The firmware optimizations in SCSI drives, along with the architecture, are engineered for multiple user access. You can see some of the differences here:
http://storagereview.com/php/benchm...&devID_1=254&devID_2=221&devID_3=250&devCnt=4
This is a head to head comparison of the fastest of four basic types of drives, 15K SCSI, 10K SCSI, 10K IDE and 7.2K IDE. The 74GB Raptor keeps up well with the 15K SCSI for general tasks and end user applications on a single user basis. This changes completely when running the server suite applications where the similarly low level rated 10K SCSI completely outperforms the 10K 74GB Raptor. This is due to the firmware and control optimizations in SCSI. If someone were to hack the firmware for single user optimization and disable some of the data safety features in SCSI(to make it equivalent to IDE), then the performance level of the 15K SCSI would rise significantly. As to being faster loading, what do you mean? SCSI will increase boot time and will have approx the same sustained transfer rate as a 10K Raptor. The primary differences will be in disk access time where the 15K SCSI is about 25% faster due to reduced disk latency.
Generally, if it's not a dual or higher processor system(or in some cases single processor server boards without AGP), it will not have SCSI onboard. There are some exceptions, but getting fewer and farther between.
I avoid RAID-0 primarily for the same reasons that I run SCSI for business, reliability. I typically use enterprise level RAID controllers. The purpose of the RAIDs that I run are reliability and data protection. RAID-0 loses all data in the event of a drive failure. This is unacceptable to my thinking. Anything I save is worth running a CRC on in SCSI by default to ensure that it got written properly. Anything for long term storage is worth both backing up and having some degree of redundancy for protection against the inevitable hardware failure.
There are several different types of speed for HDDs. There is no single value for speed. Rotational velocity, seek times, controller latency, firmware and command set optimizations, drive geometry among other factors play a part in determining the rate at which a drive can deliver data. The type and arrangement of data on the disk makes a large difference as well.
Adding RAID-0 changes the picture somewhat. The STR values for RAID are slightly less than that of 2 drives, while reading larger files. The disk access time(time necessary to find and begin reading the data) are usually slightly slower than single drives. RAID-0 typically works better for larger files, though stripe size can have an effect as well.