SATA Vs IDE

[Anonymous]

I have run a series of disk tests on 2 SATA disks and 1 Ultra IDE disk.

They are all formatted with NTFS but the SATA disks are reading a MUCH lower throughput (see the results below). Is this an issue with PT5 or have

SATA 1 = Custom Test = 6.2 MB/s
SATA 1 = Workstation Test = 3.9 MB/s
UIDE 1 = Workstation Test = 155 MB/s!!!

Surely I must be doing something wrong here

Any advice?

Thanks in advance

BB

[David (PassMark)]

The workstation configuration in the disk test is mostly reading and mostly random access. Which makes it VERY sensitive to any caching.

Our software requests to the Windows O/S not to cache, but the device driver or the hard drive itself is pobably doing caching.

Clearly a result of 155MB/sec is too fast for a physical drive and so the data must be coming from a RAM cache somewhere.

SATA only effects the transfer of data on the cable from the disk drive. It doesn't effect the rotation speed or data density or the drive. So the difference between the two interfaces in this test should be small.

To help reduce the cache effect.
- Try testing just after the machine has booted (the windows cache varies in size over time).
- Don't run very short tests.
- Try doing several test runs.
- Use sequential disk access rather than random
- Use large file sizes
- Use the Raw disk access method

------
David