Click for Voyager Hard Drive Docks!
One Click on the Banner shows support to my Sponsors
The Source for Mac Performance News and Reviews
Don't forget to check out all the other site features!
SoftRaid Cache Tweaking Tip
Boost Disk Performance up to 44%
By Ryan Dumperth
Ryan Dumperth wrote with a tip that he says dramatically increased performance in Softraid 2.1.5. As noted in in the past, I use SoftRaid's driver even on single drives (see my Ultimate Mac, PCI SCSI Cards Compared article (listed as Conley in the tables as they originally owned Softraid) and XLR8 G3/400 review for disk scores) as it has the lowest latency and highest performance of any driver I've seen. Ryan discovered another trick that could boost performance even further on drives that support it:
" I have an S900/200 with an Initio Miles and a brand new Barracuda 9GB drive, using SoftRAID 2.1.5 drivers. The drive has a 512K multisegmented cache, and SoftRAID allows one to change the number of segments. On my setup, it defaulted to three segments, and this was modifiable. I tried 4,2 and 1, to some dramatic results. The optimal setting was one cache segment. It didn't seem to affect anything but sustained writes and read transactions per second (although the latter has always fluctuated wildly on my machine).
According to HDT's benchmark, my average sustained write climbed from 13.6MB/sec to 19.6MB, a 44% increase. The tests were done with all my regular extensions enabled. The drive has been noticably faster than my 9GB Atlas II. I can't imagine what a single Cheetah, let alone a RAID setup with them would be like. Here's a screensot of the before and after test results. Ryan Dumperth-"
Note that the R/W Transactions/Sec also improved significantly.
I welcome feedback from other readers on this tip. Before making any change to disk drivers, make sure your data is backed up just in case.
Ken Wehr reported a 50% increase in sustained write speeds on his Ultra2 Seagate Cheetah using the above tip.
Jose Azevedo wrote that he saw a decrease in MacBench scores (see his MB5 scores) using the Cache tweak. Remember that you can't compare MacBench's disk mix tests to FWB's Benchtest scores which were used to show the before/after IO rates for the Tip. Often there is a 'waterbed' effect (pushing down in one area pops up in another) to many disk tweaks - what's good for large file IO may negatively affect small file IO and vice-versa. That's why some Raid/Disk drivers have options for video capture and pre-press use for instance.
For other articles on SCSI, IDE and PCI Controlers, see my SCSI/IDE features page.
Back to XLR8YOURMAC.COM
Your Source for the best in CPU/SCSI/VIDEO card reviews, daily news, and more!
Copyright © Mike, 1999.
No part of this site's content is to be reproduced in any form without permission.
All brand or product names mentioned here are properties of their respective companies.
Disclaimer: Users must read and are bound by the Site Terms & Conditions of Use.