Re: IDE performance drop between 2.4.23 and 2.6.0 - pinpointed!

[Pawel Dziekonski]

hi,

same problem here however it seems that i have pinpointed it )at least
for me):

root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  1.60 seconds = 39.96 MB/sec
root@localhost:~ # 
root@localhost:~ # hdparm -a 256 /dev/hdb

/dev/hdb:
 setting fs readahead to 256
 readahead    = 256 (on)
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  1.60 seconds = 39.93 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  1.74 seconds = 36.77 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  1.60 seconds = 39.91 MB/sec

this is OK! (under 2.4.23-pre8 i get a little bit more than 40)
also changing redaahead to even 8192 gives simmilar results.
also changing it for hda.

now look at this:

modprobe eagle-usb (my adsl modem)
adictrl -w (load firmware to modem , synchronize etc..)

root@localhost:~ # adictrl -w
Sending options to device /proc/bus/usb/001/002
Options successfully sent to driver.
Sending DSP code to device /proc/bus/usb/001/002
DSP code successfully loaded.
Waiting for synchro...
OK .. Modem is synchronized.
root@localhost:~ # 
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.84 seconds = 16.67 MB/sec
root@localhost:~ # 
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.67 seconds = 17.42 MB/sec
root@localhost:~ # 
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.74 seconds = 17.09 MB/sec
root@localhost:~ # umount /proc/bus/usb/
root@localhost:~ # 
root@localhost:~ # 
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.30 seconds = 19.39 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  2.71 seconds = 23.62 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.57 seconds = 17.91 MB/sec
root@localhost:~ # hdparm -a 8192 /dev/hdb

/dev/hdb:
 setting fs readahead to 8192
 readahead    = 8192 (on)
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.09 seconds = 20.74 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.44 seconds = 18.59 MB/sec
root@localhost:~ # hdparm -a 8192 /dev/hda

/dev/hda:
 setting fs readahead to 8192
 readahead    = 8192 (on)
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  2.94 seconds = 21.78 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  2.58 seconds = 24.80 MB/sec
root@localhost:~ # hdparm -ft /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  2.96 seconds = 21.63 MB/sec
root@localhost:~ # hdparm -t /dev/hdb

/dev/hdb:
 Timing buffered disk reads:  64 MB in  3.59 seconds = 17.82 MB/sec
root@localhost:~ # hdparm -t /dev/hdb

disaster...

funny (? ;-) thing is that under 2.4 i have same issue...

my chipset is kt133 (via), i use uhci.

root@localhost:~ # cat /proc/interrupts 
           CPU0       
  0:    2283127          XT-PIC  timer
  1:       4896          XT-PIC  i8042
  2:          0          XT-PIC  cascade
  8:          1          XT-PIC  rtc
  9:          0          XT-PIC  acpi
 10:          0          XT-PIC  eth0, VIA686A
 11:     271497          XT-PIC  uhci_hcd, uhci_hcd
 12:     171840          XT-PIC  bttv0, nvidia
 14:      33427          XT-PIC  ide0
 15:          1          XT-PIC  ide1
NMI:          0 
ERR:          4

tested under pure 2.6.0.  ac scheduler, same under noop and dead.

regards, Pawel

ps. i read the list via www so cc my the replies please. thanks.
-- 
Pawel Dziekonski <pawel.dziekonski|@|pwr.wroc.pl>, KDM WCSS avatar:0:0:
Wroclaw Networking & Supercomputing Center, HPC Department
-> See message headers for privacy policy info.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/