X86's strength will forever be backward compatibility. While for consumer software this isn't a huge deal (often a recompile for a new ISA is all you need to do), for business and gamers, who are the key early adopters, this is huge. As an ISA, X86 is ugly. As implemented, X86 is ugly. But the world runs on X86.
Again, I view everything as moot, since we are heading the way of virtualization in the long run, so the ISA won't matter.
X86's strength will forever be backward compatibility. While for consumer software this isn't a huge deal (often a recompile for a new ISA is all you need to do), for business and gamers, who are the key early adopters, this is huge. As an ISA, X86 is ugly. As implemented, X86 is ugly. But the world runs on X86.
Again, I view everything as moot, since we are heading the way of virtualization in the long run, so the ISA won't matter.
Yes, because, ultimately, regardless of what anyone thinks of M$ (I hate them too...), the world does run on Windows. Though, that dynamic is changing from literally nearly everyone is on Windows, to only a vast majority. Their market share advantage is eroding (in small % chunks, but no less eroding from the 97% they once enjoyed).
Now, does that mean we are going to see wide scale transition from Windows to some other OS? I, for one, seriously doubt it's going to be an overnight transition to something else from Windows.
Will it happen eventually? Possibly, if M$ continues to screw the pooch release, after release, after release.
They have not yet acknowledged that their consumer base does not necessarily want "their vision" of what Windows is or should be. If that enlightenment doesn't come soon enough, then there is a good chance we see further erosion of their once powerful monopoly...(though, let's not jump the gun...80+% is nothing to sneeze at still...)
Its safe to say that AMD are at least trying to change their position rather than stay in the bubble that was essentially going nowhere.
er... not exactly. they're trying to branch into the ARM bubble that's reached it's saturation pressure.
while i agree with arm having the capability to supplant x86 in the long run, the hierarchy in ARM ecosystem will be quite hostile towards amd, both in high volume sector and high margin sector unless amd has something very attractive to offer. right now, they don't. they licensed vanilla arm cores instead of uarch(understandable, considering amd's funds and deadlines), their flagship gpu uarch isn't scalable down to ultramobile level to compete with the likes of adreno (qualcomm) and powervr(imagination tech.)... or even mali - without shrinking and tuning for 20nm, at least.
this isn't even the worst part. that starts when arm finds a foothold in the high margin sector with their a64 uarch and successors, then starts to raise their fees. even right now, smaller vendors quickly fall off of arm competition due to pressure from established players.
I agree with sarinaide, AMD is trying to change their situation.
Which is the alternative that you offer? To remain in a dying x86/Windows market up to being completely crushed by Intel from one side and by ARM from the other?
Considering GIMP is freeware on Linux platforms, it would probably run on anything you could get to emulate, closely enough, x86.
Though, blackkstar's point is valid. If you are looking for this timeline to happen in the next...oh...20 years. You're losing your mind.
Everyone said PPC was going to destroy x86 all the way back 20 years ago. Look at PPC now...not doing so hot in the consumer space right? x86 is far stronger than you suspect...like any other ISA, it has it's own sets of issues. Though, those are not insurmountable challenges.
GIMP is not freeware but FOSS and, as a consequence, has been ported to about a dozen of different architectures, including ARM.
Changes will come much before 20 years. AMD prediction is that big data centers will start porting software to ARM in 2015.
The relation between PPC and x86 and is different between ARM and x86. Did Microsoft release a Windows version for PPC? Sure no.
It discusses why special-purpose processors from Cray, MasPar, Convex, HP were replaced by DEC Alpha, MIPS, and SPARC and how latter those were replaced by x86. Here by "replace" I don't mean completely. According to their figure #1 about a 90% of the supercomputers in the top500 list are x86 today, when x86 was less than a 1% of the list in the year 2000.
There is interesting technical and economical comparison of why the reasons that lead to x86 to replace previous processors are the same why ARM will replace x86.
Among the reasons they mention faster evolution and technical innovation. They even speculate about ARM SoCs are "likely to become early adopters of 3D memory packaging". In fact, I can see ARM people is participating in a talk about future exascale-level memory technologies
One or two days ago, I quoted AMD head saying explicitly why they will not compete with Intel in the high-performance x86 market anymore. I will quote now AMD about their server strategy and why are moving away from x86 to ARM (bold emphasis from mine):
The company said its first ARM server processors - which will be released in the second half of next year - will be faster and more powerful than its existing low-power x86 server processors.
AMD shared initial details on its 64-bit ARM chips, code-named Seattle, which will have up to 16 CPU cores. The chips will be up to four times faster and more power efficient than the quad-core Opteron X-series chips, which draw up to 11 watts of power and are based on the x86 architecture.
AMD will sell its first ARM processors alongside x86 server processors, which were also updated with new high-end Opteron chips due next year. But AMD expects ARM processors to outrun x86 chips in the long term.
Observers have said the ARM-based chips could give AMD an edge over Intel, and the move could be as significant as its introduction of 64-bit server chips in 2003, and dual-core chips in 2004. Both moves gave AMD a competitive advantage over Intel, but server chip delays and the failure of chips based on the Bulldozer core ultimately cost AMD market share.
"ARM servers hit the market for real in 2014, that's the try year," said Andrew Feldman, corporate vice president and general manager of the server business unit at AMD.
By 2016 or 2017, ARM CPUs will have 20 percent of the server market, Feldman said. Right now the server market is dominated by Intel's x86 chips, but Feldman said large data centers will start porting software from x86 to ARM in 2015.
ARM CPUs, which are cheaper and ship in higher volumes compared to the more expensive x86 chips, will prevail in the end, Feldman said.
"In the history of our industry, over the past 40 years, smaller, lower-cost, higher volume CPUs have always won," Feldman said.
AMD has been losing server processor market share to Intel, and its Seattle server chips signify a big change in direction for the company. AMD broke its reliance on the x86 architecture in October last year when it announced that it had licensed ARM's 64-bit architecture and would sell ARM-based server processors next year.
AMD's ARM-based server chips will be targeted at single-socket servers and also dense servers like the company's own SeaMicro servers. AMD will start shipping test units of the chips in the first quarter of 2014.
One or two days ago, I quoted AMD head saying explicitly why they will not compete with Intel in the high-performance x86 market anymore. I will quote now AMD about their server strategy and why are moving away from x86 to ARM (bold emphasis from mine):
The company said its first ARM server processors - which will be released in the second half of next year - will be faster and more powerful than its existing low-power x86 server processors.
AMD shared initial details on its 64-bit ARM chips, code-named Seattle, which will have up to 16 CPU cores. The chips will be up to four times faster and more power efficient than the quad-core Opteron X-series chips, which draw up to 11 watts of power and are based on the x86 architecture.
AMD will sell its first ARM processors alongside x86 server processors, which were also updated with new high-end Opteron chips due next year. But AMD expects ARM processors to outrun x86 chips in the long term.
Observers have said the ARM-based chips could give AMD an edge over Intel, and the move could be as significant as its introduction of 64-bit server chips in 2003, and dual-core chips in 2004. Both moves gave AMD a competitive advantage over Intel, but server chip delays and the failure of chips based on the Bulldozer core ultimately cost AMD market share.
"ARM servers hit the market for real in 2014, that's the try year," said Andrew Feldman, corporate vice president and general manager of the server business unit at AMD.
By 2016 or 2017, ARM CPUs will have 20 percent of the server market, Feldman said. Right now the server market is dominated by Intel's x86 chips, but Feldman said large data centers will start porting software from x86 to ARM in 2015.
ARM CPUs, which are cheaper and ship in higher volumes compared to the more expensive x86 chips, will prevail in the end, Feldman said.
"In the history of our industry, over the past 40 years, smaller, lower-cost, higher volume CPUs have always won," Feldman said.
AMD has been losing server processor market share to Intel, and its Seattle server chips signify a big change in direction for the company. AMD broke its reliance on the x86 architecture in October last year when it announced that it had licensed ARM's 64-bit architecture and would sell ARM-based server processors next year.
AMD's ARM-based server chips will be targeted at single-socket servers and also dense servers like the company's own SeaMicro servers. AMD will start shipping test units of the chips in the first quarter of 2014.
Technically, under US Law, a monopoly of X86 processors wouldn't be recognized as a monopoly, since the CPU market is quite vibrant right now. You have plenty of Windows alternatives, and nothing is stopping you from Windows Apps via WINE on Linux using an ARM CPU.
As for why there is no ARM port of Photoshop, it's because there's no suitable desktop operating system for ARM. Which is why AMD going ARM only puts them in an awkward situation as they would have to basically bow out of the desktop and laptop market completely.
GIMP is already ported to ARM and can be used in an ARM desktop machine (if you have one powerful enough). Photoshop doesn't port to ARM by the same reason that they don't port to any archs/OSs and offer the "Not supported on this architecture" error to some Mac desktop users.
The ARM servers will run an ARM OS that is virtually identical to the x86 version in features. Neither AMD nor someone else will be migrating today to ARM to miss desktop/laptop markets. The migration will be made when those markets are ready for the change, not before.
I already explained this migration is not something that will happen today nor tomorrow. I also offered a possible schedule, with desktops being the last in migrating.
blackkstar :
It would also completely invalidate their dGPU lineup. Firepro would be DOA after AMD abandons x86.
Nope. They will release GCN accelerators for ARM chips. I am said (but cannot verify it) that AMD Seattle is already HSA compatible.
Note also that Nvidia has already ported CUDA to ARM. CUDA is not for playing Android games. It is for High-Performance-Computing.
blackkstar :
Intel would more than likely slowly start phasing out PCIe in HEDT. They would have no reason to keep it around, those slots would only be used to sell competitor's parts. In workstation it would be the difference between customers buying Xeon Phi and customers buying Firepro or Quadro/Tesla.
Intel is already moving the next Xeon Phi to socket package for blocking customers to use FirePro/Teslas with Xeon CPUs.
blackkstar :
When you're speaking of AMD going ARM only in the future,
I didn't say "only", true?
You can't compare FLOSS projects to proprietary ones when it comes to compilation and ports.
I have a fully optimized version of GIMP for my FX 8350. Do you think Adobe would ever release a Piledriver enhanced version of Photoshop?
FLOSS projects are ported because a niche sector needs them to work while proprietary software is only ported when it is deemed profitable. Look how long Linux users have been barking for native Photoshop ports and look how mum Adobe has been on it. Yet GIMP makes it to ARM GNU/Linux of all things faster.
ARM + HSA is only going to be good for HPC workloads where you have access to compiling the software you want to run on your own. It won't ever be a replacement for a Windows laptop or desktop. It won't even be viable in an environment where there's a little bit of proprietary solution that isn't working with HSA and hasn't been ported to ARM.
It is a mess. Maybe I am just bitter because I am a Linux guy and Linux people who wanted x86 ports of software had to go as far as implementing Windows API in Linux because the software vendors who they have been begging to support x86 Linux won't do it.
So you can see why I'm sceptical of ARM ports of anything proprietary coming out. Companies are sluggish to port to x86 Linux, for them to port to something that's a fraction of a fraction of that (HSA enabled ARM Linux) seems like you'd have better luck winning the lottery.
If you want my opinion on this, we will see AMD have a huge contract win where AMD sells them a bunch of ARM APUs with GCN HSA enabled and that will be about the end of it. It makes more sense that way since AMD has been focusing on delivering products that actually server a purpose to a customer as opposed to creating parts and then waiting for companies to build something with them.
I do think that AMD is going to keep their x86 hardware lineup unchanged. I suppose what I'm getting at is you can think of ARM APU on AMD as a product like 8 core Jaguar for PS4 or xbone. Special product ordered by company that doesn't affect the lineup of other existing "legacy" products.
I would even go as far as to say that "legacy" market could even refer to a legacy market where customers don't order chips and instead buy what fits their needs best from the AMD lineup while the new customers with more money order specialized parts, like ARM APU with HSA.
Also, your pcmag link is only discussing servers. Which, IMHO, gives credibility to the fact that we're done seeing AMD sell CPUs that are released on server first and then HEDT second and instead just designing chips that work around HEDT gaming PCs first. It frees AMD from having to worry about how the chip will perform in server workloads while maintaining gaming performance or workstation performance.
I remain sceptical of ARM APU servers. They have potential but they will only work in situations where the client can use programs they can compile themselves AND where the software they want to use has HSA features or they can add their own.
For traditional microserver workloads, an ARM APU is completely useless as there's nothing there to take advantage of HSA and clients would be better off with a non-APU for that if it's a need served by traditional microservers.
Java is getting HSA which will help a ton, but it still is going to require HSA enabled code, so you end up waiting on ports if you're a proprietary software guy.
Due to x86 monopoly, AMD cannot license x86 technology to third parties, only sell the x86 silicon they make. Nvidia couldn't even obtain a x86 license.
I think this may help answer some questions you may have.
It did help to understand that you are trolling when you are not too busy bashing AMD.
Im not bashing AMD, im bashing people who think an APU is going to be an awesome replacement for HEDT and only listens to marketing hype and badly written articles claiming AMD is making 100% profit on a product they are selling.
I am bashing people who post and repost articles that aren't even remotely true, hence the link I provided for any people who might need it.
You are bashing AMD with false statements. You take an article about Kaveri performance using ordinary CPU workloads and you claim it is using "APU optimized software". You take an article that doesn't claim a 100% profit (at contrary) and you claim it says 100% profit. You avoid technical discussion and pretend there is only marketing hype. You take another article with benchmarks and claim that are FAKE, because you pretend that Kaveri /Steamroller is 36% slower than Bulldozer... And lots of anti-AMD BS as that.
When have I ever claimed that those games were "APU optimized software"? This is a blatent lie made up by you. Kaveri will suck in those games I listed.
AMD sells a chip to sony for $100. Article states that. Then the article goes and sais AMD makes $100 per PS4 sold. One of those statemens is a lie.
Technical discussion about Kaveri owning DT is based on HSA enabled software only. You pretend this is all workloads and go around spreading it as ERMAGO.
Those fake numbers posted from the cosmology leak are fake. I don't know what ES chip that is, neither do you. Its someone's best GUESS that it was a kaveri. read this statement carefully. THE BD/PD NUMBERS WERE MADE UP LIES. END OF STORY I can't test kaveri, they can't test kaveri because that ES probably isn't even kaveri. I can however test BD and those BD/PD numbers are 100% FAKE, AKA LIES.
but hey, keep on believing that everything on the internet is true. It makes for a comical reading.
One or two days ago, I quoted AMD head saying explicitly why they will not compete with Intel in the high-performance x86 market anymore. I will quote now AMD about their server strategy and why are moving away from x86 to ARM (bold emphasis from mine):
The company said its first ARM server processors - which will be released in the second half of next year - will be faster and more powerful than its existing low-power x86 server processors.
AMD shared initial details on its 64-bit ARM chips, code-named Seattle, which will have up to 16 CPU cores. The chips will be up to four times faster and more power efficient than the quad-core Opteron X-series chips, which draw up to 11 watts of power and are based on the x86 architecture.
AMD will sell its first ARM processors alongside x86 server processors, which were also updated with new high-end Opteron chips due next year. But AMD expects ARM processors to outrun x86 chips in the long term.
Observers have said the ARM-based chips could give AMD an edge over Intel, and the move could be as significant as its introduction of 64-bit server chips in 2003, and dual-core chips in 2004. Both moves gave AMD a competitive advantage over Intel, but server chip delays and the failure of chips based on the Bulldozer core ultimately cost AMD market share.
"ARM servers hit the market for real in 2014, that's the try year," said Andrew Feldman, corporate vice president and general manager of the server business unit at AMD.
By 2016 or 2017, ARM CPUs will have 20 percent of the server market, Feldman said. Right now the server market is dominated by Intel's x86 chips, but Feldman said large data centers will start porting software from x86 to ARM in 2015.
ARM CPUs, which are cheaper and ship in higher volumes compared to the more expensive x86 chips, will prevail in the end, Feldman said.
"In the history of our industry, over the past 40 years, smaller, lower-cost, higher volume CPUs have always won," Feldman said.
AMD has been losing server processor market share to Intel, and its Seattle server chips signify a big change in direction for the company. AMD broke its reliance on the x86 architecture in October last year when it announced that it had licensed ARM's 64-bit architecture and would sell ARM-based server processors next year.
AMD's ARM-based server chips will be targeted at single-socket servers and also dense servers like the company's own SeaMicro servers. AMD will start shipping test units of the chips in the first quarter of 2014.
One or two days ago, I quoted AMD head saying explicitly why they will not compete with Intel in the high-performance x86 market anymore. I will quote now AMD about their server strategy and why are moving away from x86 to ARM (bold emphasis from mine):
The company said its first ARM server processors - which will be released in the second half of next year - will be faster and more powerful than its existing low-power x86 server processors.
AMD shared initial details on its 64-bit ARM chips, code-named Seattle, which will have up to 16 CPU cores. The chips will be up to four times faster and more power efficient than the quad-core Opteron X-series chips, which draw up to 11 watts of power and are based on the x86 architecture.
AMD will sell its first ARM processors alongside x86 server processors, which were also updated with new high-end Opteron chips due next year. But AMD expects ARM processors to outrun x86 chips in the long term.
Observers have said the ARM-based chips could give AMD an edge over Intel, and the move could be as significant as its introduction of 64-bit server chips in 2003, and dual-core chips in 2004. Both moves gave AMD a competitive advantage over Intel, but server chip delays and the failure of chips based on the Bulldozer core ultimately cost AMD market share.
"ARM servers hit the market for real in 2014, that's the try year," said Andrew Feldman, corporate vice president and general manager of the server business unit at AMD.
By 2016 or 2017, ARM CPUs will have 20 percent of the server market, Feldman said. Right now the server market is dominated by Intel's x86 chips, but Feldman said large data centers will start porting software from x86 to ARM in 2015.
ARM CPUs, which are cheaper and ship in higher volumes compared to the more expensive x86 chips, will prevail in the end, Feldman said.
"In the history of our industry, over the past 40 years, smaller, lower-cost, higher volume CPUs have always won," Feldman said.
AMD has been losing server processor market share to Intel, and its Seattle server chips signify a big change in direction for the company. AMD broke its reliance on the x86 architecture in October last year when it announced that it had licensed ARM's 64-bit architecture and would sell ARM-based server processors next year.
AMD's ARM-based server chips will be targeted at single-socket servers and also dense servers like the company's own SeaMicro servers. AMD will start shipping test units of the chips in the first quarter of 2014.
AMD will sell its first ARM processors alongside x86 server processors.
AMD will sell LOW POWER ARM processors.
AMD will sell high-end Opteron chips
Those are 2 different markets. Those are not the same market. Those are 2 different markets.
But AMD expects ARM processors to outrun x86 chips in the long term
I have asked before. HOW LONG TILL ARM SURPASSES HIGH END x86? 2, 5, 7, 9 years? Long term is a pretty general statement, kinda like saying that today's cpus will eventually be replaced.
(iv) Which is the efficiency of GCN compared to VLIW4?
Here is what I found using Tom's graphics cards charts for the reference HD 6970 and 7970
Extreme performance index (I thought it appropriate since these are top tier cards):
HD 6970 -- 84.57%
HD 7970 -- 120.27%
Power consumption at load:
HD 6970 -- 234 Watts
HD 7970 -- 254 Watts
Efficiency as Index % per Watt:
HD 6970 ~ 0.36
HD 7970 ~ 0.47
Efficiency gain from VLIW4 to GCN ~31%.
That is not valid, because is comparing total power consumption of two different cards made in very different processes (40nm vs 28nm).
then you can't compare the efficiency of kaveri to bd/pd. you can't compare the efficiency of sandy bridge to ivy bridge. those are invalid because they are different processes.
(iv) Which is the efficiency of GCN compared to VLIW4?
Here is what I found using Tom's graphics cards charts for the reference HD 6970 and 7970
Extreme performance index (I thought it appropriate since these are top tier cards):
HD 6970 -- 84.57%
HD 7970 -- 120.27%
Power consumption at load:
HD 6970 -- 234 Watts
HD 7970 -- 254 Watts
Efficiency as Index % per Watt:
HD 6970 ~ 0.36
HD 7970 ~ 0.47
Efficiency gain from VLIW4 to GCN ~31%.
That is not valid, because is comparing total power consumption of two different cards made in very different processes (40nm vs 28nm).
So, those numbers are inaccurate because they represent valid comparisons? Or they are inaccurate because they don't suit your purpose? I am confused.
That is basically the point. GCN > VLIW4. Even if VLIW4 was on 28nm, it would still not be as good as GCN. FYI: they went to GCN because it was good for compute, and game developers are appreciating the ease of coding for it, plus it pushes their HSA mantra by allowing easier/better GPU compute.
Add in that it consumes less power per watt, increases performance per thermal envelope, and nearly double TFLOPS in this last generation on the same node (which evidently is your holy grail of measurements judging by kaveri and the gflops quote you keep pushing).
Where are you missing the picture? It's clear as day...
GCN uses less power than VLIW4 cores. PERIOD. End of discussion.
One or two days ago, I quoted AMD head saying explicitly why they will not compete with Intel in the high-performance x86 market anymore. I will quote now AMD about their server strategy and why are moving away from x86 to ARM (bold emphasis from mine):
The company said its first ARM server processors - which will be released in the second half of next year - will be faster and more powerful than its existing low-power x86 server processors.
AMD shared initial details on its 64-bit ARM chips, code-named Seattle, which will have up to 16 CPU cores. The chips will be up to four times faster and more power efficient than the quad-core Opteron X-series chips, which draw up to 11 watts of power and are based on the x86 architecture.
AMD will sell its first ARM processors alongside x86 server processors, which were also updated with new high-end Opteron chips due next year. But AMD expects ARM processors to outrun x86 chips in the long term.
Observers have said the ARM-based chips could give AMD an edge over Intel, and the move could be as significant as its introduction of 64-bit server chips in 2003, and dual-core chips in 2004. Both moves gave AMD a competitive advantage over Intel, but server chip delays and the failure of chips based on the Bulldozer core ultimately cost AMD market share.
"ARM servers hit the market for real in 2014, that's the try year," said Andrew Feldman, corporate vice president and general manager of the server business unit at AMD.
By 2016 or 2017, ARM CPUs will have 20 percent of the server market, Feldman said. Right now the server market is dominated by Intel's x86 chips, but Feldman said large data centers will start porting software from x86 to ARM in 2015.
ARM CPUs, which are cheaper and ship in higher volumes compared to the more expensive x86 chips, will prevail in the end, Feldman said.
"In the history of our industry, over the past 40 years, smaller, lower-cost, higher volume CPUs have always won," Feldman said.
AMD has been losing server processor market share to Intel, and its Seattle server chips signify a big change in direction for the company. AMD broke its reliance on the x86 architecture in October last year when it announced that it had licensed ARM's 64-bit architecture and would sell ARM-based server processors next year.
AMD's ARM-based server chips will be targeted at single-socket servers and also dense servers like the company's own SeaMicro servers. AMD will start shipping test units of the chips in the first quarter of 2014.
Except the wall street, google, blizzard, Verizon, and amazon wins...right? All those were in the last 6 months (i.e. since that statement was made).
What are you trying to say?
That recent SeaMicro Opteron design wins are stirring the server segment for AMD. It's not huge gains, but it is likely helping them out greatly because the margins are big.
Technically, under US Law, a monopoly of X86 processors wouldn't be recognized as a monopoly, since the CPU market is quite vibrant right now. You have plenty of Windows alternatives, and nothing is stopping you from Windows Apps via WINE on Linux using an ARM CPU.
As for why there is no ARM port of Photoshop, it's because there's no suitable desktop operating system for ARM. Which is why AMD going ARM only puts them in an awkward situation as they would have to basically bow out of the desktop and laptop market completely.
GIMP is already ported to ARM and can be used in an ARM desktop machine (if you have one powerful enough). Photoshop doesn't port to ARM by the same reason that they don't port to any archs/OSs and offer the "Not supported on this architecture" error to some Mac desktop users.
The ARM servers will run an ARM OS that is virtually identical to the x86 version in features. Neither AMD nor someone else will be migrating today to ARM to miss desktop/laptop markets. The migration will be made when those markets are ready for the change, not before.
I already explained this migration is not something that will happen today nor tomorrow. I also offered a possible schedule, with desktops being the last in migrating.
blackkstar :
It would also completely invalidate their dGPU lineup. Firepro would be DOA after AMD abandons x86.
Nope. They will release GCN accelerators for ARM chips. I am said (but cannot verify it) that AMD Seattle is already HSA compatible.
Note also that Nvidia has already ported CUDA to ARM. CUDA is not for playing Android games. It is for High-Performance-Computing.
blackkstar :
Intel would more than likely slowly start phasing out PCIe in HEDT. They would have no reason to keep it around, those slots would only be used to sell competitor's parts. In workstation it would be the difference between customers buying Xeon Phi and customers buying Firepro or Quadro/Tesla.
Intel is already moving the next Xeon Phi to socket package for blocking customers to use FirePro/Teslas with Xeon CPUs.
blackkstar :
When you're speaking of AMD going ARM only in the future,
I didn't say "only", true?
You can't compare FLOSS projects to proprietary ones when it comes to compilation and ports.
I have a fully optimized version of GIMP for my FX 8350. Do you think Adobe would ever release a Piledriver enhanced version of Photoshop?
FLOSS projects are ported because a niche sector needs them to work while proprietary software is only ported when it is deemed profitable. Look how long Linux users have been barking for native Photoshop ports and look how mum Adobe has been on it. Yet GIMP makes it to ARM GNU/Linux of all things faster.
ARM + HSA is only going to be good for HPC workloads where you have access to compiling the software you want to run on your own. It won't ever be a replacement for a Windows laptop or desktop. It won't even be viable in an environment where there's a little bit of proprietary solution that isn't working with HSA and hasn't been ported to ARM.
It is a mess. Maybe I am just bitter because I am a Linux guy and Linux people who wanted x86 ports of software had to go as far as implementing Windows API in Linux because the software vendors who they have been begging to support x86 Linux won't do it.
So you can see why I'm sceptical of ARM ports of anything proprietary coming out. Companies are sluggish to port to x86 Linux, for them to port to something that's a fraction of a fraction of that (HSA enabled ARM Linux) seems like you'd have better luck winning the lottery.
If you want my opinion on this, we will see AMD have a huge contract win where AMD sells them a bunch of ARM APUs with GCN HSA enabled and that will be about the end of it. It makes more sense that way since AMD has been focusing on delivering products that actually server a purpose to a customer as opposed to creating parts and then waiting for companies to build something with them.
I do think that AMD is going to keep their x86 hardware lineup unchanged. I suppose what I'm getting at is you can think of ARM APU on AMD as a product like 8 core Jaguar for PS4 or xbone. Special product ordered by company that doesn't affect the lineup of other existing "legacy" products.
I would even go as far as to say that "legacy" market could even refer to a legacy market where customers don't order chips and instead buy what fits their needs best from the AMD lineup while the new customers with more money order specialized parts, like ARM APU with HSA.
Also, your pcmag link is only discussing servers. Which, IMHO, gives credibility to the fact that we're done seeing AMD sell CPUs that are released on server first and then HEDT second and instead just designing chips that work around HEDT gaming PCs first. It frees AMD from having to worry about how the chip will perform in server workloads while maintaining gaming performance or workstation performance.
I remain sceptical of ARM APU servers. They have potential but they will only work in situations where the client can use programs they can compile themselves AND where the software they want to use has HSA features or they can add their own.
For traditional microserver workloads, an ARM APU is completely useless as there's nothing there to take advantage of HSA and clients would be better off with a non-APU for that if it's a need served by traditional microservers.
Java is getting HSA which will help a ton, but it still is going to require HSA enabled code, so you end up waiting on ports if you're a proprietary software guy.
I am not comparing FOSS to proprietary in that way. I was showing how one can run Photoshop-like software in ARM hardware, because there is nothing 'magical' on x86 ISA.
Adobe doesn't release a version of Photoshop for the FX-8350 because there is no demand. If the FX-8350 was in a 90% of all the PCs, Adobe would optimize its software for it.
Adobe will release full Photoshop for ARM when there is enough ARM desktop/workstations for that kind of move. For instance, if Apple migrates to all ARM. Adobe will release a version of Photoshop for Apple-ARM.
ARM has joined the HSA foundation for spreading HSA from phones to HPC/servers. The PC desktop market is also in their target.
Again, I am not saying all this will happen tomorrow or the day after, only that it will happen.
Technically, under US Law, a monopoly of X86 processors wouldn't be recognized as a monopoly, since the CPU market is quite vibrant right now. You have plenty of Windows alternatives, and nothing is stopping you from Windows Apps via WINE on Linux using an ARM CPU.
As for why there is no ARM port of Photoshop, it's because there's no suitable desktop operating system for ARM. Which is why AMD going ARM only puts them in an awkward situation as they would have to basically bow out of the desktop and laptop market completely.
GIMP is already ported to ARM and can be used in an ARM desktop machine (if you have one powerful enough). Photoshop doesn't port to ARM by the same reason that they don't port to any archs/OSs and offer the "Not supported on this architecture" error to some Mac desktop users.
The ARM servers will run an ARM OS that is virtually identical to the x86 version in features. Neither AMD nor someone else will be migrating today to ARM to miss desktop/laptop markets. The migration will be made when those markets are ready for the change, not before.
I already explained this migration is not something that will happen today nor tomorrow. I also offered a possible schedule, with desktops being the last in migrating.
blackkstar :
It would also completely invalidate their dGPU lineup. Firepro would be DOA after AMD abandons x86.
Nope. They will release GCN accelerators for ARM chips. I am said (but cannot verify it) that AMD Seattle is already HSA compatible.
Note also that Nvidia has already ported CUDA to ARM. CUDA is not for playing Android games. It is for High-Performance-Computing.
blackkstar :
Intel would more than likely slowly start phasing out PCIe in HEDT. They would have no reason to keep it around, those slots would only be used to sell competitor's parts. In workstation it would be the difference between customers buying Xeon Phi and customers buying Firepro or Quadro/Tesla.
Intel is already moving the next Xeon Phi to socket package for blocking customers to use FirePro/Teslas with Xeon CPUs.
blackkstar :
When you're speaking of AMD going ARM only in the future,
I didn't say "only", true?
You can't compare FLOSS projects to proprietary ones when it comes to compilation and ports.
I have a fully optimized version of GIMP for my FX 8350. Do you think Adobe would ever release a Piledriver enhanced version of Photoshop?
FLOSS projects are ported because a niche sector needs them to work while proprietary software is only ported when it is deemed profitable. Look how long Linux users have been barking for native Photoshop ports and look how mum Adobe has been on it. Yet GIMP makes it to ARM GNU/Linux of all things faster.
ARM + HSA is only going to be good for HPC workloads where you have access to compiling the software you want to run on your own. It won't ever be a replacement for a Windows laptop or desktop. It won't even be viable in an environment where there's a little bit of proprietary solution that isn't working with HSA and hasn't been ported to ARM.
It is a mess. Maybe I am just bitter because I am a Linux guy and Linux people who wanted x86 ports of software had to go as far as implementing Windows API in Linux because the software vendors who they have been begging to support x86 Linux won't do it.
So you can see why I'm sceptical of ARM ports of anything proprietary coming out. Companies are sluggish to port to x86 Linux, for them to port to something that's a fraction of a fraction of that (HSA enabled ARM Linux) seems like you'd have better luck winning the lottery.
If you want my opinion on this, we will see AMD have a huge contract win where AMD sells them a bunch of ARM APUs with GCN HSA enabled and that will be about the end of it. It makes more sense that way since AMD has been focusing on delivering products that actually server a purpose to a customer as opposed to creating parts and then waiting for companies to build something with them.
I do think that AMD is going to keep their x86 hardware lineup unchanged. I suppose what I'm getting at is you can think of ARM APU on AMD as a product like 8 core Jaguar for PS4 or xbone. Special product ordered by company that doesn't affect the lineup of other existing "legacy" products.
I would even go as far as to say that "legacy" market could even refer to a legacy market where customers don't order chips and instead buy what fits their needs best from the AMD lineup while the new customers with more money order specialized parts, like ARM APU with HSA.
Also, your pcmag link is only discussing servers. Which, IMHO, gives credibility to the fact that we're done seeing AMD sell CPUs that are released on server first and then HEDT second and instead just designing chips that work around HEDT gaming PCs first. It frees AMD from having to worry about how the chip will perform in server workloads while maintaining gaming performance or workstation performance.
I remain sceptical of ARM APU servers. They have potential but they will only work in situations where the client can use programs they can compile themselves AND where the software they want to use has HSA features or they can add their own.
For traditional microserver workloads, an ARM APU is completely useless as there's nothing there to take advantage of HSA and clients would be better off with a non-APU for that if it's a need served by traditional microservers.
Java is getting HSA which will help a ton, but it still is going to require HSA enabled code, so you end up waiting on ports if you're a proprietary software guy.
I am not comparing FOSS to proprietary in that way. I was showing how one can run Photoshop-like software in ARM hardware, because there is nothing 'magical' on x86 ISA.
Adobe doesn't release a version of Photoshop for the FX-8350 because there is no demand. If the FX-8350 was in a 90% of all the PCs, Adobe would optimize its software for it.
Adobe will release full Photoshop for ARM when there is enough ARM desktop/workstations for that kind of move. For instance, if Apple migrates to all ARM. Adobe will release a version of Photoshop for Apple-ARM.
ARM has joined the HSA foundation for spreading HSA from phones to HPC/servers. The PC desktop market is also in their target.
Again, I am not saying all this will happen tomorrow or the day after, only that it will happen.
Except adobe just finished optimizing all their products for Piledriver architecture and GCN...or did you miss that announcement on AMD's page?
Off subject but WTF is the point of the xbox one again? Gets reviewed with a lower score than the PS4 has less power that is not even at question games like dead rising 3 gets 20FPS at times if not 16 BF4/COD has a lower res than the PS4. Ryse of rome is considered to be horrible by reviews. Have to admit though its cute when Microsoft tries.
Off subject but WTF is the point of the xbox one again? Gets reviewed with a lower score than the PS4 has less power that is not even at question games like dead rising 3 gets 20FPS at times if not 16 BF4/COD has a lower res than the PS4. Ryse of rome is considered to be horrible by reviews. Have to admit though its cute when Microsoft tries.
Microsoft is trying to make money...they don't want to sell you a gaming machine. They want you to buy a set top box that happens to play games. They want you to buy streaming services, and all sorts of other subscription add on crap to milk you for every penny.
EDIT: Another Opteron design win, this one from the US Government, giving a contract for development of a HPC based on AMD technology:
It's a research award, not a server solution. The brand name won't matter if what comes out of that research is that they need to develop another Qbit computing machine for the task.
Notice 2 keywords from the article/news: "power consumption" and "extreme-scale computing".
What do you think those keywords might point at?
GPU power first and either ARM or X86 supporting it. Now, I would say ARM has a bigger change of being "extreme-scalar"; you just need something that can push/feed the GPU array.
But oh well, that's with my own interpretation of the thingy.
Technically, under US Law, a monopoly of X86 processors wouldn't be recognized as a monopoly, since the CPU market is quite vibrant right now. You have plenty of Windows alternatives, and nothing is stopping you from Windows Apps via WINE on Linux using an ARM CPU.
As for why there is no ARM port of Photoshop, it's because there's no suitable desktop operating system for ARM. Which is why AMD going ARM only puts them in an awkward situation as they would have to basically bow out of the desktop and laptop market completely.
GIMP is already ported to ARM and can be used in an ARM desktop machine (if you have one powerful enough). Photoshop doesn't port to ARM by the same reason that they don't port to any archs/OSs and offer the "Not supported on this architecture" error to some Mac desktop users.
The ARM servers will run an ARM OS that is virtually identical to the x86 version in features. Neither AMD nor someone else will be migrating today to ARM to miss desktop/laptop markets. The migration will be made when those markets are ready for the change, not before.
I already explained this migration is not something that will happen today nor tomorrow. I also offered a possible schedule, with desktops being the last in migrating.
blackkstar :
It would also completely invalidate their dGPU lineup. Firepro would be DOA after AMD abandons x86.
Nope. They will release GCN accelerators for ARM chips. I am said (but cannot verify it) that AMD Seattle is already HSA compatible.
Note also that Nvidia has already ported CUDA to ARM. CUDA is not for playing Android games. It is for High-Performance-Computing.
blackkstar :
Intel would more than likely slowly start phasing out PCIe in HEDT. They would have no reason to keep it around, those slots would only be used to sell competitor's parts. In workstation it would be the difference between customers buying Xeon Phi and customers buying Firepro or Quadro/Tesla.
Intel is already moving the next Xeon Phi to socket package for blocking customers to use FirePro/Teslas with Xeon CPUs.
blackkstar :
When you're speaking of AMD going ARM only in the future,
I didn't say "only", true?
You can't compare FLOSS projects to proprietary ones when it comes to compilation and ports.
I have a fully optimized version of GIMP for my FX 8350. Do you think Adobe would ever release a Piledriver enhanced version of Photoshop?
FLOSS projects are ported because a niche sector needs them to work while proprietary software is only ported when it is deemed profitable. Look how long Linux users have been barking for native Photoshop ports and look how mum Adobe has been on it. Yet GIMP makes it to ARM GNU/Linux of all things faster.
ARM + HSA is only going to be good for HPC workloads where you have access to compiling the software you want to run on your own. It won't ever be a replacement for a Windows laptop or desktop. It won't even be viable in an environment where there's a little bit of proprietary solution that isn't working with HSA and hasn't been ported to ARM.
It is a mess. Maybe I am just bitter because I am a Linux guy and Linux people who wanted x86 ports of software had to go as far as implementing Windows API in Linux because the software vendors who they have been begging to support x86 Linux won't do it.
So you can see why I'm sceptical of ARM ports of anything proprietary coming out. Companies are sluggish to port to x86 Linux, for them to port to something that's a fraction of a fraction of that (HSA enabled ARM Linux) seems like you'd have better luck winning the lottery.
If you want my opinion on this, we will see AMD have a huge contract win where AMD sells them a bunch of ARM APUs with GCN HSA enabled and that will be about the end of it. It makes more sense that way since AMD has been focusing on delivering products that actually server a purpose to a customer as opposed to creating parts and then waiting for companies to build something with them.
I do think that AMD is going to keep their x86 hardware lineup unchanged. I suppose what I'm getting at is you can think of ARM APU on AMD as a product like 8 core Jaguar for PS4 or xbone. Special product ordered by company that doesn't affect the lineup of other existing "legacy" products.
I would even go as far as to say that "legacy" market could even refer to a legacy market where customers don't order chips and instead buy what fits their needs best from the AMD lineup while the new customers with more money order specialized parts, like ARM APU with HSA.
Also, your pcmag link is only discussing servers. Which, IMHO, gives credibility to the fact that we're done seeing AMD sell CPUs that are released on server first and then HEDT second and instead just designing chips that work around HEDT gaming PCs first. It frees AMD from having to worry about how the chip will perform in server workloads while maintaining gaming performance or workstation performance.
I remain sceptical of ARM APU servers. They have potential but they will only work in situations where the client can use programs they can compile themselves AND where the software they want to use has HSA features or they can add their own.
For traditional microserver workloads, an ARM APU is completely useless as there's nothing there to take advantage of HSA and clients would be better off with a non-APU for that if it's a need served by traditional microservers.
Java is getting HSA which will help a ton, but it still is going to require HSA enabled code, so you end up waiting on ports if you're a proprietary software guy.
I am not comparing FOSS to proprietary in that way. I was showing how one can run Photoshop-like software in ARM hardware, because there is nothing 'magical' on x86 ISA.
Adobe doesn't release a version of Photoshop for the FX-8350 because there is no demand. If the FX-8350 was in a 90% of all the PCs, Adobe would optimize its software for it.
Adobe will release full Photoshop for ARM when there is enough ARM desktop/workstations for that kind of move. For instance, if Apple migrates to all ARM. Adobe will release a version of Photoshop for Apple-ARM.
ARM has joined the HSA foundation for spreading HSA from phones to HPC/servers. The PC desktop market is also in their target.
Again, I am not saying all this will happen tomorrow or the day after, only that it will happen.
Yeah, but you were making the example that GIMP runs on ARM so it's ok. I can tell you as a user of both Photoshop and GIMP that GIMP is no where near a replacement for Photoshop. GIMP is good software and I love it, but there are things that just don't work as well or are completely missing.
But here's the thing, even if ARM did gain marketshare it would take years upon years for it to get any sort of decent marketshare on desktop, laptop, and workstation.
And therein lies the problem with ARM going anywhere on desktop. You're going to have to convince people to go ARM and then wait for the software to show up.
It's pretty hard to sell someone a computer and tell them "you can use GIMP for now but maybe in the future if enough people buy ARM you can use Photoshop" when they can go out and buy x86 right away.
AMD is dealing with getting the hardware out that works on Mantle and HSA by selling people hardware under the premise of PC gaming. A day will come when software shows up and software vendors go "oh, by the way, you have GCN so you can use HSA and your program is now significantly faster". And in the meantime the people who own GCN products will be enjoying them for gaming.
ARM DT has absolutely no inroads at this time for that. They won't offer anything different than x86 and yet will want to supplant x86? It's impossible. ARM will forever be relegated to tablets and phones and at best ARM will get lucky and land some roles in servers, but that's it.
Let me put this into perspective. There's going to have to be a time when ARM DT and laptops are sold and there are none out there.
So imagine I am Lenovo, and I have three 11.6in laptops to choose from. Which one do you take?
A. The Intel Pentium or i3
B. The AMD Kabini APU
C. The ARM 116in laptop
Or even better yet, if you're so sure of ARM supplanting x86, tell me why you would buy an ARM laptop for your next device instead of x86. A group of people are going to have to make that jump and buying an ARM DT or laptop has no benefits over x86 at all. Who is going to go "oh golly gee the x86 ISA is really bloated so I'm going to get this ARM, I don't need Photoshop, recent versions of Flash, etc"
There will probably be a time (if it hasn't already) when there are more ARMs than x86s but ARM will never fully replace x86. Not unless ARM comes up with something that's significantly better than x86 at x86 dominated tasks like workstation type things and gaming.
I don't know of a single person who owns a tablet who has not asked me about either replacing it or getting a Windows laptop that can run regular programs they're used to.
Facebook
Twitter
Instagram