The feeling is that SATA is still the best connection for video editing. SSD devices are getting larger, more reliable and cheaper, but they really only benefit you for boot-up, and a bit for the programs.
Right now,were I looking at SSD, it would be only for the boot drive w/ the programs, and possibly the Page File. My other HDD's would still be SATA, and probably one of the RAID 0 configurations for my media drives and my Export drives.
Things are changing, and only time will sort it out. When the prices fall to about US$500 for a 650 GB SSD, I might get serious about going in that direction.
Just my opinion,
If a system has a 6 gb/second ssd 3.0 drive as the OS, and a hard disk sata drive (which according to my research goes about 1.7 gb/s)
it seems that you should NOT put page files on the sata drive, but keep them on the much faster SSD drive.
One would be best to experiment and benchmark the placement of the Page File. It will differ by OS, version, and then by the I/O setup.
With 64-bit OS's, and a full compliment of RAM, the Page File will likely be used less, than with a 32-bit OS.
Good luck, and do not hesitate to test. I am a big fan of setting up static Page Files, and then locating it/them per my testing. Each of my 32-bit OS's does better with a slightly different Page File.
how do you recommend testing.
Sorry, by GB , i should have written gbits.
but hear is a real world example
the crucial ssd 300 reads at :Read speeds up to 355MB/s
the sata The VelociRaptor yielded a 105MB/sec average sequential
So the ssd is three times as fast
so keeping everything on the SSD drive makes sense if the other drive is a hard disk drive.
On the other hand, if the PE is using all of the 350 bandwidth, and could use the extra 105 from the sata hard drive, then it would make sense to use it.
Since there has not been a benchmark program, with Assets, done for PrE, I would create a test Project with common Assets, Effects and operations. Choose a combo that will test the full system, and then do a Save_As, so one can go back and test several configurations.
Locate the Page Files in one place, say on the SSD, and run through a lits of tests, like Rendering the Timeline, then Export/Share to something like H.264, or HD WMV, that takes processing power. Time the various operations, and make note of exactly how long each took.
Then, relocate the Page File, and Open the original Save_As, then perform the same operations, making note of the times. Compare those times. Do you see a difference? Which works better for you?
With PrPro CS5, there is a full PrPro Benchmark, the database of results is linked above, so things are a bit easier, and that benchmark program will do the calculating for you.