Simple perf testing for SIMD delayed operations

Initial results

$ ./build/testing -t 1       # scalar
Access time: 242, yielding 7.27553e+06

$ ./build/testing -t 1 -s    # AVX2
Access time: 192, yielding 7.27553e+06

$ ./build/testing -t 4       # scalar
Access time: 74, yielding 7.27553e+06

$ ./build/testing -t 4 -s    # AVX2
Access time: 62, yielding 7.27553e+06

What about floats?

$ ./build/testing -t 1 -f    # scalar
Access time: 203, yielding -556276

$ ./build/testing -t 1 -s -f # AVX2
Access time: 137, yielding -556276

$ ./build/testing -t 4 -f    # scalar
Access time: 55, yielding -556271

$ ./build/testing -t 4 -s -f # AVX2
Access time: 40, yielding -556271

Build instructions

Just use the usual CMake process:

cmake -S . -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build