Linux-Hardware Digest #92, Volume #14 Thu, 28 Dec 00 02:13:04 EST
Contents:
Re: Performance difference between UDMA33 and UDMA66 (David W. Studeman)
USB optical mouse (Logitech) (Apostata)
Re: "COM 5"? ("D. Stimits")
Re: PCI configuration in Linux 2.2.18 : IRQ assignment? ("D. Stimits")
Re: Performance difference between UDMA33 and UDMA66 (Maxim)
Re: Guillemot isis and linux ("Jason Byrne")
Re: Performance difference between UDMA33 and UDMA66 (Markus Kossmann)
----------------------------------------------------------------------------
From: David W. Studeman <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.mandrake
Subject: Re: Performance difference between UDMA33 and UDMA66
Date: Wed, 27 Dec 2000 21:07:01 -0800
Reply-To: [EMAIL PROTECTED]
You didn't miss anything, just wasted money is all. If your previous
reading from your drive was less than 33mbs, no faster interface is going
to make your drives absolute transfer speed any faster. Very few drives on
the market can even match UDMA33 except for some 10,000rpm drives which
barely surpass it at 34mbs. I have the same drive on this machine and
that's about the transfer speed it delivers. Most 10,000rpm drives sustain
transfer at 28 to 29mbs with very few exceptions. On my faster machine, the
A7V mobo came with ATA100 which I disabled because I use SCSI anyway and
never even considered the IDE controller in the purchase. The ATA66 spec
you are looking for is only the bandwidth available on the controller, not
the drive.
In real life, transfer speeds are slower yet because files are not read as
a continous stream of raw data as you do in the tests. Keep the motherboard
controller and keep in mind that drive transfer speed is about three years
behind controller speed.
I'm amazed at salesman that believe as most people do about controllers
giving you faster drive performance.
The drives data sheets will tell you what the sustained transfer rate is
but you have to fish their website pretty hard usually to find it.
Dave
Maxim wrote:
> I have WD136BA hard drive (Ultra ATA/66, 7200 rpm) on ASUS P2B-F MBoard
> running Mandrake-7.0. This motherboard doesn't support UDMA66 but still
> WD136BA performs quite well using UDMA33 mode; testing it with "hdparm
> -t" I got about 21.8 MB/sec. Recently I bought and installed Promise
> UDMA66 PCI card with hope to enable UDMA66 mode. I ungraded kernel to
> 2.2.18 and patched it from
> http://www.kernel.org/pub/linux/kernel/people/hedrick/. After a little
> problems I successfully booted Linux from HD connected to UDMA66 PCI
> card. Now "hdparm -i" shows that HD in UDMA mode4 but it DID NOT improve
> performance AT ALL! I still get the same 21.8 MB/sec with "hdparam -t"!
> What did I miss? How 7200 rpm HD should perform in UDMA mode 4? What
> else should I do? Below I enclosed "hdparm -i" output for old and new
> condiguration. THANKS!
>
> Maxim
>
> -----------connected to MBoard IDE
> (UDMA33)------------------------------------------------
> /dev/hda:
>
> Model=WDC WD136BA, FwRev=P74OA30A, SerialNo=WD-WM7261272469
> Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
> RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34
> BuffType=3(DualPortCache), BuffSize=1961kB, MaxMultSect=16,
> MultSect=off
> DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
> CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=26712000
> WARNING 10197936 ORPHANED SECTORS :: KERNEL REPORTING ERROR
> tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
> IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
> UDMA modes: mode0 mode1 *mode2 mode3 mode4
> Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3
> ATA-4
>
> # /sbin/hdparm -t /dev/hda
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 2.95 seconds =21.69 MB/sec
> # /sbin/hdparm -t /dev/hda
>
> /dev/hda:
> Timing buffered disk reads: 64 MB in 2.96 seconds =21.62 MB/sec
>
=========================================================================================================
>
> --------------------------------connected to UDMA66 Promise PCI
> card-------------------
> /dev/hde:
>
> Model=WDC WD136BA, FwRev=P74OA30A, SerialNo=WD-WM7261272469
> Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
> RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34
> BuffType=3(DualPortCache), BuffSize=1961kB, MaxMultSect=16,
> MultSect=off
> DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
> CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=26712000
> WARNING 10197936 ORPHANED SECTORS :: KERNEL REPORTING ERROR
> tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
> IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
> UDMA modes: mode0 mode1 mode2 mode3 *mode4
> Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3
> ATA-4
>
> # /sbin/hdparm -t /dev/hde
>
> /dev/hde:
> Timing buffered disk reads: 64 MB in 2.95 seconds =21.69 MB/sec
> # /sbin/hdparm -t /dev/hde
>
> /dev/hde:
> Timing buffered disk reads: 64 MB in 2.93 seconds =21.84 MB/sec
>
============================================================================================================
>
------------------------------
From: Apostata <[EMAIL PROTECTED]>
Subject: USB optical mouse (Logitech)
Date: Thu, 28 Dec 2000 05:30:04 -0000
Hello,
Does anyone know where I could find (if one exists) a driver for my USB
optical wheel mouse from Logitech. Sounds like a bit of a tall order, but
I figured I'd ask :)
Thanks in advance!
P.S. I'm planning to run Mandrake
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
Date: Wed, 27 Dec 2000 22:35:43 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: "COM 5"?
[EMAIL PROTECTED] wrote:
>
> >
> > Is your BIOS properly set to "o/s is NOT plug-n-play aware"? There are
> > some quite good 56k PCI US Robotics modems, but for PCI, the hardware
> > must be initialized, and only then can the software be set up. For
> > example, setserial -a on a /dev/ttyS? will list its current settings;
> > setserial can also be used to set them. File /proc/pci will list
> current
> > pci device setup, which must match the setserial settings on the
> > particular serial port. Either the modem can be made to match the port
> > with setpci, or the port can be made to match the modem (with
> > setserial). Look at the base address and irq. Samples:
> >
>
> I can't set my BIOS to "o/s is NOT plug-n-play aware"--or anything
> related to that: the BIOS on my machine doesn't have any setting or
> option related to pnp on its menu or submenus. I did manage to get rid
> of the winmodem, though.
This is rather unusual on a modern motherboard. Even on a semi-modern
one. Unless this is an original pentium or 486, you might want to
double-check that option...it might have some different naming to it.
>
> Is there any way to get around the pnp issue? I've tried using
> setserial already and that didn't work. After doing everything you
> suggested--barring the crucial first requirement--wvdial still couldn't
> find my modem. However: the setserial -v command output all of the
> info about the modem, including items that I hadn't entered as args. I
> have not tried the setpci command: I looked at its man page and I don't
> think that I understand how to use it.
Is /etc/wvdial.conf set up properly? Does it have a "Modem0" and "Dialer
Defaults" sections? A sample for my USR is (note it names /dev/ttyS2, so
setserial on /dev/ttyS2 must match lspci -v for the modem):
[Modem0]
Modem = /dev/ttyS2
Baud = 115200
Init1 = ATZ
Init2 = ATQ0 V1 E1
[Dialer Defaults]
Modem = /dev/modem
Baud = 57600
Init1 = ATZ
Init2 = ATQ0 V1 E1
SetVolume = 1
Dial Command = ATDT
Init4 = ATM1L3
Also, a sample dial section for a specific phone number (made up) is:
[Dialer myisp]
Username = myusername
Password = apassword
Phone = 5551212
Area Code = 202
Inherits = Dialer Defaults
Stupid mode = 0
This would allow root to do:
wvdial myisp
>
> Does the lame BIOS on my machine make it impossible for Linux to
> recognize the pci modem, or is there some work-around?
Something has to initialize the pci hardware. If not the bios, then
manually. This is where setpci comes in handy. Once you know a serial
port and its settings, you can use setpci to initialize the modem to
that address and irq. Otherwise there wouldn't be any ability to use pci
slots at all. If you have any pci cards running now, and your bios
doesn't have such an ability, I'd be surprised that any of them are
working. One of the really significant differences between ISA
plug-n-play versus PCI cards, is that PCI cards have their own internal
bios that the motherboard bios or the operating system can use to set
them up...whereas ISA pnp offers no clues to the system, and must always
be set up by an outside program or manually (all PCI is plug-n-play, but
it has automated setup possible).
So a recap of the issue:
PCI hardware can be initialized by the bios if set for o/s not
pnp-aware; if not, then it is up to the operating system or other
program to do this. The serial port must match the PCI card
settings...that leaves the choices that either the serial port can be
made to match the modem, or the modem can be made to match the serial
port (or a mix). To set or view a serial port, use setserial. To view a
PCI card, use lspci or view /proc/pci. To set a PCI card, either enable
the bios to do this by telling it the o/s won't be doing it, or use
setpci.
>
> Sent via Deja.com
> http://www.deja.com/
------------------------------
Date: Wed, 27 Dec 2000 22:38:02 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: PCI configuration in Linux 2.2.18 : IRQ assignment?
PCI card irq can be altered with setpci. See also lspci.
Chris Rankin wrote:
>
> Hi,
> I have an old Pentium 90 PC with one PCI slot free, and I am trying to
> replace the old 28.8 Kbps ISA modem with a modern 56 Kbps PCI modem. The
> new modem will *have* to work in both Linux 2.2.18 and Windows 95.
>
> I have read the Plug-and-Play HOWTO. I also appreciate that I *could*
> use a 56 Kbps external modem at a pinch, except that I have enough plugs
> and adapters on my single power-point as it is. One more would cross the
> line from crazy to dangerous.
>
> The ISA modem is on IRQ 4, which I freed up from one of the
> motherboard's serial ports. The other serial port was on IRQ 3, which I
> have also freed. IRQs 5 and 9 belong to an ISA-PnP soundcard, 7 belongs
> to the motherboard's parallel port, 10 is used by a PCI card providing 2
> USB ports, and 11 is used by a PCI video card. (Typically, anyway!) I
> boot as a PnP OS for the sake of Windows 95.
>
> Basically, I will have IRQs 3 and 4 available for the new modem.
> However, I'm concerned that any new PCI card will not get configured to
> use either, because I tested adding another PCI card and it was
> configured to use IRQ 9 instead! This caused a conflict with the
> soundcard of course.
>
> I have also experimented with setpci to change the INTERRUPT_LINE value
> for the PCI USB-port card (from 10 to 3), and discovered that it didn't
> make the slightest bit of difference which IRQ I wrote into the bus. Is
> this information written by the BIOS purely for reference? The output
> from lspci did not change, and /proc/interrupts showed that the USB uhci
> module always used IRQ 10 anyway. Is there any way that I could actively
> reassign the IRQs to ensure that Linux gets it right? I am assuming that
> Windows is PnP-enough to sort itself out! (yeah, right ...)
>
> To my knowledge, the BIOS has never assigned any device to use IRQ 3,
> despite this IRQ currently being free. As I mentioned earlier, the USB
> hub uses 10, the video card uses 11 and any new PCI card was given IRQ 9
> where it conflicted with the ISA-PnP soundcard. Is there something
> special about IRQ 3? I *did* check with the BIOS, and it regards both
> IRQs 3 and 4 as available for PnP.
>
> The soundcard is an Ensoniq SoundScape, and uses the sscape module with
> these options:
>
> sscape irq=5 dma=1 io=0x338 mpu_io=0x330 mpu_irq=9
>
> This module refuses to load without the mpu_io and mpu_irq options.
> However, IRQ 9 never appears in /proc/interrupts even when it is loaded.
> Could the BIOS somehow be unaware that the soundcard needs it? If so,
> why? Marking an IRQ as "reserved for a legacy ISA card" is not an option
> when booting as a PnP OS, BTW.
>
> Thanks for any help with any of this. And the sooner the ISA bus bites
> the dust, the better :-)!!!
>
> Cheers,
> Chris
------------------------------
From: Maxim <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.mandrake
Subject: Re: Performance difference between UDMA33 and UDMA66
Date: Thu, 28 Dec 2000 05:59:27 GMT
Wow, I'd like to know it before I bought this UDMA66 controller card. Fortunately,
it's just $20. What is then
the difference between ATA/33 and ATA/66 HDs?
Maxim
"David W. Studeman" wrote:
> You didn't miss anything, just wasted money is all. If your previous
> reading from your drive was less than 33mbs, no faster interface is going
> to make your drives absolute transfer speed any faster. Very few drives on
> the market can even match UDMA33 except for some 10,000rpm drives which
> barely surpass it at 34mbs. I have the same drive on this machine and
> that's about the transfer speed it delivers. Most 10,000rpm drives sustain
> transfer at 28 to 29mbs with very few exceptions. On my faster machine, the
> A7V mobo came with ATA100 which I disabled because I use SCSI anyway and
> never even considered the IDE controller in the purchase. The ATA66 spec
> you are looking for is only the bandwidth available on the controller, not
> the drive.
> In real life, transfer speeds are slower yet because files are not read as
> a continous stream of raw data as you do in the tests. Keep the motherboard
> controller and keep in mind that drive transfer speed is about three years
> behind controller speed.
> I'm amazed at salesman that believe as most people do about controllers
> giving you faster drive performance.
> The drives data sheets will tell you what the sustained transfer rate is
> but you have to fish their website pretty hard usually to find it.
>
> Dave
>
> Maxim wrote:
>
> > I have WD136BA hard drive (Ultra ATA/66, 7200 rpm) on ASUS P2B-F MBoard
> > running Mandrake-7.0. This motherboard doesn't support UDMA66 but still
> > WD136BA performs quite well using UDMA33 mode; testing it with "hdparm
> > -t" I got about 21.8 MB/sec. Recently I bought and installed Promise
> > UDMA66 PCI card with hope to enable UDMA66 mode. I ungraded kernel to
> > 2.2.18 and patched it from
> > http://www.kernel.org/pub/linux/kernel/people/hedrick/. After a little
> > problems I successfully booted Linux from HD connected to UDMA66 PCI
> > card. Now "hdparm -i" shows that HD in UDMA mode4 but it DID NOT improve
> > performance AT ALL! I still get the same 21.8 MB/sec with "hdparam -t"!
> > What did I miss? How 7200 rpm HD should perform in UDMA mode 4? What
> > else should I do? Below I enclosed "hdparm -i" output for old and new
> > condiguration. THANKS!
> >
> > Maxim
> >
> > -----------connected to MBoard IDE
> > (UDMA33)------------------------------------------------
> > /dev/hda:
> >
> > Model=WDC WD136BA, FwRev=P74OA30A, SerialNo=WD-WM7261272469
> > Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
> > RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34
> > BuffType=3(DualPortCache), BuffSize=1961kB, MaxMultSect=16,
> > MultSect=off
> > DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
> > CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=26712000
> > WARNING 10197936 ORPHANED SECTORS :: KERNEL REPORTING ERROR
> > tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
> > IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
> > UDMA modes: mode0 mode1 *mode2 mode3 mode4
> > Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3
> > ATA-4
> >
> > # /sbin/hdparm -t /dev/hda
> >
> > /dev/hda:
> > Timing buffered disk reads: 64 MB in 2.95 seconds =21.69 MB/sec
> > # /sbin/hdparm -t /dev/hda
> >
> > /dev/hda:
> > Timing buffered disk reads: 64 MB in 2.96 seconds =21.62 MB/sec
> >
>
>---------------------------------------------------------------------------------------------------------
> >
> > --------------------------------connected to UDMA66 Promise PCI
> > card-------------------
> > /dev/hde:
> >
> > Model=WDC WD136BA, FwRev=P74OA30A, SerialNo=WD-WM7261272469
> > Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
> > RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34
> > BuffType=3(DualPortCache), BuffSize=1961kB, MaxMultSect=16,
> > MultSect=off
> > DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
> > CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=26712000
> > WARNING 10197936 ORPHANED SECTORS :: KERNEL REPORTING ERROR
> > tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
> > IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
> > UDMA modes: mode0 mode1 mode2 mode3 *mode4
> > Drive Supports : ATA/ATAPI-4 T13 1153D revision 17 : ATA-1 ATA-2 ATA-3
> > ATA-4
> >
> > # /sbin/hdparm -t /dev/hde
> >
> > /dev/hde:
> > Timing buffered disk reads: 64 MB in 2.95 seconds =21.69 MB/sec
> > # /sbin/hdparm -t /dev/hde
> >
> > /dev/hde:
> > Timing buffered disk reads: 64 MB in 2.93 seconds =21.84 MB/sec
> >
>
>------------------------------------------------------------------------------------------------------------
> >
------------------------------
From: "Jason Byrne" <[EMAIL PROTECTED]>
Subject: Re: Guillemot isis and linux
Date: Wed, 27 Dec 2000 22:27:22 -0800
"Jason Byrne" <[EMAIL PROTECTED]> wrote in message
news:FLz26.8936$[EMAIL PROTECTED]...
> "Marty" <[EMAIL PROTECTED]> wrote in message
> news:92dqn3$1l0m$[EMAIL PROTECTED]...
> > Hi!
> > does anybody have the maxi studio isis by guillemot and knows if there
are
> > any drivers so it'll work under linux?
>
> I'm not familiar with all the cards from Guillemot... but I have a
> 'Guillemot Maxi-Gamer Phoenix' (16mb PCI) - which is just a 3dfx Voodoo
> Banshee (chipset) .. so I follow 3dfx tips for setup.
>
> If you can figure out which chipset the card actually uses... you're
> probably in luck.
oohhh... it's a really fancy soundcard setup ;-) - just checked the
guillemot site...
I noticed http://www.alsa-project.org has one Guillemot soundcard listed
(MaxiSound Fortissimo) - don't see any mention of any other Guillemot cards.
>
> - Jason
>
> >
> > if not WHO CAN MAKE ONE!!!
> >
> > thanx there!
> >
> > Marty
> >
> >
> >
>
>
------------------------------
From: Markus Kossmann <[EMAIL PROTECTED]>
Subject: Re: Performance difference between UDMA33 and UDMA66
Date: Thu, 28 Dec 2000 06:50:30 +0100
Maxim wrote:
>
> I have WD136BA hard drive (Ultra ATA/66, 7200 rpm) on ASUS P2B-F MBoard
> running Mandrake-7.0. This motherboard doesn't support UDMA66 but still
> WD136BA performs quite well using UDMA33 mode; testing it with "hdparm
> -t" I got about 21.8 MB/sec. Recently I bought and installed Promise
> UDMA66 PCI card with hope to enable UDMA66 mode. I ungraded kernel to
> 2.2.18 and patched it from
> http://www.kernel.org/pub/linux/kernel/people/hedrick/. After a little
> problems I successfully booted Linux from HD connected to UDMA66 PCI
> card. Now "hdparm -i" shows that HD in UDMA mode4 but it DID NOT improve
> performance AT ALL! I still get the same 21.8 MB/sec with "hdparam -t"!
> What did I miss?
Nothing. In your case the data transfer rate is limited by the drives
hardware parameters ( rotational speed and and data density) and not by
transfer rate of the bus.
To see a difference , you will probably need a IDE drive of the latest
generation ( i.e WD400BB or IBM DTLA ) which is fast enough to saturate
UDMA 33 .
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.hardware.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Hardware Digest
******************************
