Not that it's any of my business, but I've found buying for the future is never very cost effective. Today's high-end technology is likely to cost half as much a year from now.
I'm a big fan of only upgrading when I'm falling behind -- and then not buying the top of the line. Otherwise, it's too heartbreaking to see the system I spent $1200 bucks on last year now selling for less than $500.
> "What would be the best for rendering - an AMD 16 core or a very fast i7?"
I believe that Intel outperforms AMD by a wide margin. In looking at the PassMark chart that Steve pointed to, my 6 core Intel Core i7-3930K @ 3.20GHz has a PassMark score of 12,147 while the AMD FX-9590 Eight-Core has a score of only 10,426. So Intel is faster with 2 less cores. Also there are no AMD processors with a faster score. The entire upper 1/3 of the chart is all Intel. I wouldn't even consider AMD a contender on the high end.
Get yourself an Intel Core i7.
As for buying for the future, I agree with Steve. I made that mistake once and bought the latest Intel Core 2 Quad Extreme for $1,000 (that's just the processor) thinking this will last me for 5 years. Guess what? I was wrong. After 5 years it was incredibly slow. I should have spent $500 on a regular Core 2 Quad processor and then 2.5 years later spent another $500 on a Core i7 and in that 5 years I would have spent the same $1000 but had a much faster processor in the end. I will never buy the "fastest" processor again.
Guy i used to work with would assemble a hot new gaming machine about every 18 months, full works with water cooling. He'd typically spend about $4500 to $5000 and then brag about how fast his system was. I'd wait 8 or 10 months and then buy the motherboard & cpu being obsoleted and get a faster system than his for $600 using standard air cooling. Then 8 or 10 months later he'd go spend another $4500 again ...
This is a really good question and a really hard one to answer simply. The real question here is which is faster for rendering. The PassMark scores are interesting, but do not necessarily translate to shorter rendering time in Movie Studio. This has a lot to do with how well the software takes advantage of multiple threads (multiple cores) and other CPU optimizations.
To make things worse, the answer might be different for each rendering codec. Some codec might be more optimized for a particular CPU architecture.
So the question remains: for rendering in Movie Studio, what is best? More cores? Or a smaller number of faster cores?
A bonus question: Does this change for VMS 11, 12 and 13? Is the response any different for the 32bit or 64bit version?
Few years ago I had spent some time benchmarking rendering time for the same clip on different machines, testing both CPU and GPU rendering combinations. The results below were for rendering 1080-30p AVC. As you can see the i5 performed better despite a slower clock and less cores. The Xeon machine was disappointing, 8 cores didn't translate into much faster rendering.
(note, these are physical cores, all systems had hyperthreading enabled)
And, cores are not necessarily the thing when it comes to rendering. Although an application (Vegas) is capable of using several cores, the internal filters and codecs are still tied to antiquated architecture, in many cases. So a particular point in the rendering chain may (and often does) throttle the whole thing back to one-core performance.