G
Guest
Guest
Is it the way of the future? Will Nvidia and ATI follow suit? Or is it just pure coincidence that Sony and Nintendo chose this design philosophy?
It should be the way of the future. Bandwidth (EFFECTIVE, not actually PHYSICAL) has been the constraint of 3D graphics for a very long time.
R300 and NV25 finally seem to have efficient enough bandwidth optimisation that the effective bandwidth is just about high enough to max out the cores... but not quite. (pure core OC still gives perf. boost)
The idea with an embedded buffer is that you can have a gargantuan path to the memory controller - a mind-numbing 2,560-bit pipe in PS2, and an almost as staggering 1,024-bit (512 each on two buffers) in GCN.
By comparison, the newest PC graphics cards have just started to have 256-bit memory buses - 1/10th that of the GS's embedded buffer.
It makes even more sense in a console environment, since the coder can optimise the cache work very closely, and the frame buffer will never exceed any standard TV resolutions (at least until new standards appear
), so a fixed-size cache can be used without too much deficit.
R300 and NV25 finally seem to have efficient enough bandwidth optimisation that the effective bandwidth is just about high enough to max out the cores... but not quite. (pure core OC still gives perf. boost)
The idea with an embedded buffer is that you can have a gargantuan path to the memory controller - a mind-numbing 2,560-bit pipe in PS2, and an almost as staggering 1,024-bit (512 each on two buffers) in GCN.
By comparison, the newest PC graphics cards have just started to have 256-bit memory buses - 1/10th that of the GS's embedded buffer.
It makes even more sense in a console environment, since the coder can optimise the cache work very closely, and the frame buffer will never exceed any standard TV resolutions (at least until new standards appear
), so a fixed-size cache can be used without too much deficit. Embedded RAM is a trade off like anything else.
It can basically increase frame/ZBuffer bandwidth, at the cost of silicon real estate and clock speed.
Obviously Sony and Nintendo thought that the tradeoff was worthwhile, NVidia didn't.
IMO -- Sony and to a lesser extent Nintendo sacrificed too much in their respective feature sets for the benefits of the embedded memory. However this is just one opinion.
At some point embedded memory will most likely be ubiquitous on graphics processors, when the tradeoffs make that both practical and desireable is open to debate.
It can basically increase frame/ZBuffer bandwidth, at the cost of silicon real estate and clock speed.
Obviously Sony and Nintendo thought that the tradeoff was worthwhile, NVidia didn't.
IMO -- Sony and to a lesser extent Nintendo sacrificed too much in their respective feature sets for the benefits of the embedded memory. However this is just one opinion.
At some point embedded memory will most likely be ubiquitous on graphics processors, when the tradeoffs make that both practical and desireable is open to debate.
Similar to the PowerVR approach, put frequently accessed (read/write) data in a buffer with fast access & high bandwidth.
PowerVR uses low cost TILE approach, Nintendo's Flipper as the whole buffer in, BitBoys have a whole buffer inside too...
(PowerVR has plenty of other reasons, I'm not sure exactly what was the main reason of the tile based approach...)
PowerVR uses low cost TILE approach, Nintendo's Flipper as the whole buffer in, BitBoys have a whole buffer inside too...
(PowerVR has plenty of other reasons, I'm not sure exactly what was the main reason of the tile based approach...)
G
Guest
Guest
I think Sony have made the EE and GS mucher smaller now and plan to make them smaller(60nm)!
Will that be cool?
Will that be cool?