Feel free to call me an idiot for asking. I have little if any understanding of these things but if I remember correctly, new technologies were said to allow the gpu to do more of the cpu's work than ever before (dx 11, sm5.0, opencl). Is any of this being done or anything similar?
Dolphin, the GameCube and Wii emulator - Forums
›
Dolphin Emulator Discussion and Support
›
Support
Is there a possibility that more of dolphins code could be pushed from the cpu to gpu
Is there a possibility that more of dolphins code could be pushed from the cpu to gpu
Thread Rating:
|
Is there a possibility that more of dolphins code could be pushed from the cpu to gpu
|
|
01-08-2012, 04:27 AM
(This post was last modified: 01-08-2012, 04:31 AM by NaturalViolence.)
Short answer: No
Longer answer: A gpu is only faster than a cpu at running extremely parallel vector operations. Very few applications involve this type of logic (3D rendering, physics simulations, AI simulations, database/accounting applications, decryption, other scientific applications, etc.). The GC/Wii CPU and DSP emulation is neither highly parallel nor highly vectorized and even if they could port most of the code to opencl/directcompute/cuda (they can't due to numerous limitations present in all three frameworks) it would likely run MUCH slower. GPU emulation on the other hand is highly parallel and involves lots of vector code however dolphin already emulates the GC/Wii gpu using shaders which run on your gpu.
"Normally if given a choice between doing something and nothing, I’d choose to do nothing. But I would do something if it helps someone else do nothing. I’d work all night if it meant nothing got done."
-Ron Swanson "I shall be a good politician, even if it kills me. Or if it kills anyone else for that matter. " -Mark Antony |
|
« Next Oldest | Next Newest »
|
Users browsing this thread: 1 Guest(s)