4 Minute Chia 32k Phase 1 Plotting

I have been working on a RAM-only optimized k32 plotter that is lock-free and scales linearly with each thread. It is extremely cache friendly, and as you can see from the htop output, threads spend barely any time in kernel mode. There is a bit more room for improvement. Moving on to Phase 3 now with an optimization plan already in hand. The video shown is running on an 64-thread ARM CPU. Fx calculation is not using any SIMD. I did not see any significant gain (sometimes regression) on the NEON versions of blake3 for the way it is used in the plotting mechanism.