<br><br><p><DEFANGED_div><DEFANGED_span class="gmail_quote">On 10/22/07, <b class="gmail_sendername">Justin Denick</b> <<a href="mailto:justin.denick@gmail.com">justin.denick@gmail.com</a>> wrote:</DEFANGED_span><blockquote class="gmail_quote" DEFANGED_style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br><br><p></p><p><DEFANGED_div><DEFANGED_span class="q" id="q_115c876ef99e2a2e_1">On 10/20/07, <b class="gmail_sendername">Bob Kline</b> <<a href="mailto:bob.kline@gmail.com" target="_blank" DEFANGED_onclick="return top.js.OpenExtLink(window,event,this)">
bob.kline@gmail.com</a>> wrote:<blockquote class="gmail_quote" DEFANGED_style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br><br><br><p></p><p>On 10/20/07, <b class="gmail_sendername">Michael Mol</b> <<a href="mailto:mikemol@gmail.com" target="_blank" DEFANGED_onclick="return top.js.OpenExtLink(window,event,this)">
mikemol@gmail.com</a>> wrote:</p><blockquote class="gmail_quote" DEFANGED_style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On 10/20/07, Bob Kline <<a href="mailto:bob.kline@gmail.com" target="_blank" DEFANGED_onclick="return top.js.OpenExtLink(window,event,this)">bob.kline@gmail.com</a>> wrote:<br>> On 10/20/07, Bob Kline <<a href="mailto:bob.kline@gmail.com" target="_blank" DEFANGED_onclick="return top.js.OpenExtLink(window,event,this)">
bob.kline@gmail.com</a>> wrote:<br>> > On 10/20/07, Michael Mol <
<a href="mailto:mikemol@gmail.com" target="_blank" DEFANGED_onclick="return top.js.OpenExtLink(window,event,this)">mikemol@gmail.com</a>> wrote:<br><br>[snipped]<br><br>> > > How about:<br>> > ><br>> > > mount -t ext3 -o async /dev/X /disk2
<br>> ><br>> ><br>
> ><br>> > That would seem to make sense, but the<br>> > man page for mount suggests that the<br>> > default is async.<br>> ><br>> > One can use<br>> ><br>> > mount -t ext3 -o defaults /dev/X /disk2
<br>> ><br>> > but its unclear why this is useful. I'm<br>> > assuming that one doesn't need it, but<br>> > will try it just to be sure.<br>> ><br>> > More later.<br><br>[snipped]
<br><br>><br>> The async option is apparently the default,<br>> and has no affect on the slow transfer rates<br>> I'm seeing.<br>><br>> Some have wondered whether the connection<br>> is somehow defaulting to a USB
1.1 mode. But<br>> if so, I see no way to do anything about that.<br><br>What kind of performance do you get if you try:<br><br>dd if=/dev/X of=/dev/null bs=1M count=1000<br><br> </blockquote><p></p><p><br>On the Mandriva machine it took 231 seconds,
<br>
for a bit rate of 4.5 MBps, or 36 Mbps. That's<br>
about 50% faster than the mounted transfer<br>
rate using cp. <br>
<br>
I'll check the Kubuntu machine after I finish doing<br>
something else.<br>
<br>
-Bob</p></blockquote></DEFANGED_span></p><DEFANGED_div><p><br>Be sure to see what the DMA specs are too.<br><br>hdparm -cuda /devx<br><br>As you can see it makes a big differnece<br><br></p>justin@da-boxx ~ $ sudo hdparm -t /dev/hda<br>
<br>/dev/hda:
<br> Timing buffered disk reads: 166 MB in 3.02 seconds = 54.92 MB/sec<br>justin@da-boxx ~ $ sudo hdparm -d0 /dev/hda<br><br>/dev/hda:<br> setting using_dma to 0 (off)<br> using_dma = 0 (off)<br>justin@da-boxx ~ $ sudo hdparm -t /dev/hda
<br><br>/dev/hda:<br> Timing buffered disk reads: 18 MB in 3.26 seconds = 5.53 MB/sec<br>justin@da-boxx ~ $ sudo hdparm -d1 /dev/hda<br><br>/dev/hda:<br> setting using_dma to 1 (on)<br> using_dma = 1 (on)<br>justin@da-boxx
~ $ sudo hdparm -t /dev/hda<br><br>/dev/hda:<br> Timing buffered disk reads: 166 MB in 3.02 seconds = 54.88 MB/sec<br>justin@da-boxx ~ $<br><br>It's a kernel option<br>CONFIG_IDEDMA_AUTO=y</blockquote><p><DEFANGED_div><br>Thanks. I'll check this out.
<br><br> -Bob<br> <br></p><DEFANGED_div><br></p><DEFANGED_div><br>