Forum Discussion
Altera_Forum
Honored Contributor
16 years agoI would use the latest. What they probably mean is most of the testing is done for the newer chips, and new algorithms are probably not put in or tested for older devices. At the end of the day, I've never heard of any large disadvantage going to a newer version, but have seen many disadvantages sticking with an old version.
One thing that does occur, is that each release has a lot of changes. Ideally they make the fitter better, but they might fix a bug or something like that. Altera will run a suite of designs for each version, and hopefully the average result gets better, and at worst stays the same. If things get worse they try to investigate why and weigh the pros/cons of whether to go with whatever hurts, assuming it can't be fixed. Anyway, the point is that it's always an average, and that means many designs get faster and those people never complain and assume that's just natural progress, while some designs get slower. There's always a slew of complaints like this. It would be nice if every design got faster with every release, but probably just not possible. So you might hear people say a new version is doing a worse job, but quite often that's because it's a single data point.