http://www.gamechup.com/ps4-will-have-many-launch-titles-trying-to-replicate-ps3-spu-runtime-system-cerny/
He also revealed that they are trying to replicate the SPU Runtime System of the PS3. If you remember the SPUs were able to do little tasks very efficiently and at lightning speed, reducing the load off the GPU. Since the PS4 doesnt have the Cell processor, there needs to be a different way achieve the same effect. Cerny explains how they are trying to accomplish this. Were trying to replicate the SPU Runtime System (SPURS) of the PS3 by heavily customizing the cache and bus, he said. SPURS is designed to virtualize and independently manage SPU resources. For the PS4 hardware, the GPU can also be used in an analogous manner as x86-64 to use resources at various levels. This idea has 8 pipes and each pipe(?) has 8 computation queues. Each queue can execute things such as physics computation middle ware, and other proprietorially designed workflows. This, while simultaneously handling graphics processing. He also mentioned that developers probably wont be using this functionality early on in the PS4s lifecycle but later on this could become an extremely important function. The PS4 will be launched this holiday season and more information will be released at E3. Most developers reacted positively at PS4 specs, and majority of them praised the 8GB GDDR5 RAM of the system.
so less stress on gpu more grahics more Gflops?
Whiteknight19
It basically says, AMD GCN can be use like multi-core X86-64 CPU (AMD GCN itself includes AMD's X86 IP).
AMD designed it's own Intel Larrabee type solution without completing the full X86 compatibility i.e. there's no need to waste transistors on full X86 legacy.
Statements from AMD
http://www.tomshardware.com/reviews/fusion-hsa-opencl-history,3262-6.html
"That effort resulted in a couple of things," adds Macri. "One, we ended up with the best architecture out there thats unifying scalar and vector compute. It blows away what [Intel] did with Larrabee. The Nvidia guys have only attacked part of the problem, because they only have the IP portfolio to attack part of the problem. What theyve done isnt bad. Its actually good for having one hand tied behind their back. But with [Fusion], we had the full IP capability, and it truly is the first unified architecture, top to bottom."
http://www.tomshardware.com/reviews/fusion-hsa-opencl-history,3262-8.html
In addition to unified memory, AMD notes that HSA establishes cache coherency between the CPU and GPU, eliminating the need to do a DMA flush every time the programmer wants to move data between the CPU and GPU. The GPU is also now allowed to reference pageable memory, so the entire virtual memory space is available. Not least of all, HSA adds context switching, enabling quality of service. With these features in hardware, an HSA platform becomes very similar in programming style to that of a CPU.
Log in to comment