On Thu, 2003-11-06 at 18:37, Javier Villavicencio wrote: > > > > The initial SATA drive performance isn't bad, but isn't that great. I > > haven't been bold enough yet to turn on any specific optimizations in > > this new machine yet, so it will likely get better: > > > > Gandalf root # hdparm /dev/hda > > > > /dev/hda: > > multcount = 16 (on) > > IO_support = 0 (default 16-bit) > > unmaskirq = 0 (off) > > using_dma = 1 (on) > > keepsettings = 0 (off) > > readonly = 0 (off) > > readahead = 8 (on) > > geometry = 155061/16/63, sectors = 156301488, start = 0 > > Gandalf root # > > > > Gandalf root # hdparm -tT /dev/hde > > > > /dev/hde: > > Timing buffer-cache reads: 128MB in 0.34 seconds=376.47MB/sec > > Timing buffered disk reads: 64MB in 1.78 seconds= 35.96MB/sec > > Gangalf root # > >
> Uhm, i configured this with 2.6 kernel on my girlfriend's computer (not an ASUS, bur > an ABIT with same characteristics) and the hdparm gives around 1000Mb/s form > buffer-cache and 56Mb/s from buffered disk reads. > You may try this hdparm params for your SATA drive: > > hdparm -a255 (-a4096 with 2.6kenrel) -d1 -m16 -c1 -X69 -M254 -u1 > > This should give better results to you from the Siimage SATA 3112 chip. > Javier, I'll probably end up starting another thread on this after I do some web browsing on the subject, but the only one of these parameters that had any effect was -d0 (dma off) which made the performance very slow. (1.3MB/S) Other than that none of these seem to have any effect at all on this system. I'll have to make sure there are no driver updates or anything like that as I am using 2.4.22-aa1. I'll also try your 2.4.23 kernel as that could certainly have updates for this chip. Mark -- [EMAIL PROTECTED] mailing list
