AMD CPU speculation... and expert conjecture

[juanrga]

Juan...You are comparing the maxed out A10-7850k to what the 3rd highest clocked part?

Wrong, FLAGSHIP parts are the FX 9590 @ 4.7 GHz base clock/5.0 GHz Turbo Core 3.0.

Now, your part overclocked does not hit the stock clocks of the old flagship part.

Show me a production Kaveri @ 4.5 GHz and we can have your discussion...otherwise...your crystal ball is cracked.

It seems you don't understand that I compared average frequencies. Some few (golden-chips) FX-8350 surpass the 4750MHz, are labeled as FX-9000 series, and sold at higher base frequency by AMD. You can also find some few Kaveri A10 that surpass the 4775MHz. That the latter are not labeled special neither sold at higher frequencies doesn't mean really anything.

I understand your difficulty to accept that I predicted 4.5GHz and that Kaveri hits 4.5GHz. I understand your frustration because my broken crystal ball did work better than the "facts" from your 'engineers' at AMD. But you would stop spreading FUD against Kaveri frequencies.

In fact you are the first person that I meet in a forum that ask me to "show" a "production Kaveri @ 4.5 GHz" before admitting that the chip can hit that frequency. I suppose that you will also solicit me to "show" a "production i5-4670k @ 4.5 GHz" before admitting that the Haswell i5 can hit 4.5GHz without FD-SOI.

 

[blackkstar]

@gamerK, yes you can run the same benchmarks on Windows and on Linux. And yes there's not always an improvement. But sometimes there is. And it can be massive. Blender is a huge one. In fact, even just compiling it in Windows with MinGW and GCC makes it significantly faster. Windows has a lot of cruft slowing it down, but that doesn't mean it's always going to make every task slower. But it does make it easier for outliers to stand out which make it convenient for certain people to cherry pick things and push an agenda.

And as far as server needs go and computing needs go, at the time they were extremely different. Gaming CPUs still basically needed single core and you were lucky if you got two big threads in a game. Servers scale to many cores extremely easily. DBMS, web server, etc are all designed to use tons and tons of cores. Look at Power and SPARC designs, they don't care about single thread performance first, they care about a chip with a boat load of threads.

Times are changing for multi-threading, because as some of you have mentioned, we're reaching the end of silicon shrinking and we can't seem to do much better than 2600k single thread performance and we instead move the goal posts for new products from performance to power per watt or some other statistic that's meaningless for people who use a single computer in a place that doesn't have to pay a ton of money for electricity.

I don't know how you are going to argue against Bulldozer not being designed as a server chip.

By having the shared architecture, reducing the size and sharing things that aren’t commonly used in their peak capacity in server workloads, “Bulldozer” is actually very well aligned with server workloads now and on into the future. In fact, a great deal of the trade-offs in Bulldozer were made on behalf of servers, and not just one type of workload, but a diversity of workloads.

Chuck Moore, Chief Architect AMD (http://www.anandtech.com/show/5058/amds-opteron-interlagos-6200)

Your version of what bulldozer is designed to do is more accurate than the Chief Architect of AMD's was at the release of Bulldozer? I'm impressed! You know what AMD is designing chips for better than AMD does!

AMD is done with big cores in servers. They tried with Bulldozer to go server first and consumer second and it was a massive disaster. Now we are seeing Jaguar and ARM show up primarily in servers. And it's no surprise, servers love many weak cores. And ARM and Jaguar are absolutely perfect for a situation where you need many weak cores. But I don't think you will ever grasp that. You will forever think that because the games you work on don't scale to multiple cores that the only other tasks that scale to multiple cores are transcoding, rendering, and compiling.

 

[jdwii]

Juan, I said it was "comparatively low" for an early ES.

Amazingly you manage to twist words time and again. As I recall, at that time you said it was not an issue and clockspeed should be 4.0 GHz as your "predictions" stated.

Now, as I recall, when Kaveri was revealed, someone had to "fix" his predictions to accomodate for the loss in clockspeed...correct?

Now maybe you should go fix your crystal ball.

You said that the ES at 1.8GHz did mean that final Kaveri couldn't pass the 3GHz. I said you that Kaveri could hit 4.5Ghz and guess what the average frequency on air is of 4775MHz

4775MHz = 4.775GHz = 4.5GHz.

just as predicted by my broken crystal ball.

As explained before Kaveri CPU is clocked lower because has less TDP and more GPU on die. Your FUD about clocks was so accurate as your "facts" about FX-Steamroller coming to AM3+, Kaveri FD-SOI 28nm, AM4 on store, extra ALU per core, 30% single thread performance increase, Steamroller being faster than Haswell...

I find it funny he says higher OC factor although its only by 1%? While its clocked way lower than the other chip.

What is really funny is the double standard where lower base frequency is attributed to bulk, but higher OC factor is then dismissed.

You would not look at absolute differences, because the difference between 98% and 97%) and between 3% and 2% is of 1%, but in the last case you are adding a 50% more.

If percentages and relative values generate some problem try with frequencies:

The top Kaveri APU hits 4775MHz on air.
The FX-8350 CPU hits 4750MHz on air.

Yes I can see how Kaveri cannot hit 3Ghz. It is a complete disaster.

Do we really want to go into that? Please tell me if so

 

[jdwii]

Juan...You are comparing the maxed out A10-7850k to what the 3rd highest clocked part?

Wrong, FLAGSHIP parts are the FX 9590 @ 4.7 GHz base clock/5.0 GHz Turbo Core 3.0.

Now, your part overclocked does not hit the stock clocks of the old flagship part.

Show me a production Kaveri @ 4.5 GHz and we can have your discussion...otherwise...your crystal ball is cracked.

It seems you don't understand that I compared average frequencies. Some few (golden-chips) FX-8350 surpass the 4750MHz, are labeled as FX-9000 series, and sold at higher base frequency by AMD. You can also find some few Kaveri A10 that surpass the 4775MHz. That the latter are not labeled special neither sold at higher frequencies doesn't mean really anything.

I understand your difficulty to accept that I predicted 4.5GHz and that Kaveri hits 4.5GHz. I understand your frustration because my broken crystal ball did work better than the "facts" from your 'engineers' at AMD. But you would stop spreading FUD against Kaveri frequencies.

In fact you are the first person that I meet in a forum that ask me to "show" a "production Kaveri @ 4.5 GHz" before admitting that the chip can hit that frequency. I suppose that you will also solicit me to "show" a "production i5-4670k @ 4.5 GHz" before admitting that the Haswell i5 can hit 4.5GHz without FD-SOI.

At the end of the day price/performance is all what matters and in this case your weak little Kaveri losses hard real hard. So again i ask you do you really want to do this?

 

[jdwii]

juanrga i'm also waiting for those benchmarks that shows the I5 2500K losing or even meeting a A10 Kaveri on average instead of a few benchmarks that show that

 

[juanrga]

At the end of the day price/performance is all what matters and in this case your weak little Kaveri losses hard real hard. So again i ask you do you really want to do this?

I like the morphing of your pseudo-arguments The FUD started with Kaveri cannot hit 3Ghz due to bulk (your imagined friends at AMD said so). Then it changed to Kaveri is clocked at 3.7GHz due to bulk. Next it changed to show me a "production Kaveri @ 4.5GHz" or you will never accept that Kaveri can OC. Now it is changed to but but but but the price...

juanrga i'm also waiting for those benchmarks that shows the I5 2500K losing or even meeting a A10 Kaveri on average instead of a few benchmarks that show that

I see that you continue confused about this one.

 

[colinp]

Mobile Kaveri

http://fudzilla.com/home/item/34659-mobile-kaveri-tips-up-in-hp-laptop

This is where Kaveri should put in a better showing, after being a complete load of "meh" on the desktop.

 

[juanrga]

AMD is done with big cores in servers. They tried with Bulldozer to go server first and consumer second and it was a massive disaster. Now we are seeing Jaguar and ARM show up primarily in servers. And it's no surprise, servers love many weak cores. And ARM and Jaguar are absolutely perfect for a situation where you need many weak cores.

AMD server strategy includes big cores, as those in Berlin Opteron. Also jaguar servers have been replaced by ARM servers. AMD server strategy:

Warsaw CPU --> Piledriver cores --> Legacy customers
Berlin APU/CPU --> Steamroller/GCN cores --> HSA workloads
Seattle CPU --> A57 cores --> new paradigm

Bulldozer fiasco have little to see with being a big core design, but it is a consequence of its terrible inefficiency. Xeon CPUs are excellent server chips and use cores that are bigger than AMD.

 

[blackkstar]

AMD is done with big cores in servers. They tried with Bulldozer to go server first and consumer second and it was a massive disaster. Now we are seeing Jaguar and ARM show up primarily in servers. And it's no surprise, servers love many weak cores. And ARM and Jaguar are absolutely perfect for a situation where you need many weak cores.

AMD server strategy includes big cores, as those in Berlin Opteron. Also jaguar servers have been replaced by ARM servers. AMD server strategy:

Warsaw CPU --> Piledriver cores --> Legacy customers
Berlin APU/CPU --> Steamroller/GCN cores --> HSA workloads
Seattle CPU --> A57 cores --> new paradigm

Bulldozer fiasco have little to see with being a big core design, but it is a consequence of its terrible inefficiency. Xeon CPUs are excellent server chips and use cores that are bigger than AMD.

The point I was getting at is that Bulldozer was designed for lots of weak threaded workloads. Besides it being a poor implementation of that ideal, even if it was good AMD would have gone for more weak threads over a few strong ones. They wanted to rule servers and then they basically just ramrodded Bulldozer into consumer markets, where single thread performance has far more value to the end user.

The result was something that needed software tailored to the architecture to perform well instead of taking advantage of software that already exists. AMD has done what they can to sway developers in key markets that it cares about, which seems to be gaming. So to me it seems like AMD still cares about the gaming PC market and they do have plans further down the line.

But the thing that sticks out to me about the new Mobile Kaveri models is how they are referring to the chips as having a total number of compute units that includes GCN clusters and CPU cores.

Which brings me to my point. How is AMD going to market a straight up CPU against that? An 8m/16c dCPU wouldn't look so well against mobile Kaveri to a consumer.

It completely blurs the line between CPU and GPU core. So 8m/16c would end up with 16 "compute units" while FX 7600P (a 4c APU with 8 GCN clusters) gets branded as 12 "compute units"

So it seems either like a marketing disaster in the works or that AMD is going to abandon CPU only and if/when we do see a big CPU from AMD It's going to have GCN attached regardless.

I hate to admit it, but to me, AMD referrencing their APU models as having "compute core" counts that include GCN and CPU cores makes me really think there is no more dCPU coming. But it doesn't rule out HSA between multiple APUs or even a dGPU yet. But it seems a dCPU is very unlikely if AMD is going to start marketing their APUs as having "compute core" totals that include GCN clusters and traditional CPU cores.

 

[Cazalan]

Your link is completely irrelevant. And Skylake CGPU rumors are covered in many places:

The ONLY Intel link is irrelevant? LOL. While the speculation links are relevant. Got it!

Up is down, down is up!

 

[juanrga]

Another report that confirms the shrinking of the traditional PC market and that explains AMD/Intel strategies to move away from this market

http://www.digitimes.com/news/a20140505PR203.html

Below an interview to AMD CEO. It repeats again that the traditional PC market is declining, that he expects a decline of about 7--10% for this year and that AMD is accelerating its transformation to move production away from that declining market. He mentions again the objective of up to a 50% of revenues from outside the traditional PC market. This is the second person from AMD that openly admits that Bulldozer didn't work. Recognizing the mistake is a first step to avoid repeating it.

He confirms by first time in public that Keller and Koduri are working together in next gen server arch. I already said a pair of posts ago that Keller and Koduri are working in a new high-performance APU (Basilisk) that will replace the FX/Opteron series in 2016 (if everything goes as scheduled and there is no adjustments/delays). Keller is working in a new CPU architecture (post excavator) and Koduri is working in a new GCN architecture that complements it. The fact are working together is a sign of more HSA features

http://blogs.wsj.com/digits/2014/05/05/amd-ceo-beyond-the-unhealthy-duopoly-of-pc-chips/

Any bet that we will see the usual posters negating all what the AMD head says? Any bet that those posters will try to convince us about AMD releasing an inexistent 12-core FX CPU on a magic 20nm FD-SOI process because their imagined friend-working-at-AMD said so, but only to them?

 

[juanrga]

Review of A10-7700k

http://semiaccurate.com/2014/05/05/amds-a10-7700k-k/

 

[juanrga]

Your link is completely irrelevant. And Skylake CGPU rumors are covered in many places:

The ONLY Intel link is irrelevant? LOL. While the speculation links are relevant. Got it!

Up is down, down is up!

You don't get it. Your link was irrelevant because its content was irrelevant not because it is from Intel.

If instead deleting the links provided to you, you had read them, you would discover that some of them cite Intel sources. The Intel website mentions that the same architect designing the Phi cores is also designing the new GPU cores for Skylake and successors.

This quote from one of the links it pretty much summarize the situation:

So, over a year later, we were proven right about the lineage of Sky Lake’s GPU. Time to go buy a mole a new set of rubber booties, Intel waters their lawns enough to make his little toes muddy far too often. We told you that Larrabee as a GPU wasn’t dead, but would you listen? Noooo…….

 

[juanrga]

But the thing that sticks out to me about the new Mobile Kaveri models is how they are referring to the chips as having a total number of compute units that includes GCN clusters and CPU cores.

Which brings me to my point. How is AMD going to market a straight up CPU against that? An 8m/16c dCPU wouldn't look so well against mobile Kaveri to a consumer.

It completely blurs the line between CPU and GPU core. So 8m/16c would end up with 16 "compute units" while FX 7600P (a 4c APU with 8 GCN clusters) gets branded as 12 "compute units"

So it seems either like a marketing disaster in the works or that AMD is going to abandon CPU only and if/when we do see a big CPU from AMD It's going to have GCN attached regardless.

I hate to admit it, but to me, AMD referrencing their APU models as having "compute core" counts that include GCN and CPU cores makes me really think there is no more dCPU coming. But it doesn't rule out HSA between multiple APUs or even a dGPU yet. But it seems a dCPU is very unlikely if AMD is going to start marketing their APUs as having "compute core" totals that include GCN clusters and traditional CPU cores.

When AMD released the FX-9000 series I was one of the first suggesting that it would imply no Steamroller FX series, because the notation rule 8100 (Bulldozer) --> 8300 (Piledriver) --> 8500 (Steamroller) was broken. Mr I-have-a-friend-at-AMD disagreed and claimed that AMD was going to release a FX Steamroller CPU 28nm FDSOI for the AM3+ platform.

As I mentioned before the laws of physics are the same for AMD, Intel, and Nvidia. A dCPU+dGPU doesn't scale up to exascale; as a consequence, AMD, Nvidia, and Intel are preparing APUs for replacing the slow dCPU+dGPU.

As I mentioned before 14nm gives space enough to merge 8 big cores and about 2048 GCN cores in the same die. Once you have that APU you don't need a separate 4/6/8 cores dCPU anymore and can kill the FX-4000/6000/8000/9000 series.

HSA between APUs or between APU and dGPU is very inefficient from a latency/power viewpoint. Any possible advantage from using extra cores is lost in the communication bottleneck. This is the same reason why intel want to merge Xeon cores and Phi cores in same die:

Xeon CPU + Phi(PCIe) --> Xeon CPU + Phi(socket) --> Xeon + Phi on same die.

 

[8350rocks]

Another report that confirms the shrinking of the traditional PC market and that explains AMD/Intel strategies to move away from this market

http://www.digitimes.com/news/a20140505PR203.html

Below an interview to AMD CEO. It repeats again that the traditional PC market is declining, that he expects a decline of about 7--10% for this year and that AMD is accelerating its transformation to move production away from that declining market. He mentions again the objective of up to a 50% of revenues from outside the traditional PC market. This is the second person from AMD that openly admits that Bulldozer didn't work. Recognizing the mistake is a first step to avoid repeating it.

He confirms by first time in public that Keller and Koduri are working together in next gen server arch. I already said a pair of posts ago that Keller and Koduri are working in a new high-performance APU (Basilisk) that will replace the FX/Opteron series in 2016 (if everything goes as scheduled and there is no adjustments/delays). Keller is working in a new CPU architecture (post excavator) and Koduri is working in a new GCN architecture that complements it. The fact are working together is a sign of more HSA features

http://blogs.wsj.com/digits/2014/05/05/amd-ceo-beyond-the-unhealthy-duopoly-of-pc-chips/

Any bet that we will see the usual posters negating all what the AMD head says? Any bet that those posters will try to convince us about AMD releasing an inexistent 12-core FX CPU on a magic 20nm FD-SOI process because their imagined friend-working-at-AMD said so, but only to them?

When I talk with them in June I will let you know what they say about your speculation

 

[8350rocks]

Another report that confirms the shrinking of the traditional PC market and that explains AMD/Intel strategies to move away from this market

http://www.digitimes.com/news/a20140505PR203.html

Below an interview to AMD CEO. It repeats again that the traditional PC market is declining, that he expects a decline of about 7--10% for this year and that AMD is accelerating its transformation to move production away from that declining market. He mentions again the objective of up to a 50% of revenues from outside the traditional PC market. This is the second person from AMD that openly admits that Bulldozer didn't work. Recognizing the mistake is a first step to avoid repeating it.

He confirms by first time in public that Keller and Koduri are working together in next gen server arch. I already said a pair of posts ago that Keller and Koduri are working in a new high-performance APU (Basilisk) that will replace the FX/Opteron series in 2016 (if everything goes as scheduled and there is no adjustments/delays). Keller is working in a new CPU architecture (post excavator) and Koduri is working in a new GCN architecture that complements it. The fact are working together is a sign of more HSA features

http://blogs.wsj.com/digits/2014/05/05/amd-ceo-beyond-the-unhealthy-duopoly-of-pc-chips/

Any bet that we will see the usual posters negating all what the AMD head says? Any bet that those posters will try to convince us about AMD releasing an inexistent 12-core FX CPU on a magic 20nm FD-SOI process because their imagined friend-working-at-AMD said so, but only to them?

Have you changed your tinfoil hat in a while?

Of course AMD is saying that the PC market is an unhealthy main product line with average consumer sales shrinking slightly faster than PC gamers are upgrading. The overall trend is downward. That does NOT mean that they are skipping out entirely, only that they are trying to achieve a foothold in other markets to better supplement their PC component business and relations with OEMs will allow them to get those products into other markets in competitive form factors.

But only your speculation links are accurate about things going on in the real world.

Perhaps you take a deep breath, stop playing defensive about everything and look at your argument? I know you come here because the guys at S|A quit listening and you think you can foist this on us, but really? Be objective. You only see the world in 2 shades...Your side of the coin is "clearly right", while the other side of the coin is "clearly wrong".

 

[de5_Roy]

looks like arm socs are coming to AM1 platform in the near future
http://www.tomshardware.com/news/amd-ambidextrous-x86-arm,26709.html
i wonder what this new "matching x86 design" will be.. certainly doesn't look like replacement, rather coexist.

AMD demos Seattle, its first ARM-based server chip
http://techreport.com/news/26419/amd-demos-seattle-its-first-arm-based-server-chip
AMD is working on K12, brand-new x86 and ARM cores
http://techreport.com/news/26417/amd-is-working-on-k12-brand-new-x86-and-arm-cores

moar promo slides
http://www.anandtech.com/show/7990/amd-announces-k12-core-custom-64bit-arm-design-in-2016
http://www.anandtech.com/show/7989/amd-announces-project-skybridge-pincompatible-arm-and-x86-socs-in-2015

 

[juanrga]

Another report that confirms the shrinking of the traditional PC market and that explains AMD/Intel strategies to move away from this market

http://www.digitimes.com/news/a20140505PR203.html

Below an interview to AMD CEO. It repeats again that the traditional PC market is declining, that he expects a decline of about 7--10% for this year and that AMD is accelerating its transformation to move production away from that declining market. He mentions again the objective of up to a 50% of revenues from outside the traditional PC market. This is the second person from AMD that openly admits that Bulldozer didn't work. Recognizing the mistake is a first step to avoid repeating it.

He confirms by first time in public that Keller and Koduri are working together in next gen server arch. I already said a pair of posts ago that Keller and Koduri are working in a new high-performance APU (Basilisk) that will replace the FX/Opteron series in 2016 (if everything goes as scheduled and there is no adjustments/delays). Keller is working in a new CPU architecture (post excavator) and Koduri is working in a new GCN architecture that complements it. The fact are working together is a sign of more HSA features

http://blogs.wsj.com/digits/2014/05/05/amd-ceo-beyond-the-unhealthy-duopoly-of-pc-chips/

Any bet that we will see the usual posters negating all what the AMD head says? Any bet that those posters will try to convince us about AMD releasing an inexistent 12-core FX CPU on a magic 20nm FD-SOI process because their imagined friend-working-at-AMD said so, but only to them?

Have you changed your tinfoil hat in a while?

Of course AMD is saying that the PC market is an unhealthy main product line with average consumer sales shrinking slightly faster than PC gamers are upgrading. The overall trend is downward. That does NOT mean that they are skipping out entirely,

Are you really unable to understand the difference between 50% and 0%?

 

[colinp]

looks like arm socs are coming to AM1 platform in the near future
http://www.tomshardware.com/news/amd-ambidextrous-x86-arm,26709.html
i wonder what this new "matching x86 design" will be.. certainly doesn't look like replacement, rather coexist.

AMD demos Seattle, its first ARM-based server chip
http://techreport.com/news/26419/amd-demos-seattle-its-first-arm-based-server-chip
AMD is working on K12, brand-new x86 and ARM cores
http://techreport.com/news/26417/amd-is-working-on-k12-brand-new-x86-and-arm-cores

moar promo slides
http://www.anandtech.com/show/7990/amd-announces-k12-core-custom-64bit-arm-design-in-2016
http://www.anandtech.com/show/7989/amd-announces-project-skybridge-pincompatible-arm-and-x86-socs-in-2015

Very interesting. So now we know what Jim Keller was hired for, and it wasn't HEDT.

Juanrga, 8350rocks, will you two just get a room somewhere?

 

[juanrga]

Keller is working in both x86 and ARM as predicted.

x86 new architecture for 2016 (aka post-Carrizo) as predicted.

New ARM custom core called "K12 core" (there are lots of irony about this name, you know why).

Jaguar cores were replaced by A57 in servers. Now A57 pin compatible with puma+ AM1 for replacement in embedded/tablets/notebooks...

K12 core likely to come in 14/16nm FinFET.

Update: Jim Keller added some details on K12. He referenced AMD's knowledge of doing high frequency designs as well as "extending the range" that ARM is in. Keller also mentioned he told his team to take the best of the big and little cores that AMD presently makes in putting together this design.

By "extending the range" he means scaling it up from phones to desktops and high-performance servers.

It is funny to see another set of predictions confirmed.

 

[juanrga]

Very interesting. So now we know what Jim Keller was hired for, and it wasn't HEDT.

Except it was for AMD future HEDT because he has experience on both high-performance ARM and high-performance x86.

 

[Cazalan]

You don't get it. Your link was irrelevant because its content was irrelevant not because it is from Intel.

If instead deleting the links provided to you, you had read them, you would discover that some of them cite Intel sources. The Intel website mentions that the same architect designing the Phi cores is also designing the new GPU cores for Skylake and successors.

This quote from one of the links it pretty much summarize the situation:

So, over a year later, we were proven right about the lineage of Sky Lake’s GPU. Time to go buy a mole a new set of rubber booties, Intel waters their lawns enough to make his little toes muddy far too often. We told you that Larrabee as a GPU wasn’t dead, but would you listen? Noooo…….

You don't call something that's been completely redesigned by some old name somewhat related to the project. And the link was perfectly valid because they're able to get another 40% performance advantage without even changing the underlying architecture. That's fairly significant for 1 generation, when they can also easily double the transistor count. Broadwell has 48 EU, so Skylake could easily go to 64 EU for the same power.

 

[de5_Roy]

Very interesting. So now we know what Jim Keller was hired for, and it wasn't HEDT.

hedt isn't mentioned anywhere. check the list of markets amd is aiming at - all high volume, semicustom markets. hedt is saturated, mature and strongly dominated by intel's high end cpus, that's not something an arm soc or cat-core soc can overturn. amd is taking the usual route, start with the new stuff at the bottom and working their way up. with bd, they dropped that at the high end and then the skus settled at the middle-bottom of their respective ranges.
the interesting part may be the arm+gcn soc and x86 soc coexisting - major achievement of modular design philosophy. this indicates major overhaul in the unified northbridge and in all it's components. a full-blown hybrid would be an soc running amd-64 and arm v8 instructions in the same silicon. i wonder if that's what amd will aim at achieving....
mantle api boosting such socs may not be mere speculations for long.

finally, low power cat cores rule! power effieciency always prevails!

 

[griptwister]

looks like arm socs are coming to AM1 platform in the near future
http://www.tomshardware.com/news/amd-ambidextrous-x86-arm,26709.html
i wonder what this new "matching x86 design" will be.. certainly doesn't look like replacement, rather coexist.

AMD demos Seattle, its first ARM-based server chip
http://techreport.com/news/26419/amd-demos-seattle-its-first-arm-based-server-chip
AMD is working on K12, brand-new x86 and ARM cores
http://techreport.com/news/26417/amd-is-working-on-k12-brand-new-x86-and-arm-cores

moar promo slides
http://www.anandtech.com/show/7990/amd-announces-k12-core-custom-64bit-arm-design-in-2016
http://www.anandtech.com/show/7989/amd-announces-project-skybridge-pincompatible-arm-and-x86-socs-in-2015

Well fudge me... I was about to go Intel and upgrade my ram....

I'll wait for these new AMD x86 CPUs, my FX8350 does what I need it to do for now...

I think I'll definately hit the GPU upgrade cycle later this year though. I was wondering, any thoughts on this rumored 128bit Windows 9 OS? Would it be worth buying?

Oh, I just thought about this... what if AMD actually tracked this thread and bumped a couple of your guys ideas? Now wouldn't that be something? I'm curious to see what they're planning.

 

[de5_Roy]

I'll wait for these new AMD x86 CPUs, my FX8350 does what I need it to do for now...

I think I'll definately hit the GPU upgrade cycle later this year though. I was wondering, any thoughts on this rumored 128bit Windows 9 OS? Would it be worth buying?

Oh, I just thought about this... what if AMD actually tracked this thread and bumped a couple of your guys ideas? Now wouldn't that be something? I'm curious to see what they're planning.

i don't think 64-bit is running out of steam any time soon. i read somewhere that 64-bit o.s. can address something-exabytes memory, so 128-bit addressing won't be helping with current tasks....