A whole separate set of registers
Integers ride in r3, r4, … — but floating-point values get their own bank, f1 through f13. The first float argument arrives in f1, the second in f2, and the result comes back in f1 (the way r3 works for ints). Further float arguments continue in f3, f4, … just as integer arguments march up r5, r6 — so a three-float function takes its third argument in f3.
A float is single-precision (32-bit), and PowerPC has single-precision arithmetic that ends in s. Addition is fadds:
fadds f1, f1, f2 # f1 = a + b, single precision
blr
Same three-operand shape as add: destination first, then the two sources.
Your task
Write add_f, returning the sum of two f32s.