> Well, I can't say it looks the same : > > # ptime dd if=/root/test.rar of=/dev/rdsk/c3t0d0p0 bs=8k count=6400 > 6400+0 records in > 6400+0 records out > > real 1:38.689 > user 0.007 > sys 0.242 > # ptime dd if=/root/test.rar of=/dev/rdsk/c3t0d0p0 bs=4k count=12800 > 12800+0 records in > 12800+0 records out > > real 1:48.955 > user 0.013 > sys 0.293 > # ptime dd if=/root/test.rar of=/dev/rdsk/c3t0d0p0 bs=2k count=25600 > 25600+0 records in > 25600+0 records out > > real 2:10.504 > user 0.026 > sys 0.495 > > > Nothing unexpected, other than the SLOW timing... > (this was under 32bit kernel on the same hardware > than previously).
Yep, that is very slow for an usb 2.0 storage device. In each case we get less than 500 kbytes/secs; no surprise that writing through pcfs is slow, too. Did you test reading from the raw usb disk device using "dd", too? That is a lot faster, correct? Did you test bigger write block sizes than 8K? What happens when you write 32K or 64K blocks to the flash device? What is reported by "iostat -xnzc 5" while you write to the raw usb flash device using once of the above "dd" command? Is the system 100% idle? Or does the kernel consume a significant amount of system cpu time? This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
