Andrew Milson

Thanks @Pratyush. Yeah keen to take a stab!

@Pratyush unless I'm missing something I think it might be easier going straight to the second approach @ValarDragon mentioned. With the first approach It's easy to obtain the values for...

With the first approach I mean additional memory on top of resizing the coefficients to the size of the domain. The second approach has better locality and no need to...

Are these changes still wanted? If so I'm happy to submit a pull request for this

I'm also having issues with FixedBufferAllocator at comptime (zig 0.11.0-dev.2165+8f481dfc3). There's an additional issue when the types don't share the same alignment as u8. to reproduce: ``` test "comptime alloc"...

cc @mlugg

Thanks Wei. Online I read OpenCL was ~40% slower than metal and ~30% slower than CUDA. I'd be really curious what kinds of performance an OpenCL implementation of miniSTARK would...

Hahah the CUDA legend himself! This is super helpful. Thanks mate

@beauwilliams thanks for the trace. What Mac do you have? For now the `--feature gpu` is only supported on M1 Macs. In theory the gpu stuff should work on older...

Thanks Pratyush. Yeah I'm keen to take a look