4 ms·
I'd be really curious to see how it optimizes for ARM. x86 is great for kitchen-sink optimizations like this, but with ARM you have a much smaller extension set
by smoldesu 4y ago
I'd be really curious to see how it optimizes for ARM. x86 is great for kitchen-sink optimizations like this, but with ARM you have a much smaller extension set to target. Hell, you'd probably be stuck with NEON unless you figured out a way to GPU-accelerate the process.