A Moogle  🏳️‍🌈 💻 🎮 🎶 🇸🇪

@therealian Real question, have audio coders and graphics coders ever tried to team up and offload DSP of multiple streams onto the GPU?

It seems like a clear win, there're fullscreen NTSC shaders can grind through an 81-sample-wide FIR in the video domain, audio should be EZ.

I'm not aware of anyone doing it, but I've long thought that it's an obvious application, esp. since DSP paralleizes well.

Newer AMD cards have a dedicated programmable audio DSP, which seemed odd to me since there's an ALU monster already there.

Problem w/ GPUs, AIUI, is latency, since that kills you in audio tasks. Not sure of the details though. But that's from asking GPU peeps.