AMD CPU speculation... and expert conjecture

[juanrga]

There is no refresh with higher clocks. Don't expect a 4.1GHz 'Kaveri'. Don't expect a new 'Kaveri' with full HSA. Don't expect a new 'Kaveri' with HBM. The only 'refresh' will be firmware changes, different steeping, and minor changes like that. A more mature process could reduce power consumption a bit.

There will be new models above Carrizo but there is absolutely nothing really new under the hood.

First you say there will be no Kaveri refresh. Then when a refresh is confirmed you change to "there will be no higher clocks". Now higher frequency updates appear to be planned.

Sometime in the second quarter of the year AMD is also expected to release new APUs for desktops, which are currently known as the “Kaveri Refresh”. The new chips will operate at higher clock-rates than the currently available “Kaveri” and will increase performance of AMD’s APUs a bit.
-http://www.kitguru.net/components/graphic-cards/anton-shilov/amd-we-will-release-new-products-starting-in-the-second-quarter-of-2015/

I change my posts so fast as AMD change the plans.

Initially AMD planned Carrizo on FM2+ as replacement for Kaveri. I posted the old roadmap showing up to 65W Carrizos. That was canceled.

Then AMD planned a Carrizo coming to desktop only in BGA format and up to 35W, with rest of desktop line being covered by a "Kaveri refresh" consisting of higher clocked Kaveris. I have been said that was was canceled as well.

My last info is that "Kaveri refresh" now consists of new models with essentially nothing new beyond minor changes such as firmware fixes. I am said that there is no higher clocks.

I see that Anton Silov claims otherwise. Cannot we just wait and see who is right, Anton or my source?

 

[8350rocks]

What Anton says mirrors what I am hearing as well...but what do engineers know?

 

[truegenius]

This is my last word about this issue.

I was the only one that provided real data for a concrete processor and the empirical fits. My point was to use the well-known cubic law to explain why AMD is reducing base clocks on Carrizo to compete with Broadwell.

You can avoid the cubic law approximation if you don't like it and you can wait to Carrizo to hit the market, then measure final silicon to get the empirical relations between V and F and between P and F.

Instead the well-known generic cubic law proportionality, you will get a more complex power series expression with lots of empirical parameters valid only for a concrete Carrizo processor (I have said "processor", not model), a concrete ambient temperature..., but the conclusion will be the same. Both the more complex expression and the simple cubic law approximation predict that efficiency improves by reducing clocks. This is why we don't see phones clocked at 4GHz, for instance, or why the more efficient Intel Xeons are clocked so low as 1.8GHz.

I have tried to explain why AMD does what it does with Carrizo. I have tried to explain why FX-8350 @4GHz is rated at 125W, but FX-9590 @4.7GHz is rated at 220W.

Any time I have tried to explain why AMD does what it does (since the epoch when I tried to explain why AMD chose bulk instead SOI) my points have been either ignored or blatantly misinterpreted.

Intel has a nice blog entry explaining the relation between power and frequency to broad audiences

https://software.intel.com/en-us/blogs/2014/02/19/why-has-cpu-frequency-ceased-to-grow

They also use the well-known V~F approx to derive the cubic law except that they write it as P ~ V^3 but evidently it is equivalent to P ~ F^3, and then write

Linear frequency growth causes power dissipation to be increasingly cubed! If the frequency is raised only twice, there will be eight times greater heat that must be accommodated or the processor will melt or shutdown.

It’s obvious that this method of increasing the frequency is not suitable for processor manufacturers because of low efficiency. However, it is used by extreme overclockers.

As said, this is my last word about this issue. People can feel free to misread me misinterpret me or just ignore my point or pretend again that the cubic law is my invention or pretend that I wrote "V=F", when I never did. I don't care. I suspect which will be the next misunderstanding of my point (hint: "without voltage changing"). I will not even care to reply that. What is really interesting is that I wrote virtually the same thoughts about why AMD is reducing frequencies (cubic law), in another forum, and my thoughts were well-received by the people therein, including the ones with internal data about Carrizo

http://semiaccurate.com/forums/showpost.php?p=227588&postcount=182
http://semiaccurate.com/forums/showpost.php?p=227607&postcount=184

As both mention my only mistake about Carrizo was in the "pipeline depth". I believed AMD would reduce the pipeline in Excavator core but it is the same inefficient pipeline that Steamroller. I hope that Zen includes a shorther pipeline.

i7-3770k same speed varying voltage

oh look man, i can't believe it, power consumption is approximately same throughout the voltage range
juan was right about p~v^3

 

[truegenius]

oh look man, they are approaching 2x frequency
it is going to melt
where is my fire extinguisher, where is my fire extinguisher.......

 

[noob2222]

Pardon my stupidity but isnt x^3 a parabolic and not a straight line? Its been soo long since i had jr high algebra that i simply dont remember.

Can anyone help me.

 

[Embra]

A little something:
MSI Adds All-in-One PCs based on AMD Beema A4-6210 Quad Core APU
http://www.guru3d.com/news-story/msi-adds-all-in-one-pcs-based-on-amd-beema-a4-6210-quad-core-apu.html

 

[Embra]

New AMD GPUs and APUs after March
http://www.guru3d.com/news-story/new-amd-gpus-and-apus-after-march.html

 

[juanrga]

AMD is already designing products on the 14nm node

http://hexus.net/tech/news/cpu/79758-amd-already-designing-products-14nm-node/

FinFET and on Samsung/Glofo node, as I said.

No FD-SOI, no TSMC...

 

[etayorius]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

 

[szatkus]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

Probably they are finishing Zen and need for money to complete additional designs. I mean they haven't released a new enthusiasts and server chips for few years.

 

[-Fran-]

It was about time they thought about taking some of the pie from Intel in the high end server markets.

Cheers!

 

[Reepca]

Hopefully this focus on the server market will include some HSA servers. With Carrizo they're supposed to reach HSA 1.0 compliance, it'd be great to have that in some servers where efficiency and data-parallelism can be very important.

 

[juanrga]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

 

[jdwii]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

Probably they are finishing Zen and need for money to complete additional designs. I mean they haven't released a new enthusiasts and server chips for few years.

This makes me very happy.

TO the point i feel like saying YES OMG like a little school girl. I'll say this like Nasa Amd can spend their money well when they get very little!

Although this wasn't always the case

 

[jdwii]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

 

[Reepca]

Pardon my stupidity but isnt x^3 a parabolic and not a straight line? Its been soo long since i had jr high algebra that i simply dont remember.

Can anyone help me.

Functions with a degree of 2 are parabolic (that is, the highest power x is raised to in any term is 2).

y = x^3 isn't parabolic or a straight line - it increases from the left, intersects the y-axis at x=0, then increases as x increases. You could get a better look at it in wolframalpha or some similar math-centred site.

 

[gamerk316]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

*In specific workloads.

 

[juanrga]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

Yes, they are the same site that leaked the Baeca nonsense, and the inexistent 16-core Steamroller Opteron, and are the same site that claimed recently that Carrizo will be 2x faster than Kaveri and 3x faster than Iris Pro: nonsense^3

I didn't comment if what Lisa Su said is god or bad but I will share my beliefs.

AMD is divided into two groups: the enterprise, embedded, and semi-custom group and the computing and graphics group. Lisa Su said explicitly that will increase the R&D of the first group, but she didn't even mentioned the R&D for the second group (aka the group leaded by John Byrne before his departure). I only see two possibilities:

(i) The R&D of the the computing and graphics group remains frozen at its current low levels or
(ii) the increase in the R&D of the enterprise, embedded, and semi-custom group is made at cost of reducing still more the R&D on the computing and graphics group.

What interests me more are the parts that WCCFTECH didn't quote. Lisa Su said that part of the current problems of AMD are due to its "over-concentration in consumer PCs" and then mentioned that her long-term strategy will consist on reducing that concentration, including reduction of the "downstream desktop and AIB channel inventory"

She said "The computing and graphics segment decreased 15% sequentially, largely driven by weak channel sales" and then praised the embedded processor revenues.

It all together looks as AMD is not planning a comeback on the "consumer PC", but turning into an enterprise, embedded, and semi-custom company.

I think the K12 and Zen cores are being designed for their future enterprise, embedded, and semi-custom business and their appearance in some PCs will be a secondary effect. In fact, what I hear is that Zen/K12 are small and efficient cores (Jaguar style) instead big and powerful. Zen and K12 will be suitable for dense servers, kiosks, game consoles, network storage, avionics, and medical devices...

 

[logainofhades]

Typically, AMD's server CPU tech ends up in the desktop space. While they will not particularly focus on consumer PC, it isn't exactly going anywhere, either.

 

[juanrga]

There is no refresh with higher clocks. Don't expect a 4.1GHz 'Kaveri'. Don't expect a new 'Kaveri' with full HSA. Don't expect a new 'Kaveri' with HBM. The only 'refresh' will be firmware changes, different steeping, and minor changes like that. A more mature process could reduce power consumption a bit.

There will be new models above Carrizo but there is absolutely nothing really new under the hood.

<snip>

I change my posts so fast as AMD change the plans.

Initially AMD planned Carrizo on FM2+ as replacement for Kaveri. I posted the old roadmap showing up to 65W Carrizos. That was canceled.

Then AMD planned a Carrizo coming to desktop only in BGA format and up to 35W, with rest of desktop line being covered by a "Kaveri refresh" consisting of higher clocked Kaveris. I have been said that was was canceled as well.

My last info is that "Kaveri refresh" now consists of new models with essentially nothing new beyond minor changes such as firmware fixes. I am said that there is no higher clocks.

I see that Anton Silov claims otherwise. Cannot we just wait and see who is right, Anton or my source?

What Anton says mirrors what I am hearing as well...but what do engineers know?

We don't have to wait. I asked for clarification. I was half wrong and half right. I copy and paste:

The clocks will not raise beyond the current high-end levels (7850K / 860K) so essentially only some of the lower-end models are affected. The "upgrade" makes the gap between the high-end and main stream models little smaller.

Thus some basic Kaveri models will receive a refresh "clocked slightly higher" but the current flagship model, the A10-7850K will remain as the higher-clocked Kaveri ever. As I wrote before Don't expect a 4.1GHz 'Kaveri'. And I can confirm my previous claims that Kaveri refresh doesn't include new HSA features, nor HBM, nor anything: "there are no physical or technical changes in them."

 

[8350rocks]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Think you are glossing over a key part of that R&D budget. While, in this case, wccftech is making some inferences from the announcement...they are not unjustified in doing so.

 

[cdrkf]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

Phenom I was a 'superior' design to Intel's quad cores of the time from an engineering perspective, given that Intel didn't have a proper quad core, rather 2 dual cores on one package.

Unfortunately for AMD, the Phenom I shipped with a defect seriously hampering it's performance, it didn't have enough cache due to transistor budgets, it didn't clock as high as it should have and there was no properly threaded software at the time to expose the benefit of a monolithic quad core over the technically inferior solution Intel had slapped together to beat them to the punch.

Intel pretty much validated AMD's design of Phenom when they neigh on copied it with the first Core processor. Also once AMD fixed the issues with Phenom I and re-released it they had a superb chip in Phenom II. It was definitely competitive with anything Intel had at the time, and the X6 did quite well against first gen i7. It was only when Sandy came out and then Bulldozer fell flat that things really went pair shaped for AMD.

 

[jdwii]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

Phenom I was a 'superior' design to Intel's quad cores of the time from an engineering perspective, given that Intel didn't have a proper quad core, rather 2 dual cores on one package.

Unfortunately for AMD, the Phenom I shipped with a defect seriously hampering it's performance, it didn't have enough cache due to transistor budgets, it didn't clock as high as it should have and there was no properly threaded software at the time to expose the benefit of a monolithic quad core over the technically inferior solution Intel had slapped together to beat them to the punch.

Intel pretty much validated AMD's design of Phenom when they neigh on copied it with the first Core processor. Also once AMD fixed the issues with Phenom I and re-released it they had a superb chip in Phenom II. It was definitely competitive with anything Intel had at the time, and the X6 did quite well against first gen i7. It was only when Sandy came out and then Bulldozer fell flat that things really went pair shaped for AMD.

I'm pretty sure the first Phenom cores were also the time when they feel behind performance per clock. But lets hold on here the Phenom II was great for the money but architecture wise Amd was late and the Phenom II was meant to compete with the quad cores Intel had not the I3-I5-I7 series. Its been kinda like this now for years now we can only wish and hope.

Now we have Intel locking everything down to save them money on chips that are already smaller and cheaper to make then Amds products.

 

[jdwii]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

*In specific workloads.

It's like when a car manufacturer claims they can go from 0-60 in 4.6 seconds down a hill with fast wind. Or how Audio manufacturers claim 100 watts per channel at 1 khz.

 

[gamerk316]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Either way juan it still means a bit more money for for those said things i noticed the same thing you did and i was like well its from some troll site that claimed bulldozer would be 50% faster and the same nonsense about the first gen phenom being like superior to Intel's quad cores.

Phenom I was a 'superior' design to Intel's quad cores of the time from an engineering perspective, given that Intel didn't have a proper quad core, rather 2 dual cores on one package.

Unfortunately for AMD, the Phenom I shipped with a defect seriously hampering it's performance, it didn't have enough cache due to transistor budgets, it didn't clock as high as it should have and there was no properly threaded software at the time to expose the benefit of a monolithic quad core over the technically inferior solution Intel had slapped together to beat them to the punch.

Intel pretty much validated AMD's design of Phenom when they neigh on copied it with the first Core processor. Also once AMD fixed the issues with Phenom I and re-released it they had a superb chip in Phenom II. It was definitely competitive with anything Intel had at the time, and the X6 did quite well against first gen i7. It was only when Sandy came out and then Bulldozer fell flat that things really went pair shaped for AMD.

Anyone remember the arguments that Phenom I was better then C2Q because "it was a REAL Quad Core"? Those were the days.