dimag0g
Very Active Member
Can you tell me which FPU that was? 32*0.375 should be exactly 12 on any FPU that supports ANSI floats, rounding rules are part of the standard (let alone there is no rounding in your example). That FPU can't even be used by a C/C++ compiler, so I'm really curious here.I had something in my code which did something like '32 * 0.375' which is 12, only the FPU I was doing it on decided the result was actually 11.9999999
More to the topic, do you guys use -mfpu=vfp or -mfpu=neon? I know it's silly, but might this be the problem?
Last edited by a moderator: