An FYI / Correction to SandMan -
Maybe for pure concrete models, ETAB's may have an advantage over RISA for dynamic solution. But, it should be much more modest than what it used to be.
If you're talking about true eigen modes, then RISA cannot take more modes to get to 90% mass participation unless you have modeled your mass differently. Hence, why there may still be some different with concrete building using a semi-rigid floor slab. But, for modeling of steel building with a rigid floor diaphragm, the mass modeling should be virtually identical between the programs.
Now, ETABs used to have an advantage because of their "Ritz Vector" solver for dynamics. This type of solver allows you to use fewer "modes" because you're biasing the solution to specifically exclude modes that do not participate in the direction requested. What you get are not the true dynamic modes of the structure, though they are usually close.
However, RISA added a Ritz Vector dynamic solution option a year or so ago. So, while our Ritz Vector algorithm won't be exactly the same as CSI's / ETABs, it should be in the same ball park. Thereby eliminating most reasons why ETABs might require significantly fewer modes than RISA.