UMPS (Universal MicroProcessor Simulator) by VM Design is pretty good. The best thing is it supports, simulates and assembles for a variety of uCs other than 8051 (AVR, PIC, COP, Motorola etc). It also provides a library of "virtual circuitry" (LCDs, LEDs etc)

Check out this link:
http://www.vmdesign.com/html/p02.htm

You can download the evaluation version for free from some site (I forgot which) In case you have any problems downloading it from the VM Design website (which was where I had downloaded it initially) a google on "UMPS download" will return several mirror sites.

The evaluation version has a code limit (I think 2K) and a 90 day trial period, but for some reason the trial period limitation is not very effective ---- I will say no more on this issue!! :-D Anyway, I got past the code limit problem by simulating each of my subroutines (or set of related subroutines) individually. Once I got my sub-routines working well, it was easy getting them to work together.

kundi