sub $0x10,%rsp allocates space on the stack, it’s the same as doing pushq 0 twice except that the value in the allocated stack space will not necessarily be zero, but this is ok since the mov instructions used later replace the garbage with real data. As Michael rightly stated the mov $0x0,%eax is because printf … Read more
Your unroll doesn’t help with the FP latency bottleneck: sum + x + y + z without -ffast-math is the same order of operations as sum += x; sum += y; … so you haven’t done anything about the single dependency chain running through all the + operations. Loop overhead (or front-end throughput) is not … Read more
Yes, for gcc5.x and later specifically, that specific expression is optimized very early to just p, even with optimization disabled, regardless of any possible runtime UB. This happens even with a static array and compile-time constant size. gcc -fsanitize=undefined doesn’t insert any instrumentation to look for it either. Also no warnings at -Wall -Wextra -Wpedantic … Read more