News & Analysis
Comment
Roba66
That was really open source. Unfortunately it was not useful as a piece of IP ...
Roba66
AMD GPU technology is not "OPEN" at all, this is a marketing ploy. The support ...
Why AMD is opening up Fusion: CPU, GPU cores are the new gates
Peter Clarke
6/16/2011 6:47 AM EDT
Write software once and reuse
There is another reason why AMD's move upstream is important.
Software developers want to write once, compile as necessary, run on many platforms and to know that the code they have written can be reused over multiple generations of hardware. The advent of multicore processing risks destroying that situation. The hardware companies have to do something to preserve code portability and that probably means a broad set of standardization issues.
And so AMD is attempting to ease the path of CPU/GPU programming under the Fusion System Architecture intermediate layer, which it calls FSAIL and which it describes as a virtual instruction set architecture. That intermediate layer then calls on the underlying resources, with automated load balancing and dispatch and AMD wants FSA to be able to cope with those resources whether they are x86, ARM, Mali, PowerVR and so on.
But as the eco-system Kool-aid kicked in AMD realized that such a shift in processor ideology is big and more than it could possibly achieve on its own. To make the Fusion System Architecture something worth building it has to offer software developers the broadest possible platform, and guarantees of portability. And who knows more about assembling a broad club of chipmaking partners than ARM.
The question remains – and this Phil Rogers was not laying out for the audience – as to how AMD will make its money out of the Fusion System Architecture.
Is it by being the best implementor of FSA in silicon? This is the historical AMD model in that is has competed with other x86 chip makers. Or is it by licensing the Fusion System Architecture, which is the ARM model? And there is good reason why ARM does not try to compete with its licensees.
Probably AMD will stick to the old fashioned way by making FSA heterogeneous multicore chips. It is in its culture. So they go open and free and ask everyone to come on-board. But don't expect Intel to join in something which acts as a playing-field leveler for the next stage of many-core processor development.
Footnote: One question asked at the Fusion Developers' Summit which appeared to go unanswered, was whether the FSA could eventually see the CPU and GPU cores merge together into a single building block. It is my interpretation that there was nothing in Phil Rogers presentation that either requires this or prohibits this. But of course a unified CPU/GPU ISA would, at least initially, be a novelty rather than standardized block. But, if a unified logic and graphics core is more power and area efficient then the FSA should allow it to prosper.
There is another reason why AMD's move upstream is important.
Software developers want to write once, compile as necessary, run on many platforms and to know that the code they have written can be reused over multiple generations of hardware. The advent of multicore processing risks destroying that situation. The hardware companies have to do something to preserve code portability and that probably means a broad set of standardization issues.
And so AMD is attempting to ease the path of CPU/GPU programming under the Fusion System Architecture intermediate layer, which it calls FSAIL and which it describes as a virtual instruction set architecture. That intermediate layer then calls on the underlying resources, with automated load balancing and dispatch and AMD wants FSA to be able to cope with those resources whether they are x86, ARM, Mali, PowerVR and so on.
But as the eco-system Kool-aid kicked in AMD realized that such a shift in processor ideology is big and more than it could possibly achieve on its own. To make the Fusion System Architecture something worth building it has to offer software developers the broadest possible platform, and guarantees of portability. And who knows more about assembling a broad club of chipmaking partners than ARM.
The question remains – and this Phil Rogers was not laying out for the audience – as to how AMD will make its money out of the Fusion System Architecture.
Is it by being the best implementor of FSA in silicon? This is the historical AMD model in that is has competed with other x86 chip makers. Or is it by licensing the Fusion System Architecture, which is the ARM model? And there is good reason why ARM does not try to compete with its licensees.
Probably AMD will stick to the old fashioned way by making FSA heterogeneous multicore chips. It is in its culture. So they go open and free and ask everyone to come on-board. But don't expect Intel to join in something which acts as a playing-field leveler for the next stage of many-core processor development.
Footnote: One question asked at the Fusion Developers' Summit which appeared to go unanswered, was whether the FSA could eventually see the CPU and GPU cores merge together into a single building block. It is my interpretation that there was nothing in Phil Rogers presentation that either requires this or prohibits this. But of course a unified CPU/GPU ISA would, at least initially, be a novelty rather than standardized block. But, if a unified logic and graphics core is more power and area efficient then the FSA should allow it to prosper.
Related links and articles:
AMD-authored OpenCL textbook due in August
AMD makes Fusion CPU, GPU agnostic
AMD rolls A series x86, graphics CPUs
Analysis: Why ARM-AMD makes sense
Analyst: AMD-ARM deal makes no sense
|
|
|
|
|
Navigate to related information
pchiang_calstan
6/16/2011 1:05 PM EDT
This seems somewhat similar to what Sun did with their open-source T1 processor several years ago. That didn't end up well for them ...
Sign in to Reply
Roba66
10/26/2012 2:13 PM EDT
That was really open source. Unfortunately it was not useful as a piece of IP for existing X86 based systems.
This is different, being actual silicon and cards that can be used, but the drawback is the deep obfuscation layer that AMD marketing thinks is still somehow "open".
Sign in to Reply
peter.clarke
6/17/2011 6:24 AM EDT
Similar perhaps although I didn't see Sun making any transition from one level of abstraction to another.
Sun was a workstation provider that designed processors in order to create a performance advantage. In the end Sun couldn't produce an advantage over Intel and AMD x86 processors and high-end PCs replaced workstations.
In this case AMD and all other processor developers are facing the same difficult transition from multi-core to many-core. AMD seems to be trying to invoke first mover and ecosystem advantage to try and set the agenda. If it doesn't, Intel or ARM will set the agenda anyway.
The big question is does AMD have ARM, Microsoft or anyone else on-board, or are those guys indulging in fence-sitting; getting close to see if AMD's "standard" approach might have legs without committing just yet.
ARM may wish it had thought of this first and be reluctant to join an ecosystem in which it is just another logo on a slide. On the other hand if AMD's approach gains traction better to be in early and hard.
Sign in to Reply
ash9
6/19/2011 7:04 PM EDT
a GPU compute core is the next iteration (ala NVIDIA, but different)into AMD's GPU cards, that compute core will make it into their APU's
Sign in to Reply
peter.clarke
6/20/2011 5:41 AM EDT
@ash9
It sounds plausible but what makes you think that?
And do you think that eventually GPUs will look so much like CPUs that we end up with a single logic-plus-graphics processing unit (L+GPU) to simplify many-core processing?
Sign in to Reply
Roba66
10/26/2012 2:06 PM EDT
AMD GPU technology is not "OPEN" at all, this is a marketing ploy. The support software is closed and proprietary; the architecture is closed and proprietary. If it was even as open as any processor you would have the equivalent of machine level or assembler available and people would be able to write tools for it.
Instead of having the marketers "ease the path of programming" for those poor dumb users, they could give out the information and let the users solve the problem for them. They could try to be more "open" than Nvidia rather than going the other way.
Technology mapping to GPU cores is best done by the people who understand the problem space. Nvidia is working with those people and making good progress; AMD is coming from behind and they need to open up a lot more to potential users to get them interested.
Sign in to Reply