-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Richard Fish wrote: > On 11/26/06, Chris Walters <[EMAIL PROTECTED]> wrote: >> Well, in Gentoo, the results of "hdparm -t /dev/sda" were around 65 MB / >> second. dmesg shows the drive in UDMA/133 mode, and I didn't bother >> with lspci. In Windows, however, the benchmarks are quite different - >> around 4300-4600 KB / second. At least I know the problem is not with >> the drive. > > Well one thing to keep in mind here is that disk throughput is not > constant everywhere. That 65MB/s is on low-numbered cylinders, > typically near the outside edge of the drive, where the linear > velocity is highest. That same drive may only give 20-30MB/s on the > inside cylinders. > > What are you using to benchmark the drive under windows? Is it > measuring raw disk throughput, or through the NTFS filesystem? > > -Richard
Hello Richard, I would assume that the two benchmark programs I am using to measure the drive performance under windows go through windows, and thus the NTFS interface. They are Dr. Hardware and FreshDiagnose. The second one showed a write speed as quoted above, and a read speed of about 280 MB per second... Though the same one reported a read speed of about 900+ MB per second for my USB drive (not really possible, since the maximum speed for USB 2.0 is about 480 MB / second). This is very frustrating. At first the drive was quite fast, under windows and now it is extremely slow... I asked about this in a windows xp pro group, and so far no one is touching it. Regards, Chris -----BEGIN PGP SIGNATURE----- iD8DBQFFa2XjUx1jS/ORyCsRCKLNAJ9zw1OjMaoR0jUs7HUoPtqRMm8eAQCggSuA t7LgUJ4TeZySBTFP0R76wJc= =Z/2j -----END PGP SIGNATURE----- -- gentoo-user@gentoo.org mailing list
