This depends entirely on what you’re measuring: if my goal is to make my work faster and a compiler can generate something which runs faster, most people don’t feel that it’s not fair to use extra CPU features - they just want to compare the results which they’ll actually get.
I remember people whining in the 90s x86/Alpha/POWER comparisons that some compilers were aware of fused Multiply-Add operations, too, but everyone who was trying to make a purchasing decision just tuned them out since they wanted to know realistic FLOPS/$ ratios.
I remember people whining in the 90s x86/Alpha/POWER comparisons that some compilers were aware of fused Multiply-Add operations, too, but everyone who was trying to make a purchasing decision just tuned them out since they wanted to know realistic FLOPS/$ ratios.