Break the mask out into binary first:
0000 1111 1111 1111 (I am going to assume the "d" was a typo, let me know if this is not the case)
Now look at the instruction like this:
Code:
Source: 0000 0000 0000 000[COLOR="Red"]0[/COLOR]
Mask: 0000 1111 1111 1111
Compare: 1111 1111 0111 110[COLOR="red"]1[/COLOR]
Now take each column individually, and if there is a 1 in the mask location, the source and the compare must be equal in order for the MEQ instruction to evaluate true. If the mask position is zero, then the MEQ "doesn't care" about that position.
In your example, right off the bat, we see that the least significant bits are not equal and they are "masked in" to the instruction, so you know it evaluates false without having to go any further.
Assuming the "d" in the mask was not a typo, but leaving out the leading zero:
Code:
Source: 0000 0000 0000 000[COLOR="Red"]0[/COLOR]
Mask: 1111 1111 1111 1101
Compare: 1111 1111 0111 110[COLOR="red"]1[/COLOR]
An instructor once explained it this way: If there's a big fat zero in the mask, the values from the source can't fall through to be compared, so ignore those. if there's a skinny little "1" in the mask, then the number in the source can "fall through" to be compared.