Perhaps this is a question for INTEL or likewise but i figured someone would know.
I nag about using apropriate type for integer (and other). Don’t use INT when TINYINT is enaugh. That way storage space is saved.
But last week someone mentioned that what you win in storage is lost in execution.
INT require 4 bytes and SMALLINT 2 bytes. The theory was that it’s faster to calculate with 4 bytes than with 1, 2 or 3 bytes and that the processor (?) anyway padds (or truncate) the value to 4 bytes.
Something to do with 32-bit processor and that 32/2=> 16/2=8
Anyone who can explain this?