Sinister Purpose

  • Content Count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About Sinister Purpose

  • Rank

Contact Methods

  • Website URL
  • ICQ

Profile Information

  • Location
    Tigard, Oregon, USA
  1. Sinister Purpose

    Incredible perfomance on 15K.3

    I did a little experiment. On my boot drive, which is about half full, I created a 40MB file (for good measure), and told SpeedDisk to place that file first. It put it after some system files, but definitely in the outer region. I deleted it and ran ATTO. Of course, ATTO's 32MB file was created in the vacated 40MB block. The result? 20% higher scores for the higher block sizes.Now how much would you pay?
  2. Sinister Purpose

    Constant HDD access in XP

    Your computer has a mind of its own, and it is conspiring against you!
  3. They seemed to think it was enabled on a DiamondMax Plus 160, reviewed recently at:
  4. Sinister Purpose

    Incredible perfomance on 15K.3

    Concerning ATTO scores, I don't think we should regard them too highly, and this is why: I assume that by using a 32 MB data set, we are attempting to minimize the effect of the drive's buffer. I checked, and in operation, ATTO creates a file named C:benchtst.$$$ that is 32MB long. The position on the disk would therefore depend upon the amount of data already present on the disk and the degree of fragmentation. This hardly makes good sense. Let's consider what would happen if the drive buffer does not exist, or is somehow disabled by ATTO. Theoretically, if a seek is incurred by each block read, and assuming an optimistic access time of 5mS, a maximum of 200 x 0.5KB blocks could be read in a second. That's 100 KB/s, which differs considerably from 8MB/s. Obviously, the buffer has a tremendous effect on the scores. To allow the buffer to influence the score, while crippling or disabling other such facilities certainly does not accurately reflect real performance. Although the long red and green bars are pretty, I think there are some serious flaws in the way ATTO measures performance. I notice the people at StorageReview don't use it in their testing...
  5. Sinister Purpose

    Terrible SCSI performance in Windows XP

    ...released 11-08-2002. –SP
  6. Sinister Purpose

    SCSI: Atlas 10KIII and Tekram Probs.

    Are there any other internal or external devices connected to your Tekram U160 controller? I read this in an LSI manual, in case it might apply: The autoenable/disable sensing feature on the LSI8955U may enable termination erroneously if it is directly cabled to another SCSI device or host adapter using the same sensing method. The LSI8955U senses the presence of SCSI devices by detecting the ground signal on conductor 50 of the SCSI cable.