View previous topic :: View next topic |
Author |
Message |
Guest
|
Posted: Wed May 01, 2002 2:18 am Post subject: |
|
|
whoa nice fix....i ran hdparm and got a whopin 1.80mb/s then i added the fix to the start up script and now its at 38.7mb/s....muuuuuuuch betta |
|
Back to top |
|
|
AutoBot l33t
Joined: 22 Apr 2002 Posts: 968 Location: Usually Out
|
Posted: Wed May 01, 2002 5:57 am Post subject: |
|
|
Man you guys know your stuff I got much better performance, not sure how good this is but here is my new numbers:
/dev/hda:
Timing buffer-cache reads: 128 MB in 2.11 seconds = 60.66 MB/sec
Timing buffered disk reads: 64 MB in 3.95 seconds = 16.20 MB/sec
And this is on an old 10gig 5400rpm
PS>Is there a way to enable any of this stuff on my Buslink 7200rpm USB 2.0 drive as it is SCSI emulated? _________________ This message self destructed a long time ago. |
|
Back to top |
|
|
Saravak Guest
|
Posted: Tue May 28, 2002 2:01 pm Post subject: |
|
|
I got clean 33,86 MB/sec !
But I included support for my chip set (via kt266)in kernel when i installed my system! |
|
Back to top |
|
|
DArtagnan l33t
Joined: 30 Apr 2002 Posts: 942 Location: Israel, Jerusalem
|
Posted: Tue May 28, 2002 2:27 pm Post subject: |
|
|
How do u know to get the speed?
I want to check too _________________ All for one and one for All
--
MACPRO machine... |
|
Back to top |
|
|
Malakin Veteran
Joined: 14 Apr 2002 Posts: 1692 Location: Victoria BC Canada
|
Posted: Tue May 28, 2002 3:29 pm Post subject: |
|
|
Quote: | How do u know to get the speed? |
Cached and uncached speeds:
Code: |
hdparm -T /dev/hda
hdparm -t /dev/hda
|
Fore more info:
|
|
Back to top |
|
|
lx Veteran
Joined: 28 May 2002 Posts: 1012 Location: Netherlands
|
Posted: Mon Jun 03, 2002 9:40 pm Post subject: |
|
|
You can also set it using the kernel option
Quote: |
make menuconfig
ATA,IDE...->IDE,ATA->Use PCI DMA by default when available
|
the help talks about some problems with certain boards, but well it works for me.
Cya lX
breaking 30Mb/s, well I have the quantum fireball 30Gb/AS which beats the 30Mb limit, but remember harddisk heads go from outer to inner ring, and on innerring there's less room so at the end of the harddisk I fall under the 30Mb/s think it's about 35 -> 28 Mb/s. _________________ "Remember there's a big difference between kneeling down and bending over.", Frank Zappa |
|
Back to top |
|
|
Xamot n00b
Joined: 20 May 2002 Posts: 29
|
Posted: Mon Jun 03, 2002 9:43 pm Post subject: wonderfull thread |
|
|
This whole Thread should be moved to Tips & Tricks or thoroughly summarized there. I'm MUCH happier with my system performance after reading thru this thread.
Thanks,
Xamot |
|
Back to top |
|
|
Zu` l33t
Joined: 26 May 2002 Posts: 716 Location: BE
|
Posted: Mon Jun 03, 2002 10:41 pm Post subject: |
|
|
Thanks for performance tweaks, this is really cool:
Code: |
mountaindew root # hdparm -Tt /dev/hda
/dev/hda:
Timing buffer-cache reads: 128 MB in 1.41 seconds = 90.66 MB/sec
Timing buffered disk reads: 64 MB in 2.61 seconds = 24.54 MB/sec
mountaindew root # hdparm -c 1 -d 1 /dev/hda
/dev/hda:
setting 32-bit I/O support flag to 1
setting using_dma to 1 (on)
I/O support = 1 (32-bit)
using_dma = 1 (on)
mountaindew root # hdparm -Tt /dev/hda
/dev/hda:
Timing buffer-cache reads: 128 MB in 0.95 seconds =134.98 MB/sec
Timing buffered disk reads: 64 MB in 2.65 seconds = 24.16 MB/sec
mountaindew root # hdparm -m 16 /dev/hda
/dev/hda:
setti multcount = 16 (on)
mountaindew root # hdparm -Tt /dev/hda
/dev/hda:
Timing buffer-cache reads: 128 MB in 1.47 seconds = 87.28 MB/sec
Timing buffered disk reads: 64 MB in 2.66 seconds = 24.02 MB/sec
mountaindew root # hdparm -m 16 -c 1 -u 1 -d 1 /dev/hda
/dev/hda:
setting 32-bit I/O support flag to 1
setting multcount to 16
setting unmaskirq to 1 (on)
setting using_dma to 1 (on)
multcount = 16 (on)
I/O support = 1 (32-bit)
unmaskirq = 1 (on)
using_dma = 1 (on)
mountaindew root # hdparm -Tt /dev/hda
/dev/hda:
Timing buffer-cache reads: 128 MB in 0.92 seconds = 138.78 MB/sec
Timing buffered disk reads: 64 MB in 2.70 seconds = 23.73 MB/sec
|
Timing buffer-cache reads has gone up dramatically!
But Timing buffered disk reads, seems to go down a bit? How come and what is this value actually?
Thanks all |
|
Back to top |
|
|
kode54 n00b
Joined: 16 May 2002 Posts: 42
|
Posted: Tue Jun 04, 2002 6:26 am Post subject: |
|
|
You can also enable automatic DMA on startup for specific devices, adding commands to your kernel append line.
by IDE channel, for both master and slave, ide#=dma
by drive, hdx=dma
All will automatically try for the best possible mode for the drive/controller. In case this causes any problems, I would advise you copy your current settings to a separate backup menu entry, and add the extra settings to the first. Better than digging out your rescue disk. |
|
Back to top |
|
|
jay l33t
Joined: 08 May 2002 Posts: 980
|
Posted: Tue Jun 04, 2002 6:54 am Post subject: |
|
|
Excellent tips folks!
Code: |
/dev/hdb:
Timing buffer-cache reads: 128 MB in 0.39 seconds =327.87 MB/sec
Timing buffered disk reads: 64 MB in 8.50 seconds = 7.53 MB/sec
bash-2.05a# hdparm -m 16 -c 1 -u 1 -d 1 /dev/hdb
/dev/hdb:
setting 32-bit I/O support flag to 1
setting multcount to 16
setting unmaskirq to 1 (on)
setting using_dma to 1 (on)
multcount = 16 (on)
I/O support = 1 (32-bit)
unmaskirq = 1 (on)
using_dma = 1 (on)
bash-2.05a# hdparm -Tt /dev/hdb
/dev/hdb:
Timing buffer-cache reads: 128 MB in 0.39 seconds =332.02 MB/sec
Timing buffered disk reads: 64 MB in 4.32 seconds = 14.81 MB/sec
|
_________________ Do you want your posessions identified? [ynq] (n) |
|
Back to top |
|
|
c_kuzmanic Guru
Joined: 18 Apr 2002 Posts: 488 Location: Los Angeles , California
|
Posted: Tue Jun 04, 2002 10:06 pm Post subject: |
|
|
This is for those of you with a kt266/82XX chipset:
Contrary to what you might think this chipset was not fully supported until recently, when code submitted by Vojtech Pavlik was incorporated into the 2.4.19-pre9.patch . None of the kernels that you can emerge through portage have that patch, so if you want full support for that chipset you have to build your own patched kernel. Keep in mind that this patch is a work in progress.
So what you need to do is emerge vanilla sources, then head over to ftp.kernel.org get 2.4.19-pre9.patch and apply that patch to the vanilla sources.
Optionally you can also get the pre-emptive kernel patch and apply that AFTER you applied the patch mentioned above.
I did this and my system (Soyo Dragon +, kt266a/VT8233CE) runs smoother than ever:) |
|
Back to top |
|
|
niyogi Apprentice
Joined: 10 Apr 2002 Posts: 199 Location: Austin, TX
|
Posted: Sat Jun 08, 2002 7:46 am Post subject: |
|
|
so what was the reason one would get the following error:
HDIO_SET_DMA failed: Operation not permitted
-S |
|
Back to top |
|
|
arkane l33t
Joined: 30 Apr 2002 Posts: 918 Location: Phoenix, AZ
|
Posted: Sun Jun 09, 2002 2:39 am Post subject: |
|
|
niyogi wrote: | so what was the reason one would get the following error:
HDIO_SET_DMA failed: Operation not permitted
-S |
Are you as root?
|
|
Back to top |
|
|
danielrm26 Tux's lil' helper
Joined: 18 Apr 2002 Posts: 92
|
Posted: Mon Jun 10, 2002 1:47 am Post subject: New Western Digital Special Edition Drive... |
|
|
Greetings,
I just did a new install of 1.2 on my new Special Edition WD Drive that is supposed to be as fast as SCSI in some areas. I am still getting the same numbers that I got with my old drive, which are:
Cached: +- 193MB
Uncached: +- 45MB
Shouldn't the 8MB buffer (vs. 2MB for most drives) increase my speeds? Or is it a certain kind of speeds that will be increased that isn't showing up in this benchmark?
I have done all the tweaks...I get the same numbers. FS is EXT3.
Thanks guys,
danielrm26 _________________ dmiessler.com -- grep understanding knowledge |
|
Back to top |
|
|
SpeedD n00b
Joined: 24 May 2002 Posts: 14 Location: Trabuco Canyon, CA
|
Posted: Mon Jun 10, 2002 3:05 pm Post subject: |
|
|
Is there an equivilant command to tweak SCSI performance? I have a SCSI Ultra160 drive and controller and I'd hate to think I wasn't getting full performance out of it.
TIA,
John. |
|
Back to top |
|
|
sy5tematic n00b
Joined: 31 May 2002 Posts: 15
|
Posted: Mon Jun 17, 2002 9:40 pm Post subject: |
|
|
niyogi wrote: | so what was the reason one would get the following error:
HDIO_SET_DMA failed: Operation not permitted
-S |
I had a similar (if not the same) problem
Assuming that you are running as root, and using a motherboard with a fairly recent chipset in it (in my case, a via8233a), the problem is that the linux kernels in the portage tree (I was using xfs-sources) do not recognize the new chipset. They then default to a Generic PCI IDE controller behavior, which does not support DMA.
The sign of this will be a few lines like
Code: |
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
VP_IDE: Unknown VIA SouthBridge, contact Vojtech Pavlik <vojtech@suse.cz>
|
in the output from dmesg.
My solution was to apply Pavlik's latest patch
(I got it from http://lists.insecure.org/linux-kernel/2001/Dec/5575.html
but you could also try searching for "Vojtech Pavlik patch 8233a" on google.)
Note that this patch is written against 2.5.1 and so when I applied it against the 2.4.18 kernel you get from doing an
"emerge xfs-sources"
I had to edit one file:
/usr/src/[YourLinuxVersionHere]/include/linux/pci_ids.h
by hand, but only to make sure that the
Code: |
+#define PCI_DEVICE_ID_VIA_8233A 0x3147
|
definition got in the right place.
After that, I recompiled the kernel and everything was peachy.
UDMA133 works just fine on my Maxtor 80GB.
-Sy5 |
|
Back to top |
|
|
c_kuzmanic Guru
Joined: 18 Apr 2002 Posts: 488 Location: Los Angeles , California
|
Posted: Tue Jun 18, 2002 11:34 pm Post subject: |
|
|
See my post on top of this page:) Thanks for providing detailed instructions on how to incorporate this patch.
Last edited by c_kuzmanic on Wed Jun 19, 2002 5:41 am; edited 1 time in total |
|
Back to top |
|
|
DArtagnan l33t
Joined: 30 Apr 2002 Posts: 942 Location: Israel, Jerusalem
|
Posted: Wed Jun 19, 2002 5:25 am Post subject: |
|
|
jay wrote: | Excellent tips folks!
Code: |
/dev/hdb:
Timing buffer-cache reads: 128 MB in 0.39 seconds =327.87 MB/sec
Timing buffered disk reads: 64 MB in 8.50 seconds = 7.53 MB/sec
bash-2.05a# hdparm -m 16 -c 1 -u 1 -d 1 /dev/hdb
/dev/hdb:
setting 32-bit I/O support flag to 1
setting multcount to 16
setting unmaskirq to 1 (on)
setting using_dma to 1 (on)
multcount = 16 (on)
I/O support = 1 (32-bit)
unmaskirq = 1 (on)
using_dma = 1 (on)
bash-2.05a# hdparm -Tt /dev/hdb
/dev/hdb:
Timing buffer-cache reads: 128 MB in 0.39 seconds =332.02 MB/sec
Timing buffered disk reads: 64 MB in 4.32 seconds = 14.81 MB/sec
|
|
I did these steps and my HD performance remains shity
hdparm -m16 -c1 -u1 -d1 /dev/hda
Timing buffer-cache reads: 128 MB in 0.39 seconds =79.87 MB/sec
Timing buffered disk reads: 64 MB in 8.50 seconds = 23.53 MB/sec
What's going on?
My DMA is enabled in kernel and even i chhanged the kernel line in grub menu:
kernel...bla bla...ide0=dma hda=autotune
( ide0=dma hda=autotune are described in /usr/src.../Doc.../ide.txt ) _________________ All for one and one for All
--
MACPRO machine... |
|
Back to top |
|
|
c_kuzmanic Guru
Joined: 18 Apr 2002 Posts: 488 Location: Los Angeles , California
|
Posted: Wed Jun 19, 2002 5:38 am Post subject: |
|
|
pacman wrote: |
I did these steps and my HD performance remains shity
hdparm -m16 -c1 -u1 -d1 /dev/hda
Timing buffer-cache reads: 128 MB in 0.39 seconds =79.87 MB/sec
Timing buffered disk reads: 64 MB in 8.50 seconds = 23.53 MB/sec
What's going on?
My DMA is enabled in kernel and even i chhanged the kernel line in grub menu:
kernel...bla bla...ide0=dma hda=autotune
( ide0=dma hda=autotune are described in /usr/src.../Doc.../ide.txt ) |
Is your chipset fully supported and configured in your kernel? Also, some people make the mistake of putting two harddrives on one ide controller in a master/slave configuration. One ide-channel per drive is best, otherwise you'll get decreased performance regardless of proper dma/chipset configuration. |
|
Back to top |
|
|
AutoBot l33t
Joined: 22 Apr 2002 Posts: 968 Location: Usually Out
|
Posted: Wed Jun 19, 2002 9:10 am Post subject: |
|
|
What speed is your harddrive pacman ? _________________ This message self destructed a long time ago. |
|
Back to top |
|
|
DArtagnan l33t
Joined: 30 Apr 2002 Posts: 942 Location: Israel, Jerusalem
|
Posted: Wed Jun 19, 2002 4:33 pm Post subject: |
|
|
AutoBot wrote: | What speed is your harddrive pacman ? |
liviu@mousehouse liviu $ su
Password:
mousehouse liviu # hdparm -tT /dev/hda
mousehouse liviu # hdparm -qc1 -qd1 -qS120 -qu1 -qk1 /dev/hda
/dev/hda:
Timing buffer-cache reads: 128 MB in 2.58 seconds = 49.71 MB/sec
Timing buffered disk reads: 64 MB in 2.71 seconds = 23.62 MB/sec
mousehouse liviu # exit _________________ All for one and one for All
--
MACPRO machine... |
|
Back to top |
|
|
AutoBot l33t
Joined: 22 Apr 2002 Posts: 968 Location: Usually Out
|
Posted: Thu Jun 20, 2002 1:08 am Post subject: |
|
|
Sorry what I was trying to say is it a 5400rpm or 7200rpm ? _________________ This message self destructed a long time ago. |
|
Back to top |
|
|
DArtagnan l33t
Joined: 30 Apr 2002 Posts: 942 Location: Israel, Jerusalem
|
Posted: Thu Jun 20, 2002 4:52 am Post subject: |
|
|
well, it's 40 Gb IBM so : 7200 RPM _________________ All for one and one for All
--
MACPRO machine... |
|
Back to top |
|
|
Hamshrew n00b
Joined: 20 Jun 2002 Posts: 3
|
Posted: Thu Jun 20, 2002 2:07 pm Post subject: SCSI Performance |
|
|
This has been asked once before, but is there any way to tune SCSI hard drives? The only thing I've found is a mention of using module parameters to enable TCQ, but I don't have my adapter loaded as a module.
These are my current results:
/dev/sda:
Timing buffer-cache reads: 128 MB in 0.54 seconds =239.25 MB/sec
Timing buffered disk reads: 64 MB in 1.63 seconds = 39.26 MB/sec
Is that standard?
Also, anyone know how to set up the drives on a 3ware controller properly? Or do I need to rely on the hardware to handle it? Here's the results for that, for several 7200 rpm drives in a RAID 10
/dev/sdb:
Timing buffer-cache reads: 128 MB in 0.55 seconds =234.00 MB/sec
Timing buffered disk reads: 64 MB in 2.17 seconds = 29.51 MB/sec
The RAID results seem a little low to me. _________________ Gentoo, Gnome, and LaTeX. What more do you need? |
|
Back to top |
|
|
AutoBot l33t
Joined: 22 Apr 2002 Posts: 968 Location: Usually Out
|
Posted: Thu Jun 20, 2002 8:01 pm Post subject: |
|
|
pacman wrote: | well, it's 40 Gb IBM so : 7200 RPM |
Ok I was just checking, your performance should be above what your getting indeed. _________________ This message self destructed a long time ago. |
|
Back to top |
|
|
|