Using SiSoftware Sandra as a benchmark, going from ATA100 to ATA133 results in an improvement of about 15%, IIRC. Going from a 2MB cache to an 8MB cache has a similar effect. The point I am making is that cache performance is a significant factor in overall HDD performance, and therefore interface speed is also significant.
The reason that interface speed matters is because of the hard drive's cache. Every hard drive keeps an onboard cache. When a file is accessed on the drive, the cache loads a whole block of 2 or 8 MB (depending on the drive). Then when the O/S requests the next part of the file or another file that immediately follows the first, the HDD already has that data in cache memory, and can serve it back to the system at the full transfer rate of the interface.
A hard drive's cache gets hit pretty often. That's why you see an increase when going from a standard WD drive (2MB cache) to a special edition (8MB cache). That's also why you see an increase when going from a WD (ATA100) to a Maxtor (ATA133). And that's why you will see an increase when going from ATA100 to SATA150, <b>regardless of spindle speed or sustained media transfer rates</b>.
Please understand, I'm not saying you will double your performance. We are talking 5, 10, 15 percent here.
You are saying that SATA gives <b>no</b> improvement over PATA, and I am saying that is absolutely false. There is <b>some</b> improvement, and it is up to the consumer to decide whether that amount improvement is worth the additional cost. For many it is not.
Personally, I wouldn't bother with SATA unless it were a 10,000rpm drive, but that's not because there is no gain in a 7200rpm SATA drive. It's because the gain of a 7200rpm SATA drive is not worth the money. I wouldn't go with ATA133 for the same reason: not because it isn't better, but because it isn't worth the extra cost.
A lot of the numbers are vague to me, because I didn't keep a record of my tests. It would be fun to do a benchmarking roundup of drives with various interfaces and spindle speeds, just to lay the argument to rest. What benchmarking tools have you used to determine a drive's performance?
____
P4C800-E Dlx, P4 3.0 @ 3.5MHz, 1024 Corsair @ 5550MB/s, 72GB WD 10,000rpm SATA as RAID-0 @ 92MB/s, Antec TruePower 480W, Zalman 7000 AlCu HSF, Arctic Silver III, FSB 233, CPUv 1.6, 3-4-3-7 PAT