AMD CPU speculation... and expert conjecture

[jdwii]

However Juan it seems like Intel was spending way more then Amd at the time i found this article
http://www.economist.com/node/420872
"So far, the effect of AMD's new-found success has been to knock smaller rivals out of the consumer market rather than to cut into Intel's share of the corporate market. That could change if AMD maintains the past year's triumphs. But what should be most worrying for Intel is increasing evidence that AMD, with a research-and-development budget less than 15% of Intel's, may actually be out-innovating it."

If we look at current trends Amd is spending 8.8% as much as Intel. I still say Intel is spending that in one area more then Amd is which is trying to compete with Nvidia and Intel.

 

[Cazalan]

Voltage is a function of frequency. As explained in my article about power consumption of CPUs, substituting the (very good) approximation V ~ f on the expression for the dynamic power gives the well-known cubic dependence that is quoted in many sites

https://software.intel.com/en-us/articles/the-new-era-of-tera-scale-computing

A concrete case. First the almost linear dependence of voltage with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/VoltageCurve.png

and then the almost cubic dependence of (dynamic) power with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/PtotalVccTGHzDataGraph.png

I see your rooky mistake now. You took the dumbed down example from Intel where they normalized voltage and frequency to 1, and assumed that is valid for real processors as well. It was just an example so the layman can follow along.

 

[Cazalan]

The reality is Intel is spending 10 times the amount as Amd regardless of why means nothing. Back in the day when Amd did win or make a competitive design they were spending more money and they were closer to Intel's budget. Not only that but you can't base things on history and try and claim it can repeat itself without evidence even more so with technology its illogical to do so.

Intel spends more money but they're also in more segments and a significant portion of that is on the fabs. If AMD can focus their R&D wisely they can still compete in some areas. ARM spends 1/3 of what AMD does on R&D and that isn't holding them back at all.

 

[truegenius]

intel already makes cpu's running android ? But it is not just tablets, laptops and every other ulv cpu has been bleeding money. Not to hard to receive design wins when you just about give away your product for free. forget about android it will stay in mobile until Microsoft releases a arm version of windows, if ever. But i see no need going forward both intel and amd have driven tdp down to a acceptable point to offer full windows comparability on mobile devices. Given a choice most users will take windows over android.

I would take Android over Metro.

me too
metro on my desktop makes me vomit and feel retard

i will say that a window should have glass like vista or 7 instead of cheap plastic looking thing of 8 onwards
i tried w8 and after few minutes i ran it in window basic theme to reduce its ugliness it looks that ugly

i can switch to w8 if i can find some modification software which will make it look exactly like vista or 7 ( just like softwares which used to make xp look like vista)
but then gta5 dosen't need dx12 so why even bother

 

[cdrkf]

Intel can spend as much R&D as it wants but the gains you get aren't as good as they used be able to get just by throwing money at it. Intel spent 5x anyone did on process tech and is still doing it while they are still losing their lead. Its just not going to be easy for intel to keep its lead through R&D alone in the long run. They are slowing down big time and sooner or later, the silicon will just become too volatile that even with intel's spending, they won't get more than a couple percent out of it than the competition who spend a small fraction of what intel spends.

Agreed, I think we would have to separate what money Intel is spending on Fabs and what money it is spending on CPU development to make a fair comparison.

As we've seen AMD are cutting product lines left and right (things dissapearing from the road maps) and are focusing on very specific thigns. Their stratergy to make their various IP blocks modular is also sound as once that modular framework is fully in place it should *hoepfully* last them a few generations before it becomes a bottleneck, allowing them a great deal of flexebility and also allowing them to further focus their limited R & D spend on the areas that need it most (i.e. improving the blocks- all this work on HSA is about creating the framework for said blocks and whatever you think of it, it's now complete).

I'm not suggesting this means AMD will automatically take the performance crown immediately, however I do think the 'AMD is doomed' mentality around here is premature.

1: I think Carrizo will look good in it's power envelope, +1 to hoping they get into some decent products (the business class Kaveri parts are in some nice ultra thins btw in the HP Elitebook range so that's something ).

2: I think K12 and Zen are likely to be considerably more competitive with Intel than the current position.

 

[juanrga]

The thing is the decisions taken to make the chip competitive at the 15 - 35w power range also mean it won't scale up to higher power designs. So they'd either have to make a second die on a different process node, or not release one (and it looks like the latter is the case). Whilst that's disappointing for desktop users, I'd take a decent AMD quad over a dual core i3 with HT any day if the performance is a bit better *and it has better GPU thrown into the deal*. I actually think that is possible. Kaveri looked much better at low clocks than Trinity / Richland for example (but doesn't higher up).

AMD doesn't have the money to do a second die on a different process. In fact it seems that no Kaveri refresh (a la Richland) is coming. It seems that AMD will release some new Kaveri models, but without anything really new.

 

[juanrga]

look at it this way, intel is 95x bigger than amd but there r&d budget is only 10x more, amd allocates more money as a percentage of size towards r&d than intel. is the glass half empty or half full ?

It is irrelevant if Intel spends all its money on R&D or only a part. It is irrelevant if Intel spend the rest of the money on ONGs or on their marketing dept. or what. AMD has been very competitive when the R&D budgets were close and when Intel did mistakes (Itanium, P4). Intel is not doing such mistakes with Broadwell or Skylake.

My point is that there is no hard data to believe that AMD will do a comeback. Of course anyone is free to believe the contrary. I recall people believing that Bulldozer would crush Intel and then believing that Piledriver was coming to fix evereryiting, and then believing that Steamroller was going to crush Haswell, and latter hoping that Excavator would be the real game changer.

Now we know that Excavator could even be slower than Steamroller, some people believes that Zen will be the game changer and when Zen was ready and disappointing then they will expect for post-Zen...

This thread always goes in circles. We are back to two years ago, when I said that Piledriver was the last FX CPU and that SOI was dead, and people who didn't trust me was expecting for a Steamroller 8-core FX CPUs and claimed that Kaveri was made on SOI.

Even after the truth was revealed and my claims confirmed people still didn't believe them. A group of people signed a public petition for AMD releasing FX Steamroller CPUs (the petition was nowhere), whereas I recall one of that people who saw the acronym "SHP" on one of Kaveri slides presentation and claimed that Kaveri was SOI.

I will not insist more in this issue. I will leave history to run and reveal the future as it is, not as some want it to be.

 

[juanrga]

Intel can spend as much R&D as it wants but the gains you get aren't as good as they used be able to get just by throwing money at it. Intel spent 5x anyone did on process tech and is still doing it while they are still losing their lead. Its just not going to be easy for intel to keep its lead through R&D alone in the long run. They are slowing down big time and sooner or later, the silicon will just become too volatile that even with intel's spending, they won't get more than a couple percent out of it than the competition who spend a small fraction of what intel spends.

X86 is near its limits. That is why Intel is developing new ISAs.

 

[juanrga]

Voltage is a function of frequency. As explained in my article about power consumption of CPUs, substituting the (very good) approximation V ~ f on the expression for the dynamic power gives the well-known cubic dependence that is quoted in many sites

https://software.intel.com/en-us/articles/the-new-era-of-tera-scale-computing

A concrete case. First the almost linear dependence of voltage with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/VoltageCurve.png

and then the almost cubic dependence of (dynamic) power with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/PtotalVccTGHzDataGraph.png

I see your rooky mistake now. You took the dumbed down example from Intel where they normalized voltage and frequency to 1, and assumed that is valid for real processors as well. It was just an example so the layman can follow along.

No. I gave you the link as simple example of the cubic law being quoted. The cubic law appears in books and papers dealing with CPU design, which evidently you did never read. I also give you the figures with data measured for a real processor: a SB i7. Of course the cubic law also applies to AMD designs. It is the reason why a FX-CPU @ 4GHz is rated at 125W, whereas the same CPU @ 4.7GHZ is rated at 220W.

Intel spends more money but they're also in more segments and a significant portion of that is on the fabs. If AMD can focus their R&D wisely they can still compete in some areas. ARM spends 1/3 of what AMD does on R&D and that isn't holding them back at all.

Apples to oranges. We are comparing budgets for x86 (Zen) vs x86 (Skylake). Thanks to the ISA, it is cheaper to develop an ARM CPU than a x86 CPU.

Andrew Feldman, former server head at AMD, mentioned that one could build an entirely custom chip using the ARM architecture in about 18 months for about $30 million in development costs. But it requires three or four-years and $300 million to $400 million to build an x86-based one.

That is why several companies are developing ARM server chips with same performance or even more than the best x86 designs.

 

[cdrkf]

Voltage is a function of frequency. As explained in my article about power consumption of CPUs, substituting the (very good) approximation V ~ f on the expression for the dynamic power gives the well-known cubic dependence that is quoted in many sites

https://software.intel.com/en-us/articles/the-new-era-of-tera-scale-computing

A concrete case. First the almost linear dependence of voltage with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/VoltageCurve.png

and then the almost cubic dependence of (dynamic) power with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/PtotalVccTGHzDataGraph.png

I see your rooky mistake now. You took the dumbed down example from Intel where they normalized voltage and frequency to 1, and assumed that is valid for real processors as well. It was just an example so the layman can follow along.

No. I gave you the link as simple example of the cubic law being quoted. The cubic law appears in books and papers dealing with CPU design, which evidently you did never read. I also give you the figures with data measured for a real processor: a SB i7. Of course the cubic law also applies to AMD designs. It is the reason why a FX-CPU @ 4GHz is rated at 125W, whereas the same CPU @ 4.7GHZ is rated at 220W.

Intel spends more money but they're also in more segments and a significant portion of that is on the fabs. If AMD can focus their R&D wisely they can still compete in some areas. ARM spends 1/3 of what AMD does on R&D and that isn't holding them back at all.

Apples to oranges. We are comparing budgets for x86 (Zen) vs x86 (Skylake). Thanks to the ISA, it is cheaper to develop an ARM CPU than a x86 CPU.

Andrew Feldman, former server head at AMD, mentioned that one could build an entirely custom chip using the ARM architecture in about 18 months for about $30 million in development costs. But it requires three or four-years and $300 million to $400 million to build an x86-based one.

That is why several companies are developing ARM server chips with same performance or even more than the best x86 designs.

That would suggest then, that despite AMD's relative budget shortfall, they have the money (and I'd argue expertise) to develop a good custom ARM core in K12.

 

[logainofhades]

how about tequila ? metro is a overlay and can be easily changed back to standard windows view, try win 10 i think you will like it.

Downloaded the technical preview. Gonna give it a try, as soon as I can find me a spare HDD, that I do not mind wiping.

 

[szatkus]

New benchmarks: http://www.sisoftware.eu/rank2011d/show_system.php?q=cea598aa98af9ba391b7d0edc0f1d7a598a88ee7daebcda598ad8bf3ceffd9bcd9e4d4f281bc84&l=en

This time 2.5GHz/3.5GHz

 

[szatkus]

http://semiaccurate.com/2013/05/15/amds-andrew-feldman-talks-about-arm/

By Mr. Feldman’s count it takes 400 million dollars and three years to build an x86 server chip. But only 30 Million dollars and 18 months are required to make an ARM chip. As far as AMD is concerned the ARM ecosystem offers massive cost and time-to-market advantages over building traditional x86 chips.

Entirely custom? I doubt it's possible to build anything custom for $30m. He clearly meant cores from ARM (like in Seatle).

 

[Cazalan]

No. I gave you the link as simple example of the cubic law being quoted. The cubic law appears in books and papers dealing with CPU design, which evidently you did never read. I also give you the figures with data measured for a real processor: a SB i7. Of course the cubic law also applies to AMD designs. It is the reason why a FX-CPU @ 4GHz is rated at 125W, whereas the same CPU @ 4.7GHZ is rated at 220W.

It's cubic with respect to V^2 * F. As the article and pictures you linked showed. You're the only person I've ever seen try to simplify the equation down even further to F^3.

 

[gamerk316]

Intel can spend as much R&D as it wants but the gains you get aren't as good as they used be able to get just by throwing money at it. Intel spent 5x anyone did on process tech and is still doing it while they are still losing their lead. Its just not going to be easy for intel to keep its lead through R&D alone in the long run. They are slowing down big time and sooner or later, the silicon will just become too volatile that even with intel's spending, they won't get more than a couple percent out of it than the competition who spend a small fraction of what intel spends.

X86 is near its limits. That is why Intel is developing new ISAs.

X86 is the longest lived stopgap solution in history.

 

[esrever]

New benchmarks: http://www.sisoftware.eu/rank2011d/show_system.php?q=cea598aa98af9ba391b7d0edc0f1d7a598a88ee7daebcda598ad8bf3ceffd9bcd9e4d4f281bc84&l=en

This time 2.5GHz/3.5GHz

CPU performance looks pretty weak.

 

[juanrga]

Intel can spend as much R&D as it wants but the gains you get aren't as good as they used be able to get just by throwing money at it. Intel spent 5x anyone did on process tech and is still doing it while they are still losing their lead. Its just not going to be easy for intel to keep its lead through R&D alone in the long run. They are slowing down big time and sooner or later, the silicon will just become too volatile that even with intel's spending, they won't get more than a couple percent out of it than the competition who spend a small fraction of what intel spends.

X86 is near its limits. That is why Intel is developing new ISAs.

X86 is the longest lived stopgap solution in history.

I was talking about performance.

 

[juanrga]

Apples to oranges. We are comparing budgets for x86 (Zen) vs x86 (Skylake). Thanks to the ISA, it is cheaper to develop an ARM CPU than a x86 CPU.

Andrew Feldman, former server head at AMD, mentioned that one could build an entirely custom chip using the ARM architecture in about 18 months for about $30 million in development costs. But it requires three or four-years and $300 million to $400 million to build an x86-based one.

That is why several companies are developing ARM server chips with same performance or even more than the best x86 designs.

That would suggest then, that despite AMD's relative budget shortfall, they have the money (and I'd argue expertise) to develop a good custom ARM core in K12.

About K12. I agree that R&D budget constraints are not so relevant for K12. But I don't think they have so much expertise. Keller mentioned how his team of engineers believed had found a mistake in the ARM ISA but in the end it was an mistake in AMD tools. This anecdote doesn't put the team in good shape.

A problem for K12 is that it is constrained by Skybridge. As Kanter said me the designs are very close and he expects K12 to be 10% better at best.

We can take a look at competition. We know lots of details about other's chips, but AMD remains mysteriously silent and that is very suspicious. We know that Cray chose Cavium instead AMD. We know Feldman abandoned AMD. Moreover, AMD vicepresident claiming that K12 won a single cheap design makes me think that the K12 will not be anything special.

Of course, we have to wait and see but all signs point to K12 being more a meh! than a WOW!

 

[juanrga]

New benchmarks: http://www.sisoftware.eu/rank2011d/show_system.php?q=cea598aa98af9ba391b7d0edc0f1d7a598a88ee7daebcda598ad8bf3ceffd9bcd9e4d4f281bc84&l=en

This time 2.5GHz/3.5GHz

Turbo is 3.4GHz.

 

[szatkus]

Typo, I can't edit posts here, don't know why :/

 

[szatkus]

[
About K12. I agree that R&D budget constraints are not so relevant for K12. But I don't think they have so much expertise. Keller mentioned how his team of engineers believed had found a mistake in the ARM ISA but in the end it was an mistake in AMD tools. This anecdote doesn't put the team in good shape.

ARM's design, not ARM ISA...

I can tell you few similar stories about my team. It doesn't mean that team is incompetent or something. Everyone makes mistakes, it's normal during development.
Looks like it wasn't a good idea from PR standpoint and non-technical people like you now think that holy Keller leads bunch of losers.

 

[juanrga]

[
About K12. I agree that R&D budget constraints are not so relevant for K12. But I don't think they have so much expertise. Keller mentioned how his team of engineers believed had found a mistake in the ARM ISA but in the end it was an mistake in AMD tools. This anecdote doesn't put the team in good shape.

ARM's design, not ARM ISA...

I had to refresh my mind and found the original Keller's claim

Our guys were so excited when they thought they’d found a bug in the ARM architecture using our tools [...] Turned out, it was a bug in our tools.

ISA and microarchitecture are subsets of architecture. The quote is not precise enough to know what exact bug they believed had found.

 

[juanrga]

New benchmarks: http://www.sisoftware.eu/rank2011d/show_system.php?q=cea598aa98af9ba391b7d0edc0f1d7a598a88ee7daebcda598ad8bf3ceffd9bcd9e4d4f281bc84&l=en

This time 2.5GHz/3.5GHz

CPU performance looks pretty weak.

Carrizo is considered a "GPU-centric design".

 

[de5_Roy]

The MintBox Mini is a silent, quad-core Linux Mint PC that fits in your pocket
http://www.pcworld.com/article/2871328/the-mintbox-mini-is-a-silent-quad-core-linux-mint-pc-that-fits-in-your-pocket.html

crapload of wccfudtech rumors

AMD Allegedly Preparing 28nm Trinidad GPU To Replace Curacao Pro – Arrives In March, Features 2 GB VRAM and 256-bit Bus
http://wccftech.com/amd-allegedly-preparing-28nm-trinidad-gpu-replace-curacao-pro-arrives-march-features-2-gb-vram-256bit-bus/
AMD Carrizo Benchmarks Leaked – Double The Performance of Kaveri
http://wccftech.com/amd-carrizo-graphics-benchmarks-leaked-double-performance-kaveri/

my past posts have gotten lost during submission. fingers crossed for this one....

 

[8350rocks]

2) To compete with 14nm Broadwell on efficiency they have to reduce clocks. Whereas performance increases with frequency is about linear, power increases as cube frequency: P ~ f^3

Where do you come up with this pseudo math?

P = C*F*V^2

http://en.wikipedia.org/wiki/Frequency_scaling

Power remains linear with frequency. The reason the power starts going crazy is they have to bump the core voltage to remain stable.

Voltage is a function of frequency. As explained in my article about power consumption of CPUs, substituting the (very good) approximation V ~ f on the expression for the dynamic power gives the well-known cubic dependence that is quoted in many sites

https://software.intel.com/en-us/articles/the-new-era-of-tera-scale-computing

A concrete case. First the almost linear dependence of voltage with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/VoltageCurve.png

and then the almost cubic dependence of (dynamic) power with frequency

http://i272.photobucket.com/albums/jj163/idontcare_photo_bucket/Intel%20Core%20i7-2600K/PtotalVccTGHzDataGraph.png

Cazalan 1 > juanrga 0

Real math > made up math