Quiz submission record for quiz3-2-1 at Wed Jul  7 10:26:56 2004:

Your Answer for Question 1:
If one could get a hold of the return address register $ra -one might be able to take over a program and even perhaps a computer. 

Your Answer for Question 2:
A subnormal is a floating point number that has the same exponenet as 0 but a nonzero significand. They allow a number to degrade in significance until it becomes 0, called gradual underflow. Rather than having a gap between 0 and the smallest normalized number, IEEE allows denormalized numbers. 
If our programs do not expect subnormals, it could cause an exception. 

Your Answer for Question 3:
The order of floating point instructions is important because depending on the value of x we could get different answers for the above. If x were a small number, 1.0 - x would result in 1.0. However, that might not be the case when doing (0.5 - x) + 0.5. 

Your Answer for Question 4:
The IEEE 754 currently has a gap between 0 and the smallest normalized number and subnormals tend not to be a good solution for them. It would be usefl if the next generation of IEEE 754 math co-processors included better ways to deal with those. The accuracy and range of numbers provided by using single floating point and double floating point is very satisfying and is an asset for the current standard. 


Your unique submission ID is quiz3-2-1-cs61c-cf-1089221216-3075.