I will try to answer your question at least in parts by just reduce it to some smaler equations I am able to handle (as neither astrophysics nor nuclear/quantumphysics are my speciality). Same reason is why I will not check the numbers you gave in your question. (Edit: I exactly did that now, all numbers you gave and I used I checked for mistakes. Found none. ^^)
As the ability of a Matrioshka brain using the power of our sun is given as 10^9 to 10^12 minds, I will calculate its computing power in floating operations per second (flops). I think a 'mind' is used as equivalent for one human brain. The computing power of a human brain is roundabout 10^13 flops so the Matrioshka brain has a computing power of 10^22 to 10^25 flops. Lets use the higher value here. As you want to simulate the universe in steps of femtoseconds (regarding the comments) that equals 10^10 flopfs (floating operations per femtosecond) for a MB using the sun.
The power of all stars in our galaxy is 10^32 times that of our sun, so a network of MBs using our whole galaxy is 10^42 flopfs. As we have 10^80 particles we would need 10^38 of these galaxy-MBs just to check for every timeframe if a particle interacts with any other particle or not. And that's just for the check. We still have to check what kind of interaction those particles have.
Hubble revealed an estimated number of 10^11 galaxies in his visible range. We would need 10^27 times that. Just for the checking. (Edit: I just read scientific estimations go on to 10^12 galaxies in the observable universe, so we 'just' would need 10^26 times the output of our observable universe just for checking all particles in itself for interactions. Conclusion stays the same.)
So my answer is: no. We will not be able to simulate a whole universe even with this kind of technology.
