My assumption would be that the assumption was that 64-bit arithmetic is somehow more time efficient and 32-bit is more memory efficient while still large enoug

My assumption would be that the assumption was that 64-bit arithmetic is somehow more time efficient and 32-bit is more memory efficient while still large enough to hold the results.