Hi All,
I did
1. opamp open loop simulation, folded cascode with and w/o gain boost
2. put the opamp into multiply-by-two configuration, do pss+pstb analysis, check the gain and unity freq.
3.in the open loop bode plot, DC gain of opamp with and w/o gain boosting is 66.7dB and 44.25dB, respectively. GBW is more than 500MHz and 700MHz, respectively.
4. in the SC configuration, the loop gain drops to 42.45dB and 33.45dB for with and w/o gain boosting, respectively, at 50MHz sampling rate, the GBW drops to ~56MHz and ~45MHz, respectively. seems has two poles there so that the slope falls faster.

I don't know how to add multiple pictures, i will add pictures separately.

AC sim schematic, CL=1.5p

folded cascode gain-boosted opamp schematic, 3mA current excluding gain-boosting opamps and bias generator.  
two CMFB shown, one is ideal CMFB used for AC (continuous time) simulation.

I found a problem that after i added a pair of switch (yellow circle as i indicated), the result seems good now, enough GBW but smaller PM. as shown in the picture below.
i didn't add that pair of switch is because I intend to do opamp sharing  in the pipeline ADC so that pair of switches will not be there in that case.
even without that switch, the input common mode voltage can still settles to around half vdd, but why it affect the speed that much? that means, when at sampling mode, the opamp has to be in unity feedback configuration?
any comments?