Winspice LM317 not as expected

[c5f8]

I use WinSpice and have a question about the LM317. The National Semiconductor data sheet states that Vout = 1.25V (1 + R2/R1) + Iadj * R2. I used two different LM317 netlists from this forum and I don't understand my results.

When R1=1K, R2=0, then vout=1.313708V (close to the expected 1.25V)
When R1=1K, R2=10K, then Vout=11.072751V (close to the expected 11.5V)

Anyone know why vout a little higher in the 1st case and a little lower in the 2nd case?

A second question is using capacitors across the incoming power and output leads as opposed to without. Without caps, the voltage has no ringing but it takes about 1 mili-second to approach vout. Without caps, the voltage has 10ths of a micro-volt ringing and it only takes about 0.1 mili-second to approach vout. I'm using the circuit for fan regulators in my computer case. Should I leave out the caps?

A third question is what would cause an underflow error in myftoi() for both case of R2?