I have participated in NIPS, CNS, and COSYNE at least a couple times each. In fact, I have participated in all 3 last year.
COSYNE is the smallest conference, but it's growing fast. It's a great conference because it has a good balance between experimentalists and theorists. It takes an extended abstract (2 pages). It emphasizes the systems aspect of the brain.
CNS has more physicist/mathematicians (COSYNE and NIPS also have a high number of them, but CNS has the most), and has a feeling of "European" science. The themes are more theory oriented. I heard that it is usually bigger when it's held in Europe (it alternates).
NIPS started half machine learning, half neuroscience a long time ago, but it has been overrun by ML people. But, in recent years, I see more and more computational neuroscience coming back to NIPS. It has a strict double blind reviewing with high rejection rate around 25% for an 8 page full paper (+1 page reference).
Each conference has its own niche, so the impact would depend on your area within computational neuroscience. NIPS is the only one with a full paper, so it tends to be cited more. However, a lot of networking happens in COSYNE and CNS among neuroscientists, so it would definitely have more impact to experimentalists.
Feel free to ask me more questions.