Clever Geek Handbook

How to make signed 64 bit integer using one signed integer and one unsigned integer in 32 bit cpu

On a 32 bit processor with a 32 bit compiler, I want to make a 64 bit signed integer using one of each signed and unsigned integer. Without using any predefined macros or types.

+3


source to share


1 answer


32 bit compilers will handle 64 bit numbers for you. Therefore, it is unlikely that you really need it. But I will bite. At first glance, this is a fairly simple problem.

#include <stdint.h>

static inline int64_t make_int64(uint32_t low, int32_t high) {
    return (int64_t)((uint64_t)high << 32) | low;
}

static inline void split_int64(int64_t value, uint32_t *low, int32_t *high) {
    *low = value & 0xFFFFFFFF;
    *high = (int32_t)((uint32_t)(value >> 32));
}


But its always tricky / dangerous mixing signed and unsigned integers. Manually constructing int also requires you to know how the processor formats them. We'll assume that his 2 compliments are few in number.

It would be helpful if you could provide a complete description of your requirements. For example, the above example make_int64 (0, -1) = -4294967296, but make_int64 (1, -1) = -4294967295.

0


source






More articles:


All Articles