• by talldayo on 1/22/2025, 6:24:22 PM

    This is cool, but something tells me the t/sec rate compared to native is still wildly off-base. How much less optimized is WebGPU compared to the current OpenCL/CUDA implementations?