For example, non-inverting op amp configuration where gain = 1+(R1/R2), how does the size of R1 and R2 matter? For example, for gain = 2: whats the difference between 1+(1k/1k) and 1+(10k/10k) and 1+(100k/100k)... Thank you!

 

Much of the time it will be of no matter as both input and output impeadance are functions of the op amp itself.

I had a case once when I was making a high gain amp, something like an Av of 33 at 3 MHz. For THAT amp I had to use very small resistors, R2 was something like 39 ohms (not K, just ohms) to get any operation at that frequency. Seems there is a small capacitance on each pin, and that forms a break with the feedback resistors, so since the cap is fixed you need to decrease the resistor to increase the break and the overall frequency response.

 

Most opamps have difficulty driving a feedback resistor less than 2k ohms.
With a higher value for the feedback resistor then the load current can be a little more, or the output voltage swing can be higher.

 

High resistors decrease bandwidth because stray feedback resistor capacitance forms a pole that acts like a low pass filter. High resistance causes output offset because of the input current. High resistance causes more noise. Low resistance dissipates more power and robs output drive current capability.

 

High value resistors interact with stray and device capacitance on the inverting input, adding another pole in the feedback loop, which causes high frequency peaking on the output.