On Thu, 4 Dec 2003 06:28:26 -0600, nate at refried.org wrote: > On Thu, Dec 04, 2003 at 12:06:09AM -0600, Ian Stoner wrote: > > After I copy a large file (~30 megs) my processor activity shoots up > > to 100% for 10 to 15 seconds, and I can't use my system during that > > time. > > Sounds like write caching without DMA enabled. Use hdparm to make > sure that DMA is enabled on your drives. That nailed it! Thanks, Nate. I had never heard of hdparm--I found a webpage that recommended some likely safe settings, tried them, and got a more than 5-fold increase in buffered disc reads. I've included the output of the commands I ran, in case anyone cares. Maybe I'm the only one who's never heard of hdparm? If anyone knows of any other parameters I should pay attention to, please advise. Thanks again! root at goodmanbrown:~# hdparm /dev/hda /dev/hda: multcount = 0 (off) IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 0 (off) keepsettings = 0 (off) readonly = 0 (off) readahead = 8 (on) geometry = 2495/255/63, sectors = 40088160, start = 0 root at goodmanbrown:~# hdparm -Tt /dev/hda /dev/hda: Timing buffer-cache reads: 368 MB in 2.01 seconds = 183.08 MB/sec Timing buffered disk reads: 12 MB in 3.07 seconds = 3.91 MB/sec root at goodmanbrown:~# hdparm -X66 -d1 -u1 -m16 -c3 /dev/hda /dev/hda: setting 32-bit IO_support flag to 3 setting multcount to 16 setting unmaskirq to 1 (on) setting using_dma to 1 (on) setting xfermode to 66 (UltraDMA mode2) multcount = 16 (on) IO_support = 3 (32-bit w/sync) unmaskirq = 1 (on) using_dma = 1 (on) root at goodmanbrown:~# hdparm -Tt /dev/hda /dev/hda: Timing buffer-cache reads: 376 MB in 2.02 seconds = 186.14 MB/sec Timing buffered disk reads: 64 MB in 3.04 seconds = 21.05 MB/sec -- Ian Stoner Philosophy Department University of Minnesota _______________________________________________ TCLUG Mailing List - Minneapolis/St. Paul, Minnesota http://www.mn-linux.org tclug-list at mn-linux.org https://mailman.real-time.com/mailman/listinfo/tclug-list