<br><br><div><span class="gmail_quote">On 3/30/06, <b class="gmail_sendername">Tom Lichti</b> <<a href="mailto:tom@redpepperracing.com">tom@redpepperracing.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Ryan Steffes wrote:<br>><br>><br>> Your buffered reads look more or less normal, but your cached reads<br>> are less than half of what they should be. I get over 1900 MB/sec.<br>> for cached reads.
<br>><br>> Are you sure DMA is enabled for your drives ?? Was the system doing<br>> anything else when you ran the tests?<br>><br>> You never want to put a hard drive on the same IDE bus as an optical
<br>> drive, it will slow the HDD down to the (probably 33) speed of the<br>> latter. You should put both of your hard drives on the primary IDE<br>> channel. This, however, does not explain the slow speed of your /dev/
<br>> hda drive, but it would explain problems with /dev/hdc.<br>><br>> Make sure you are using a proper 80-conductor cable for your hard<br>> drives as well, you need to be running in UDMA mode to get top
<br>> performance (assuming your Motherboard has a modern IDE interface,<br>> without that you're pretty much dead in the water).<br>><br>><br>> The system was recording during that test, I can shut myth down later
<br>> to get "clean" numbers if it would help. DMA is on for both drives.<br>What does hdparm -d /dev/hd? show?</blockquote><div><br>/dev/hda:<br> using_dma = 1 (on) <br></div><br></div>/dev/hdc:<br> using_dma = 1 (on)
<br><br>