What happens on ARM when double to int64_t conversion overflows?

258 Views Asked by At

With x86 you get the magic value 1 << 63 (MSB set, rest clear, i.e. the most-negative number, which Intel calls the "integer indefinite" value) when you convert a double to an int64_t with the x87 FPU, SSE or AVX, when the result of rounding to an integer isn't in range. In addition to that, the invalid operation FP exception is raised.

Is there a similar behaviour with ARM? That would be nice because I need portability only to x86 and ARM.

0

There are 0 best solutions below