I use EA 2.5, a little older but still quite good. I calibrated mine to the dyno tests that I've made, and the major difference is that I set the ambient air pressure to 26.5 instead of 29.92, otherwise the desktop dyno is a little optimistic! It also shows a peakier torque curve than what I get on the real dyno, so I design my electronic emulations to have a higher rpm torque peak than I really am expecting in the real engine and mentally fill in the low points.

Where it is really good is showing the effects of tuning the intake and exhaust stack lengths, an area that is overlooked by a lot of people.