Hello, all!
I have a machine with six identical IDE drives (WD1200BB),
three of them connected to the on-board controller (it is AMD 768MPX chipset),
and other three are connected to the Promise controller (PDC 20269).
All drives are UDMA 100, read speed measured by "hdparm -t /dev/hd[abcefg]"
is about 45 MBytes/s for every drive. However, the write speed seems
to differ between AMD and Promise controllers. I've tried to do
time sh -c 'dd if=/dev/zero of=/dev/hdX bs=1024k count=2048; sync'
- it takes about 50 seconds (~40 MByte/s write speed) on hda, hdb and hdc,
but 2 minutes 48 seconds (~12 MByte/s write speed) on hde, hdf and hdg.
I have 1 GB of RAM, server is dual athlon 2000+. Kernel is 2.4.20-pre5-ac1.
Is there any problem with the Promise IDE driver on Linux?
Thanks,
-Yenya
--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/ Czech Linux Homepage: http://www.linux.cz/ |
Pruning my incoming mailbox after being 10 days off-line,
sorry for the delayed reply.
On Wed, 2002-09-04 at 18:57, Jan Kasprzak wrote:
> - it takes about 50 seconds (~40 MByte/s write speed) on hda, hdb and hdc,
> but 2 minutes 48 seconds (~12 MByte/s write speed) on hde, hdf and hdg.
> I have 1 GB of RAM, server is dual athlon 2000+. Kernel is 2.4.20-pre5-ac1.
>
> Is there any problem with the Promise IDE driver on Linux?
One or two with various versions. I've seen the same thing with both IDE
and non IDE devices with the AMD chipset. The two 64bit slots and the
onboard I/O seems to have markedly higher write throughput.
Andrew D Kirch wrote:
: excuse the mailer, but I've been having similar issues with my promise
: controller... interesting huh? give me a little more info on the board and
: make sure to target a reply to [email protected], I'll keep you posted on
: anything I find to fix it... right now my speed difference isn't THAT
: dramatic... Going to a beta mandrake install from 8.2 I've lost 40% of my
: software raid-0 performance (from a hdparm of 80MB/s to 56MB/s. :(
:
: I'm going to play with jumpers later, so I'll let you know what I find.
Mainboard is MSI K7D-Master (AMD 760 MPX chipset),
the Promise controller is PDC 20269 with the following lspci
output:
02:04.0 Unknown mass storage controller: Promise Technology, Inc. 20269 (rev 02) (prog-if 85)
Subsystem: Promise Technology, Inc.: Unknown device 4d68
Flags: bus master, 66Mhz, slow devsel, latency 32, IRQ 16
I/O ports at 9000 [size=8]
I/O ports at 9400 [size=4]
I/O ports at 9800 [size=8]
I/O ports at 9c00 [size=4]
I/O ports at a000 [size=16]
Memory at f8000000 (32-bit, non-prefetchable) [size=16K]
Expansion ROM at <unassigned> [disabled] [size=16K]
Capabilities: [60] Power Management version 1
-Yenya
--
| Jan "Yenya" Kasprzak <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839 Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/ Czech Linux Homepage: http://www.linux.cz/ |
Pruning my incoming mailbox after being 10 days off-line,
sorry for the delayed reply.
On Wed, Sep 04, 2002 at 07:57:29PM +0200, Jan Kasprzak wrote:
> Hello, all!
>
> I have a machine with six identical IDE drives (WD1200BB),
> three of them connected to the on-board controller (it is AMD 768MPX chipset),
> and other three are connected to the Promise controller (PDC 20269).
> All drives are UDMA 100, read speed measured by "hdparm -t /dev/hd[abcefg]"
> is about 45 MBytes/s for every drive. However, the write speed seems
> to differ between AMD and Promise controllers. I've tried to do
>
> time sh -c 'dd if=/dev/zero of=/dev/hdX bs=1024k count=2048; sync'
>
> - it takes about 50 seconds (~40 MByte/s write speed) on hda, hdb and hdc,
> but 2 minutes 48 seconds (~12 MByte/s write speed) on hde, hdf and hdg.
> I have 1 GB of RAM, server is dual athlon 2000+. Kernel is 2.4.20-pre5-ac1.
>
> Is there any problem with the Promise IDE driver on Linux?
This looks most likely to be incorrect programming of UDMA registers on
the Promise chip. UDMA is asymmetrical by design - the hd->mb and mb->hd
speeds can be different. If you get correct read rates, it's the
harddrive that's set up correctly, for writes it's the IDE chip that has
to be programmed for the speed.
--
Vojtech Pavlik
SuSE Labs