Also, if you are saving the result to a floating point register then enter your constants as floats. N7:0|9000.0*100.0.
It looks like you are performing a percentage calculation, so if you must use integer math and N7:0 < 9000 then N7:0 | 9000 will be less than 1 - which will round to just 0 or 1 with integer math, giving you only 100 or 0 as your answer (This is what Robert was refering to). In this case you should re-arrange the equation to multiply by 100 first, then divide by 9000, or enter the constants as floats (ie 9000.0 and 100.0) even if you are storing the result in an integer.
Last, a word about constants and the CPT instruction. The CPT instruction takes longer to execute than single operation math instructions, but it is faster for longer computations because it doesn't have to take time to store and fetch intermediate results from memory. Where you are using constants, combining the two constants and using a single MUL or DIV instruction will be more efficient, ie N7:0 * 1.11111111e-2, or N7:0 | 90