On a 32-bit MIPS machine, you need to execute the following single-precision floating point arithmetic instructions:

1.) 3.5e25 * 3.5e25 / 12.0e20

2.) 13.2e-30 * 3.5e25

3.) 13.2e-20 / 3.5e25 + 5.7e8

Do any of the above operations overflow or underflow? If so, can they be rewriten so they work?

Note: all of the above expressions are in base 10, the notaton eNNN means 10 raised to the NNN power.