AMD Fusion: How It Started, Where It’s Going, And What It Means

[Guest]

You've already read about APUs, and maybe you're even using them now. But the road to creating APUs was paved with a number of struggles and unsung breakthroughs. This is the story of how hybrid chips came to be at AMD and where they’re going.

AMD Fusion: How It Started, Where It’s Going, And What It Means : Read more

 

[mayankleoboy1]

Wont the OS have to evolve along with the HSA to support it? Can those unified-memory-space and DMA be added to Windows OS just with a newer driver?

With Haswell coming next year, Intel might just beat AMD at HSA. They need to deliver a competitive product.

 

[Dupontrocks11]

Perhaps Skynet will start off as an APU manufacturer. Lol.

 

[pharoahhalfdead]

Informative article that kept my attention. Propz

 

[Reynod]

Great article ... well balanced.

I think you were being overly kind about the current CEO's ability to guide the company forward.

Dirk Meyer's vision is what he is currently leveraging anyway.

A company like that needs executive leadership from someone with engineering vision ... not a beancounter from retail sales of grey boxes.

History will agree with me in the end ... life in the fast lane on the cutting edge isn't the place for accountants and generic managers to lead ... its for a special breed of engineers.

 

[jamesyboy]

AMD is the jack-of-all-trades and the master-at-none. Even the so called "balancing" that they're supposed to be doing is already being done better by Nvidia, ARM, and now Intel with Medfield. AMD doesn't stand a chance trying to bring a ARM like balance to the x86 field. I have no idea what they were thinking when they decided that they'd rather be stuck in between mobile and desktop. They have all of this wonderful IP, all those wonderful engineers. I fear that what's best for AMD will be to leave the x86 battlefield all together, and become a company like Qualcomm or Samsung, and leverage their GPU IP into the Arm world--i fear this because a world where Intel is the only option, is one that's far worse off for the consumer.

They don't have the efficiency of Ivy Bridge, or Medfield, they don't have the power of Ivy Bridge, and they're missing out on this round of the Discrete Graphics battle (they were ahead by so far, but nvidia seems to have pulled an Ace out of their butt with the 600 series). So what exactly IS AMD doing well? HTPC CPUs? Come on! The adoption rate for the system they're proposing with HSA is between 5 and 10 years off....and because they moved too early, and won't be able to compete until then, they have to give the technology away for free to attract developers.

Financially, this a company's (and a CEOs) worst nightmare...they're too far ahead of their time, and the hardware just isn't there yet.

This will end of being just like the tablet in the late '90s, and early '00s. It won't catch on for another decade, and another company will spark, and take advantage of the transition properly, much to AMD's chagrin.

I'm not sure if it was the acquisition of ATI that made AMD feel like it was forced to do this so early, but they aren't going to force the market to do anything. This work should have been done in parallel while making leaps and bounds within the framework of the current model.

You can't lead from behind.

I've always been a fan of AMD. They've brought me so e of the nicest machines I've ever owned...the one that had me, and still have me most excited. But I have, and always will buy what's fastest, or best at the job I need the rig for. And right now...and for the foreseeable future, AMD can't compete on any platform, on any field, any where, at anytime.

AMD just bet it's entire company, the future of ATI (or what was the lovely discrete line at AMD), the future of their x86 platform, and their manufacturing business all on something that it wasn't sure it would even be around to see. They bet the farm on a dream.

Nonetheless, i disagree that you were being overly kind about the CEOs ability to lead the company. I think you're being overly kind for thinking this company has a viable business model at all. Theyll essentially have to become a KIRF (sell products that are essentially a piece o' crud, dirt cheap) f a compay to stay alive.

This is mostly me ragin at the fail. The writer of this article deserves whatever you journalist have for your own version of a Nobel.

This was a seriously thorough analysis, and by far the best tech piece i've seen all year. We need more long-form journalism in the world, for i her way too many people shouting one line blurbs, with zero understanding of the big picture.But i have to say, that while this artucle is 98% complete, you missed speaking anout the fact that this company is a company...an enterprise that survives only with revenue.

 

[A Bad Day]

I really enjoyed this article.

Now, does anyone want to play Crysis in software rendering with max eyecandy?

 

[SteelCity1981]

Kaveri looks very impressive reaching the 1 TFLOP milestone on a APU.

 

[army_ant7]

@jamesyboy:
It's not over until the fat lady sings. As I read your post, I felt that you were missing a (or the) big point of the APU and this article.

It's about how software is developed nowadays and how there is such a huge reserve of potential performance waiting to be tapped into. I could imagine that if future software bite into this "evolution" to more GPGPU programming then I would expect a huge jump in performance even on the current, or shall I see currently being phased out, Llano APU's.

Yes, current discrete GPU systems would improve in performance as well significantly I would think, but to the same degree that APU's would improve, especially with the new technologies to be implemented like unifying memory spaces, etc? I don't think so.

I'm not saying that you're totally wrong. AMD might end up croaking, but we can't say for certain 'til it happens. Don't you agree? (I'm not picking any fights BTW. Just sharing my thoughts.)

 

[blazorthon]

[citation][nom]jamesyboy[/nom]AMD is the jack-of-all-trades and the master-at-none. Even the so called "balancing" that they're supposed to be doing is already being done better by Nvidia, ARM, and now Intel with Medfield. AMD doesn't stand a chance trying to bring a ARM like balance to the x86 field. I have no idea what they were thinking when they decided that they'd rather be stuck in between mobile and desktop. They have all of this wonderful IP, all those wonderful engineers. I fear that what's best for AMD will be to leave the x86 battlefield all together, and become a company like Qualcomm or Samsung, and leverage their GPU IP into the Arm world--i fear this because a world where Intel is the only option, is one that's far worse off for the consumer.They don't have the efficiency of Ivy Bridge, or Medfield, they don't have the power of Ivy Bridge, and they're missing out on this round of the Discrete Graphics battle (they were ahead by so far, but nvidia seems to have pulled an Ace out of their butt with the 600 series). So what exactly IS AMD doing well? HTPC CPUs? Come on! The adoption rate for the system they're proposing with HSA is between 5 and 10 years off....and because they moved too early, and won't be able to compete until then, they have to give the technology away for free to attract developers.Financially, this a company's (and a CEOs) worst nightmare...they're too far ahead of their time, and the hardware just isn't there yet.This will end of being just like the tablet in the late '90s, and early '00s. It won't catch on for another decade, and another company will spark, and take advantage of the transition properly, much to AMD's chagrin.I'm not sure if it was the acquisition of ATI that made AMD feel like it was forced to do this so early, but they aren't going to force the market to do anything. This work should have been done in parallel while making leaps and bounds within the framework of the current model.You can't lead from behind.I've always been a fan of AMD. They've brought me so e of the nicest machines I've ever owned...the one that had me, and still have me most excited. But I have, and always will buy what's fastest, or best at the job I need the rig for. And right now...and for the foreseeable future, AMD can't compete on any platform, on any field, any where, at anytime.AMD just bet it's entire company, the future of ATI (or what was the lovely discrete line at AMD), the future of their x86 platform, and their manufacturing business all on something that it wasn't sure it would even be around to see. They bet the farm on a dream.Nonetheless, i disagree that you were being overly kind about the CEOs ability to lead the company. I think you're being overly kind for thinking this company has a viable business model at all. Theyll essentially have to become a KIRF (sell products that are essentially a piece o' crud, dirt cheap) f a compay to stay alive.This is mostly me ragin at the fail. The writer of this article deserves whatever you journalist have for your own version of a Nobel.This was a seriously thorough analysis, and by far the best tech piece i've seen all year. We need more long-form journalism in the world, for i her way too many people shouting one line blurbs, with zero understanding of the big picture.But i have to say, that while this artucle is 98% complete, you missed speaking anout the fact that this company is a company...an enterprise that survives only with revenue.[/citation]

Funny, but last I checked, AMD's Radeon 7970 GHz edition is the fastest single GPU graphics card for gaming right now, not the GTX 680 anymore. Furthermore, AMD can compete in many markets in both GPU and CPU performance and price. AMD's FX series has great highly threaded integer performance for its price (much more than Intel) and the high end models can have one core per module disabled to make them very competitive with the i5s and i7s in gaming performance. Going into the low end ,the FX-4100 and Llano/Trinity are excellent competitors for Intel. Some of AMD's APUs can be much faster in both CPU and GPU performance than some similarly priced Intel computers, especially in ultrabooks and notebooks where Intel uses mere dual-core CPUs that either lack Hyper-Threading or have such a low frequency that Hyper-Threading isn't nearly enough to catch AMD's APUs. Is this always the case? No, not at all. However, you ignore this when it happens (which isn't rare) and you ignore many other achievements of AMD.

As of right now, there is no retail Nvidia card that has better performance for the money (at least when overclocking is concerned) than some comparably performing AMD cards anymore. The GTX 670 ca't beat the Radeon 7950 in overclocking performance and it can't beat the 7950 in price either. The GTX 680 is no more advantageous against the Radeon 7970 and 7970 GHz Edition. I'm not saying that these cards don't compete well or that they don't have great performance for the money (that would be lying), but they don't win outside of power consumption, which, although important, isn't significant enough of an advantage when the numbers are this close.

Whether or not AMD will fail as a company remains to be seen. Maybe they will, maybe they won't. However, if you want to say that they do, then the supporting info that you give should be more accurate.

 

[army_ant7]

[citation][nom]blazorthon[/nom]As of right now, there is no retail Nvidia card that has better performance for the money (at least when overclocking is concerned) than some comparably performing AMD cards anymore. The GTX 670 ca't beat the Radeon 7950 in overclocking performance and it can't beat the 7950 in price either. The GTX 680 is no more advantageous against the Radeon 7970 and 7970 GHz Edition.[/citation]
Interesting. I didn't know that. Is this generally true about the whole GCN lineup vs. the whole Keppler line up? I'm talking about overclocking performance of course since by default, the high-end Nvidia cards are more recommended, well at least the GTX 670.

 

[doron]

@army-ant7, all comes down to current gpu prices, the games you play and whether or not you're willing to overclock.

 

[Guest]

What a great read, thanks!

Pity it failed to re-ignite my passion for AMD. After the last build of a 1055T I am afraid I am firmly entrenched again with Intel for my needs in x86.

Saying that, I use AMD for 90% of my customer builds due to platform price! So they win the Cheap Stakes!

 

[Device Unknown]

I have been with Tomshardware since Thomas Pabst stopped being a doctor and started this site. This has to be one of the best articles I have read since he left the site and BestOf took over.
-Thanks William.

 

[Device Unknown]

[citation][nom]moricon[/nom]What a great read, thanks!Pity it failed to re-ignite my passion for AMD. After the last build of a 1055T I am afraid I am firmly entrenched again with Intel for my needs in x86.Saying that, I use AMD for 90% of my customer builds due to platform price! So they win the Cheap Stakes![/citation]
Actually bud, Intel prices are comparable to AMD's if you're shooting for price vs. performance. A $200 Intel will out perform a $200 AMD. Then again, I am not familiar with the APU market prices. Those may be worth it.

We predicted the end of x86 awhile back in an article on our website. And I am sticking to it. x86 will either be dead within 5 years, or Intel will HAVE to open up licensing. ARM will be taking over the desktop segment really fast. Mark my words.

 

[ekho]

Thank you William for the article.

 

[jaquith]

Hmm...Yet another Article that pits Intel & AMD haters against each other in mass.

 

[mayankleoboy1]

[citation][nom]jaquith[/nom]Hmm...Yet another Article that pits Intel & AMD haters against each other in mass.[/citation]

and ATI..er.. AMD and Nvidia

 

[gondor]

[citation][nom]jaquith[/nom]Hmm...Yet another Article that pits Intel & AMD haters against each other in mass.[/citation]

Why would it have anything to do with haters ? I use Intel CPU and AMD's graphics card. If I was to upgarde both components today, I'd go for Intel CPU and AMD graphics again but I'm watching AMD's APU efforts very closely and if AMD manages to come up with an APU that has the graphics performance of say HD5770 and CPU performance of a quad core Core2/Nehalem (Q9xxx series Yorkfield/Lynnfield) in a decent power envelope (65W would be awesome) it will be my next purchase instead. AMD appears to be right on the amrk for this kind of performance with Kaveri.

I don't need more performance than that and I'd prefer to get rid of discrete GPU and its fan (which becomes noisy after a while). With single a item to cool (the APU itself) it's easier to get a decent (= quiet and efficient) cooling solution. The fact that HSA is going to offer great speedups with certain loads (data compression/decompression, media encoding etc.) is just an added bonus.

 

[serendipiti]

[citation][nom]jamesyboy[/nom]AMD is the jack-of-all-trades and the master-at-none. ....
This will end of being just like the tablet in the late '90s, and early '00s. " [/citation]

I don't agree. Tablets required several tecnologies (touch: both hardware and software, display, battery...) to develop, while APUs requirements are only in the hardware and software space, and both are perfectly covered by AMD-ATI.

Trinity CPUs are 15% on average slower than intel... I could understand your position if AMD CPUs where 95% slower than intel counterparts... If you look at intel portfolio, you'll find bigger (than 15%) differences in performance, so, do you still think you need the fastest CPU to perform all tasks (and again the term "fastest" is a tricky one...).
Just to compare with cars, why people simply overlook the engine specs (as long the engine fits an "standard" performance) and make its decission based on very different aspects. And why there are several models intended at different uses ?
CPUs alone doesn't make the whole experience. In most of cases I don't need an i7 (just to type this...) but what will make the difference is another factor.
APUs are already here, the timeframe is shorter than 5 - 10 years, it has already started, it will get an inflection point at the end of the next year or so, and will get mainstream (flooding the computing space) in 5 - 10 years.
if only I could guess the lottery numbers that easy

 

[tomfreak]

despite all the talks on APUs, yet they are putting the old less compute VLIW4 GPU on trinity. Why wont they use GCN instead.

on top of that they FAIL to capitalize the hybrid crossfire advantages, if all the APU is able to add 10-15% performance on a 7970 by crossfire, they could kill 2 birds(intel/Nvidia) with 1 stone. People would buy AMD with CPU/APU + discrete GPU as a single package. Yet again, having an trinity as VLIW4 when there isnt a low-end VLIW4 GPU to "properly" crossfire it. IMO with all the asset they have, they are quite mess up.

 

[Guest]

AMD FX Processors perform very well in Linux, Perform as good or better than intel.Did you see any benchmarks? Its all about software capabilities. My opinion.

 

[oomjcv]

William van Winkle sir, I commed thee.

[citation][nom]Tomfreak[/nom]despite all the talks on APUs, yet they are putting the old less compute VLIW4 GPU on trinity. [/citation]

I think the reason for not using GCN is that it is designed on the 28nm lithography, while the CPU cores are using 32nm... could be wrong though.

 

[blazorthon]

[citation][nom]Device Unknown[/nom]Actually bud, Intel prices are comparable to AMD's if you're shooting for price vs. performance. A $200 Intel will out perform a $200 AMD. Then again, I am not familiar with the APU market prices. Those may be worth it.We predicted the end of x86 awhile back in an article on our website. And I am sticking to it. x86 will either be dead within 5 years, or Intel will HAVE to open up licensing. ARM will be taking over the desktop segment really fast. Mark my words.[/citation]

I have to disagree with you. x86 has been used for decades because of backwards-compatibility and the ability to continually scale it up. ARM doesn't scale nearly as well, especially if you try to make a high a=end ARM CPU and then try improving it, the architecture simply isn't designed for that and can't compete outside of very low power deices. PowerPC might be a practical replacement, but that's unlikely. If x86 is to die, then what will replace it? GPUs can run all software and can't run all of an OS too well. They are only very fast for easily parralelized floating point work.

x86 is the cornerstone of complex software. It is complex enough to support many advanced instructions that can far more than make up in performance for its complexity when they are used. Using pretty much any other ISA would not only kill backwards compatibility, but it would also mean that it might be slower in some cases. The micro-architecture might use another ISA below x86 (current CPUs are actually RISC CPUs that decode/encode x86 code into the natice RISC code of the individual execution units), but x86 as the primary architecture as far as Windows and the software is concerned is probably here to stay.

The only semi-modern example of a CPU that didn't use RISC below the x86 layer was the Cruose and it used VLIW below x86 instead of RISC. It was decent for some things, but it didn't work out too well (not because the concept was flawed, but apparently because the company didn't understand the difficulty of the endeavor and quit because of this).

[citation][nom]army_ant7[/nom]Interesting. I didn't know that. Is this generally true about the whole GCN lineup vs. the whole Keppler line up? I'm talking about overclocking performance of course since by default, the high-end Nvidia cards are more recommended, well at least the GTX 670. [/citation]

It actually doesn't have much to do with the architectures and such. This is mostly only true because of the current prices. Overall, the 7950 can overclock equally well to the 7970 (although the 7970 GHz Edition might be able to pull ahead) or actually even slightly better than the 7970s that have the same PCB. Since it can overclock at least as well as the 670, maybe a little better, it's price is the determining factor in whether or not it has the value win.

The 7950 can be bought significantly cheaper than the 670 and often has one or more free gifts that can total over $100 in free gifts. The 670 simply can't compete with that value unless Nvidia does some serious promoting and some price drops, although it is still a great card. For a while, this was not true and Nvidia had the wins, but once AMD got through with some of the latest price drops, they took the lead just like they took the lead with the 7970 GHz Edition in performance.

[citation][nom]Tomfreak[/nom]despite all the talks on APUs, yet they are putting the old less compute VLIW4 GPU on trinity. Why wont they use GCN instead. on top of that they FAIL to capitalize the hybrid crossfire advantages, if all the APU is able to add 10-15% performance on a 7970 by crossfire, they could kill 2 birds(intel/Nvidia) with 1 stone. People would buy AMD with CPU/APU + discrete GPU as a single package. Yet again, having an trinity as VLIW4 when there isnt a low-end VLIW4 GPU to "properly" crossfire it. IMO with all the asset they have, they are quite mess up.[/citation]

AMD had to use 32nm for Trinity because there simply isn't enough 28nm fab time available and they have no access to 22nm right now. VLIW4 was the best option because it is easier to die-shrink something than it is to reverse die-shrink a more complex architecture. Also, up to the 6670 still supports CF with Trinity. I totally agree with you in that they should change this, but the architecture doesn't need to match (obviously) for CF.

They could have at least made the 7750 and 7770 compatible with it and I think that they should have, but oh well. Your point about letting it help very high end cards is a little poor in that the GPU simply isn't fast enough to help those cards more than a few percent. It probably won't be nearly a 10% improvement if it could and the CPU would probably hold everything back to a point where disabling the IGP and overclocking the CPU would probably have greater performance.

 

[busuan]

Borrowing someone else's words: AMD "jumped to a new curve" and there they will struggle and win. Intel will stay as the king of "the old curve" for quite a while but eventually it will either capitulate or fall.