 

[8350rocks]

http://hexus.net/tech/news/cpu/86585-legendary-cpu-architect-jim-keller-leaves-amd/

Jim Keller left AMD. Thoughts?

Honestly not the least bit surprised, Jim went to Apple to design chips, he went to AMD to design chips. He is more or less free lancing at this point and likes the challenge of creating new designs. With the Zen core, this is going to be the basis for fine tuning for years to come beyond inception. AMD will not be designing a new from scratch core for at least a few iterations past Zen/K12 and so the design portion Keller most enjoys is over.

This was honestly bound to happen at some point, and the timing syncs up with estimates for tape outs, etc. This is nothing negative from the perspective of design, as the work he did would have already been completed when the tape outs finish.

 

[de5_Roy]

also remember the key point to the hardware game. there is no such thing as a bad product, only bad prices.
balmer must have priced things wrong. that's it.

even if zen is ipc of ivy and single thread perf becomes close to intel, amd has ALWAYS had much better multiplier ratios than Intel for example ivy was around 4.9 when piledriver was 6. if intel core does 1.3 work and amd does .9 work the equate to close to each other in multithread perf. (4.9 x 1.3~=~6 x .9) (don't check that math I made numbers up just proves a point)

point is even if IPC of zen is poor (to ivy spec), and amd continues to hold their great multiplier ratios, they will have a very viable competitor to skylake top end i7 cpu's maybe even haswell EP

i avoid ipc discussions like i avoid believing whyreadthissite'snewsoftech.com's exterp analyses.
m$ did price their devices wrong and er... paid for it..or so to speak... but surface is really a bad product overall due to no small thanks to both m$ and intel. the thing is built for profit.


@8350rocks: i couldn't help but notice that no one, i mean almost no one was talking about k12 or it's present status while talking about keller's leave. even if anyone did, s/he slightly touched upon it instead of going into details. also weird is the very, very similar tone of all of the major sites, including techpowerup and tech report who beached and wined about not getting a fury nano sample for review ( tpu somehow did end up getting one...).
should i read into this? (geddit? read as in the former ceo..hehe amd joke)



.....

 

[8350rocks]

also remember the key point to the hardware game. there is no such thing as a bad product, only bad prices.
balmer must have priced things wrong. that's it.

even if zen is ipc of ivy and single thread perf becomes close to intel, amd has ALWAYS had much better multiplier ratios than Intel for example ivy was around 4.9 when piledriver was 6. if intel core does 1.3 work and amd does .9 work the equate to close to each other in multithread perf. (4.9 x 1.3~=~6 x .9) (don't check that math I made numbers up just proves a point)

point is even if IPC of zen is poor (to ivy spec), and amd continues to hold their great multiplier ratios, they will have a very viable competitor to skylake top end i7 cpu's maybe even haswell EP

i avoid ipc discussions like i avoid believing whyreadthissite'snewsoftech.com's exterp analyses.
m$ did price their devices wrong and er... paid for it..or so to speak... but surface is really a bad product overall due to no small thanks to both m$ and intel. the thing is built for profit.


@8350rocks: i couldn't help but notice that no one, i mean almost no one was talking about k12 or it's present status while talking about keller's leave. even if anyone did, s/he slightly touched upon it instead of going into details. also weird is the very, very similar tone of all of the major sites, including techpowerup and tech report who beached and wined about not getting a fury nano sample for review ( tpu somehow did end up getting one...).
should i read into this? (geddit? read as in the former ceo..hehe amd joke)



.....

The only thing you should read into, is that K12 as a HEDT option for a CPU is dead.

As a server option...maybe...still...

However, the big thing is going to be that the major players in servers/commercial are not pushing ARM, they are not excited about pushing ARM, and ARM is not seeing the "MOAR efficiency" argument unfold.

The excitement over ARM is over.

What happened? As was said many times before by many in this thread, by the time you make an ARM chip that competes with a complex x86-64 chip for performance, you end up with a chip that consumes just as much power for similar/worse performance.

ARM does not scale upwardly in commercial as well as many had hoped, and AMD is not wasting money on ARM for a lame duck that never was. In addition, it is taking much longer to port software, because no one is going to port software until the hardware hits critical mass. However, the hardware will not hit critical mass until the software is there...vicious cycle. AMD is not missing out by launch an ARM server processor a year later, and they and everyone else know it as well. As for ARM desktop...do not start holding your breath...Guiness Records would be on your doorstep to follow your progress long before ARM ever matures into something viable as a desktop option.

 

[juanrga]

Yes, some people always spin bad news as if were good news.

Now I start to believe that Feldman, Rory, and Keller left AMD by the same reason. I only have a doubt now. Has Lisa Su only delayed K12 or canceled K12 at Keller's departure?

If K12 was delayed, AMD still has an infinitesimal opportunity to do a comeback. If Lisa Su canceled K12, then AMD is in my opinion death.

About Zen, the problem is not the IPC. The problem will be the clocks. And the performance per watt depends seriously of the node used and Samsung/GF 14nm is clearly inferior to Intel 14nm.

And some people always spin *any* news as bad news. It always comes full circle, isn't it?

In any case, You're again shooting blanks here, Juan. Why canceling K12 would mean "doom" for AMD? Care to elaborate on that? Plus, what is it with Lisa Su? Do you think everyone hates her or something?

I think I explained this enough number of times.

 

[juanrga]

http://hexus.net/tech/news/cpu/86585-legendary-cpu-architect-jim-keller-leaves-amd/

Jim Keller left AMD. Thoughts?

Honestly not the least bit surprised, Jim went to Apple to design chips, he went to AMD to design chips. He is more or less free lancing at this point and likes the challenge of creating new designs. With the Zen core, this is going to be the basis for fine tuning for years to come beyond inception. AMD will not be designing a new from scratch core for at least a few iterations past Zen/K12 and so the design portion Keller most enjoys is over.

This was honestly bound to happen at some point, and the timing syncs up with estimates for tape outs, etc. This is nothing negative from the perspective of design, as the work he did would have already been completed when the tape outs finish.

Except that Keller left AMD without finishing his work. Also Zen is not still tapeout.

 

[juanrga]

also remember the key point to the hardware game. there is no such thing as a bad product, only bad prices.
balmer must have priced things wrong. that's it.

even if zen is ipc of ivy and single thread perf becomes close to intel, amd has ALWAYS had much better multiplier ratios than Intel for example ivy was around 4.9 when piledriver was 6. if intel core does 1.3 work and amd does .9 work the equate to close to each other in multithread perf. (4.9 x 1.3~=~6 x .9) (don't check that math I made numbers up just proves a point)

point is even if IPC of zen is poor (to ivy spec), and amd continues to hold their great multiplier ratios, they will have a very viable competitor to skylake top end i7 cpu's maybe even haswell EP

i avoid ipc discussions like i avoid believing whyreadthissite'snewsoftech.com's exterp analyses.
m$ did price their devices wrong and er... paid for it..or so to speak... but surface is really a bad product overall due to no small thanks to both m$ and intel. the thing is built for profit.


@8350rocks: i couldn't help but notice that no one, i mean almost no one was talking about k12 or it's present status while talking about keller's leave. even if anyone did, s/he slightly touched upon it instead of going into details. also weird is the very, very similar tone of all of the major sites, including techpowerup and tech report who beached and wined about not getting a fury nano sample for review ( tpu somehow did end up getting one...).
should i read into this? (geddit? read as in the former ceo..hehe amd joke)



.....

The only thing you should read into, is that K12 as a HEDT option for a CPU is dead.

As a server option...maybe...still...

However, the big thing is going to be that the major players in servers/commercial are not pushing ARM, they are not excited about pushing ARM, and ARM is not seeing the "MOAR efficiency" argument unfold.

The excitement over ARM is over.

What happened? As was said many times before by many in this thread, by the time you make an ARM chip that competes with a complex x86-64 chip for performance, you end up with a chip that consumes just as much power for similar/worse performance.

ARM does not scale upwardly in commercial as well as many had hoped, and AMD is not wasting money on ARM for a lame duck that never was. In addition, it is taking much longer to port software, because no one is going to port software until the hardware hits critical mass. However, the hardware will not hit critical mass until the software is there...vicious cycle. AMD is not missing out by launch an ARM server processor a year later, and they and everyone else know it as well. As for ARM desktop...do not start holding your breath...Guiness Records would be on your doorstep to follow your progress long before ARM ever matures into something viable as a desktop option.

K12 did never target desktops. The target was server. In fact all companies: APM, Broadcom, Qualcomm,... target servers.

And Keller disagrees with what you say about K12. Not only he stated the reasons why ARMv8 ISA is more efficient, but he even did hint that K12 was faster than Zen.

P.S.: multi-reply doesn't work for me. I have to reply individually!!!!

 

[Cazalan]

isn't the failure of the HW biz the reason ballmer [strike]got fired[/strike] left M$? however, i do see why m$ would want amd, though. it's teh patents and a deeper foothold up arm's !@##. it sounds gross but makes sense from business standpoint. also prevents p-trolls like nvidia off m$.


@JS: amd is simply not capable of building something like skylake-EX/EP. but they sure can try to undercut it and get through using TCO advantages. there's a slight chance of that... as long as zen's 'EP counterpart is up to snuff. i am cautiously optimistic that it will be for a small amount of time until intel releases broadwell-DE (xeon-d)'s successor.

Both the XBox and Surface are making money now. They are more inclined to continue those hardware products. As opposed to the phone division which got quickly hacked up.