Difference between the carry flag and overflow flag

[Fatima Hasan]

Homework Statement

Why does the carry flag not work for signed number addition and the overflow does not work for unsigned numbers?

Relevant Equations

-

Here's my explanation with an example:
(d: decimal )
7d + 2d = 9 d
If signed:
the result indicates an overflow, because it exceeds the maximum which is +7.
( 0111 + 0010 = 1001 )
However, CF = 0 indicating that there is no overflow which is wrong.
OF = 0 ⊕ 1 = 1 -> invalid result, because the result exceeds the maximum.
If unsigned:
the result is valid because min= -8< 9 < max= 7
CF = 0
OF = 0 ⊕ 1 = 1 -> indicates that the result is invalid which is incorrect.
Is there any further explanations ?

 

[jedishrfu]

Here's a discussion of the two flags:

http://teaching.idallen.com/dat2343/10f/notes/040_overflow.txt

or this video with a funny computer voice:

 

[Mark44]

Homework Statement:: Why does the carry flag not work for signed number addition and the overflow does not work for unsigned numbers?
Relevant Equations:: -

Here's my explanation with an example:
(d: decimal )
7d + 2d = 9 d
If signed:
the result indicates an overflow, because it exceeds the maximum which is +7.
( 0111 + 0010 = 1001 )
However, CF = 0 indicating that there is no overflow which is wrong.
OF = 0 ⊕ 1 = 1 -> invalid result, because the result exceeds the maximum.
If unsigned:
the result is valid because min= -8< 9 < max= 7
CF = 0
OF = 0 ⊕ 1 = 1 -> indicates that the result is invalid which is incorrect.
Is there any further explanations ?

It looks like your examples assume 4-bit numbers, and either signed or unsigned addition. As a 4-bit signed number, the possible values range from -8 to + 7, inclusive. As a 4-bit unsigned number, the possible values range from 0 to 15, inclusive.
Here's what the Intel documentation says about CF and OF, which may or may not have some bearing on your question.

Carry flag — Set if an arithmetic operation generates a carry or a borrow out of the most significant bit of the result; cleared otherwise. This flag indicates an overflow condition for
unsigned-integer arithmetic.

Overflow flag — Set if the integer result is too large a positive number or too small a negative
number (excluding the sign-bit) to fit in the destination operand; cleared otherwise. This flag
indicates an overflow condition for signed-integer (two’s complement) arithmetic.

The status flags allow a single arithmetic operation to produce results for three different data types: unsigned integers, signed integers, and BCD integers. If the result of an arithmetic operation is treated as an unsigned integer, the CF flag indicates an out-of-range condition (carry or a borrow); if treated as a signed integer (two’s complement number), the OF flag indicates a carry or borrow;

In your case, you're working with a hypothetical processor with 4-bit registers, together with signed or unsigned arithmetic.
If the operation is unsigned addition, and the numbers being added are 7 and 9, the result is too large to fit in 4 bits, so the CF flag would be set. It's true that the result overflows the capacity of a 4-bit register, but the result is also producing a carry at bit 4 (with bits numbered 0 through 3 in a 4-bit number). For unsigned subtraction, 9 - 7 wouldn't cause a problem, but subtracting 9 from 7 would cause CF to be set, because there would need to be a borrow from bit 4.

If the operation is signed addition, if the addition of two positive numbers produces a negative result, OF is set (e.g. 7 + 1 produces a bit pattern that represents a negative number). Also, if the addition of two negative numbers produces a positive result, OF is set in this case, as well (e.g., -3 + -6 produces a bit pattern that represents a positive number).