HI,I've inherited (the original author left the company) complex C code, to be compiled with Keil C V5, -o1, for a STM32F051R8. The problem: during Interrupts are FP calculations made. Timing dependent (very stringent) sometimes these FP calculations give wrong values, about a factor of 100. If I do dummy FP calculations in a routine which is called by the Interrupt routine - the FP values are of type 'volatile float' therefore the code is generated and not optimized away - and don't use these results, the above mentioned FP errors vanish.Has anybody an answer for such behavior? I don't like to include that 'fix' when I don't know how it works and cures the 'fix' the behavior for all timings?Any answer appreciated, HansPS: That was the reason for my last question about Interrupt enable/disable.