Any 3d toolpath will require more time than 2d due to the increased z movement and shear volume of lines in the tap file.
Time saving options are, as ger21 pointed out, skip roughing [I often do this but not recommended for deeper cuts in hard woods]. What specifically are "deep cuts" and "hard wood" ? Well, you learn that from experience and breaking bits. Tapered bits are tough but of course have their limits, skipping roughing carries risks.
Sometimes I compromise and only do, say, 2 roughing passes instead of the calculated 5. Can also calculate toolpath on a 45 angle which *can* greatly reduce the risk of stress and breakage [this only works with an uninterrupted toolpath from top to bottom, often 3d toolpath will jump all over the place then this won't work].
A risk free method to reduce cutting time is increasing line stepover.
For the finest detail you'd want 8 to 13% for a finish pass but you can often jack it up to, say, 15 to 18% without a perceptible difference the naked eye can see.
Go much further than that though, and you will notice a quality difference that gets worse and worse as the % increases.
Sometimes thats fine if you are only doing a rough ball shape or something but if you are doing the last supper, you want max detail regardless of time.
Finally, having quality bit makes a huge difference as to your final product and longevity of the tool.
I tinker with cheap made in china bits occasionally, by and large though I invest in quality tools for predictable, repeatable and quality results.http://www.precisebits.com/products/car ... 250b4f.asp