How to convert 2 signed 32 bit numbers to 64 bit in C

Question

a and b are upper half and lower half of the same number. Now I have to save this number to a 64 bit register. lets say a = -1(Higher bytes) and b = -50. How can I do this ?

I am using the following which works for positive numbers.

int64_t c = (a);
c = (c<<32);
c+=b;

The above does not work for -ve numbers. How to do this ?

Edit: The above code for -ve numbers give a very large value for the -50. Basically what this means is that after the operation the "c" should have the value of -50, but it should be 64 bits. As upper half i.e. "a" is -1 and acts as the signed bit. The lower half has a -ve sign which makes it value very large due to the shifting operation. I hope this is a little more clear.

In what way does it "not work"? Can you express mathematically what function you're hoping to accomplish? Or express it in terms of bit strings? — user1084944
– user1084944, Commented Mar 7, 2015 at 5:01
I repeat my previous comment. :P Are you trying to make c be the integer whose (two's complement) binary representation is the concatenation of the binary representations of a and b (both extended to be 32 binary digits long)? — user1084944
– user1084944, Commented Mar 7, 2015 at 5:13
did you mean c |= b on the last line? Better question: what do you expect the answer to be if a = -1 and b = 50? — clemej
– clemej, Commented Mar 7, 2015 at 5:21
@clemej: Yes, that is his problem. When he does addition, the 32nd bit will be zero and so he'll get a large negative number. — jschultz410
– jschultz410, Commented Mar 7, 2015 at 5:24

clemej · Accepted Answer · 2015-03-07 05:34:24Z

2

Assuming you mean to do straight bitwise replacement without sign extension, then I think you want to replace

c += b;

in your code with

c |= (uint32_t)b;

For a = -1 and b = -50,this will make c = 0xffff.ffff.ffff.ffce

For a = -1 and b = 50, this will make c = 0xffff.ffff.0000.0032

For a = 1 and b = -50, this will make c = 0x0000.0001.ffff.ffce

For a = 1 and b = 50, this will make c = 0x0000.0001.0000.0032

answered Mar 7, 2015 at 5:34

clemej

2,5732 gold badges21 silver badges28 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

clemej Over a year ago

Hat tip to @Hurkyl 's comment above that reminded me you may need a explicit cast to 32 bit to keep a negative 32-bit value from being sign extended to 64-bits before the bitwise or.

Akash Deshpande Over a year ago

Thanks you so much this was perfect. I know my question was very very confusing. But your answer was spot on. I owe my hours and hours of time saved to you !!!!!

M.M Over a year ago

note: += vs. |= makes no difference if the lower 32 bits of c are 0 and we are using 2's complement (as OP's code relies on anyway)

Collectives™ on Stack Overflow

How to convert 2 signed 32 bit numbers to 64 bit in C

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related