Linux-Hardware Digest #280, Volume #10 Thu, 20 May 99 09:13:23 EDT
Contents:
Linux POS Hardware Project (Paul Rae)
SB64 PCI ("Andre Malafaya Baptista")
Re: AMD K6-2 Problems (Justin The Cynical)
Re: Laptop touchpad movement in X window limit to a certain portion of the screen
(Mohd H Misnan)
Re: Star Office and Office 97 ... (Manfred Becker)
Re: Blur problem with X-Window ("Larry Brasfield")
Re: I got a problem with ESS ES56T-PI modem under RedHat 6.0. (Rob Clark)
A growing online shopping mall (Hyper)
Dual Celerons from ComputerNerds ([EMAIL PROTECTED])
Up to four video cards on one system? (Christof Neuberger)
Re: IDE faster than SCSI UW? (jwk)
Re: Linux compatible SMP motherboard, reviewed (The Network Wizard)
Re: Compaq Laptop Modem ([EMAIL PROTECTED])
Re: DDS-3 DAT drive (Chris Mauritz)
Re: Invalid Partition Table (help requested) (**Nick Brown)
MediaGX Motherboard ("Johnas Cukier")
Re: IDE faster than SCSI UW? (root)
Re: All the current OSes are idiotic (was Re: Is Windows for idiots?) (westprog)
----------------------------------------------------------------------------
From: Paul Rae <[EMAIL PROTECTED]>
Subject: Linux POS Hardware Project
Date: Thu, 20 May 1999 11:21:07 +0100
check out www.cromwells.co.uk
=====Original Message=====
From: Christopher Sciullo [mailto:[EMAIL PROTECTED]]
Posted At: Wednesday, May 19, 1999 5:26 AM
Posted To: hardware
Conversation: Linux POS Hardware Project
Subject: Linux POS Hardware Project
We are involved in an International Linux Point of Sale Project.
We sell the POS Hardware.
Read all about it: http://pcpos.hostme.com <http://pcpos.hostme.com>
Chris
------------------------------
From: "Andre Malafaya Baptista" <[EMAIL PROTECTED]>
Subject: SB64 PCI
Date: Thu, 20 May 1999 11:46:40 +0100
I heard that Linux doesn't support Pnp PCI cards very well...
Will I have any problems setting up a SoundBlaster64 PCI Pnp card?
TIA,
André
(remove NOSPAM to reply)
------------------------------
From: [EMAIL PROTECTED] (Justin The Cynical)
Subject: Re: AMD K6-2 Problems
Date: 20 May 1999 10:29:15 GMT
Reply-To: [EMAIL PROTECTED]
On 15 May 1999 05:18:51 GMT, Mark Hahn <[EMAIL PROTECTED]>
wrote:
->> This sounds like the problem I am getting with a new FIC VA503+ w/ K6-2 350. I am
booting successfully with a floppy now, but I wish I could get back to the
->> usual way of booting from the hard disk boot partition in /
->
->Linux does _not_ have problems with properly configured K6-2 systems.
Nope. It doesn't. This problem may be an issue with the motherboard
tho. It seems that the FIC-503+ is very picky about the RAM it will work
correctly with.
Check the FIC web page under the reports section.
--
"So does the Leatherman Micra. It has tweezers too but I find that I need
the pliers on the SuperTool in order to deal adequately with my nasal hair."
- Giles Todd, Monk
Justin The Cynical, Bastard of Smaug - [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Mohd H Misnan)
Subject: Re: Laptop touchpad movement in X window limit to a certain portion of the
screen
Date: 20 May 1999 10:54:22 GMT
On Wed, 19 May 1999 22:11:19 +0800, Hanley Loo wrote:
>Hi,
> Anybody out there can help me. I'm installing linux on my laptop. I
>can only get the touchpad to work in the upper left area of the X window
>interface. All the rest of the area is restricted. I have tried killing
>gpm and disabling it. But it doesn't help. I have check the X86config
>file, the configuration of the mouse is correct.
May be posting more info like your XF86Config on the Pointer section will really
help us diagnose your problem? If you're using gpm, you need to use MouseSystems
protocol and use /dev/gpmdata as your device. Otherwise you can just disable gpm
and use protocol = PS/2 and device = /dev/psaux.
--
|Mohd Hamid Misnan|[EMAIL PROTECTED] / [EMAIL PROTECTED] |i|
|MacOS 8.6 + |http://www.geocities.com/SiliconValley/3319/ |M|
|AMD K6-2/300 + |We want to take over the world, but we don't have |a|
|Linux 2.2.7 i586 |to do it tomorrow. It's OK by next week - Linus T.|c|
------------------------------
From: Manfred Becker <[EMAIL PROTECTED]>
Subject: Re: Star Office and Office 97 ...
Date: Wed, 19 May 1999 19:51:59 +0200
Pigou wrote:
> Can Star Office read the files from MS (ehh,ehh) Office 97 ???
>
Hi Pigou,
yes and no !
Yes it can read simple Documents.
No it can not read complex Documents with input-field, listboxes, macros and
so on. Maybe it the new release just published. I wil try it.
Gtnx
- Manfred -
--
+----------- Manfred Becker -----------+ W W
! D-53842 Troisdorf-Altenrath ! ( )___( )
! http://home.tronet.de/manfred.becker ! ( o o )
! mailto:[EMAIL PROTECTED] ! ( O )
+--------------------------------------+--oooO-`___'-Oooo---
------------------------------
From: "Larry Brasfield" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Blur problem with X-Window
Date: Thu, 20 May 1999 11:15:58 GMT
Robert Tomanek <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
> [EMAIL PROTECTED] wrote:
...
> > But then I found that the screen is
> > blurred, everything: letter, cursor, window, will have several "shadow"s
> > beside it horizontally. What's the problem with this? And what shall I do?
...
> Are you pushing the limits of your graphics card? If yes, simply
> decrease mode clock in the modeline you're using in /etc/X11/XF86Config.
I doubt such shadows will be diminished by going
to a lower resolution.
The shadow effect, (if accurately described as a
horizontally shifted, faint replica of the strong primary
image), arises due to signal reflections in the cable.
This means either: the cable has incorrect or poorly
controlled impedance; or the monitor and video card
have incorrect termination impedance. The former
is many times more probable than the latter. The
fix is to purchase a better cable. No amount of
fiddling with scan parameters will cure it. In fact,
the problem will be aggravated by more capable
video cards and faster responding CRT drivers.
--
--
Larry Brasfield
Above opinions may be mine alone.
(Humans may reply at unundered [EMAIL PROTECTED] )
------------------------------
Subject: Re: I got a problem with ESS ES56T-PI modem under RedHat 6.0.
From: [EMAIL PROTECTED] (Rob Clark)
Date: Thu, 20 May 1999 11:36:45 GMT
In article <[EMAIL PROTECTED]>,
Andrey Vlassov <[EMAIL PROTECTED]> wrote:
>I got a problem with ESS ES56T-PI modem under RedHat 6.0.
>And now my question - "Does somebody make modem ESS ES56T-PI works under
>Linux at all?"
>
>Reference to HomePage of modem:
>
>http://www.esstech.com/Product/Communications/es56pi.htm
"All of the TeleDrive? chipsets are controller-less Data/FAX V.34bis/V.17
modem solution."
"Controllerless" means controller functions are moved off the modem card
into software. So, yes, it is a winmodem...
Rob Clark, [EMAIL PROTECTED]
http://www.o2.net/~gromitkc/winmodem.html
------------------------------
From: Hyper <[EMAIL PROTECTED]>
Subject: A growing online shopping mall
Date: 19 May 99 14:11:47 -0400
Reply-To: [EMAIL PROTECTED]
For links to the best shopping on the internet go to the Hyper-Lynx
homepage at http://www.hyper-lynx.com
------------------------------
From: [EMAIL PROTECTED]
Subject: Dual Celerons from ComputerNerds
Date: Wed, 19 May 1999 18:37:08 GMT
In article <RG103.31173$[EMAIL PROTECTED]>,
bryan <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED] wrote:
> : So, my boss told me on Friday to go and get myself a new computer and
> : send him the bill. He said, as long as it came in under $3000, it was
> : okay by him.
>
> sweet ;-)
>
> : At first, I was all delirious and sweaty-palmed, figuring this was my
> : chance to rig together one of those dual-Celeron systems I've been
> : reading so much about. Not that I *need* two CPU's--it just sounds so
> : damn cool, I've been itching to try it. Or build some other, monster
> : 'dream system' of my own design.
>
> if he's footing the bill, get a dual pentium 2 at 450mhz ea. no
> hacking needed and it will perform as well as a dual celeron (why does
> this sound so funny?) ;-)
>
> : My enthusiasm was somewhat abated when I started pricing out individual
> : components and estimating the amount of *time* involved in setting up a
> : custom system compared to just plunking down $3000 for a loaded Dell.
>
> arg!!! no dell - please.
>
> : What I am in fact *supposed* to be doing with this new box is writing a
> : WinNT GUI for some simulation software, and I need to get started, like,
> : NOW (which is why I was instructed to go out and buy a new machine).
> : What if I bought a bunch of parts and it took me 2-3 weeks to get them
> : all working right? Such a delay would be _EXPENSIVE_.
>
> if you buy the right parts, there should be no such delay. the key is
> choosing parts that are known to work well together. stay with real
> brands and you should be ok.
>
> : It appears I can get (marginally?) more bang for my buck by assembling a
> : custom system, and be much more certain of being able to run Linux
> : relatively trouble-free (a *requirement*, AFAIC), but does this really
> : matter when it's not my dime anyway? If some piece of hardware doesn't
> : work with Linux as supplied, I can probably just go buy a new one and
> : *still* come in under $3K. I wonder what others would do in this
> : situation. Do you take a risk and use the opportunity to build your
> : dream system, or stick to something less exciting but, perhaps, more
> : secure?
>
> : Any advice, cautionary tales, etc. much appreciated.
>
> : Dell Dimension T Series:
> : Dimension® XPS T550, Pentium® lll Processor
> : at 550 MHz
>
> get TWO pent 2 or 3 cpus at the same time. they should be the same
> 'stepping' (hardware version). get retail boxed ones if possible with
> the same manuf. date and lot # to be sure.
>
> : Memory:
> : 128MB 100MHz SDRAM
>
> try to ensure you get ECC ram (error correcting). 2 sticks of 128meg
> should do you fine.
>
> : Keyboard:
> : Microsoft® Natural® Keyboard Elite
> : Monitor:
> : 19" (18.0" viewable, .26dp) Trinitron®
> : P990 Monitor
> : Video Card:
> : 16MB Diamond Viper TNT 3D AGP Graphics Card
>
> no - get a matrox card. millennium G200 if possible - its more compatible.
>
> : Hard Drive:
> : NEW 13.6GB Ultra ATA Hard Drive (7200 RPM)
>
> bullcrap. do yourself a favor and get a good scsi system:
>
> ibm (or seagate) 10k rpm drive, 9gig, ultra2 scsi.
> controller could be a tekram dc390u2 (forget the exact model #)
>
> : Floppy Drive:
> : 1.44MB Floppy Drive
> : Operating System:
> : Microsoft® Windows® 98
>
> yeah, right - pay a M$ tax why-dont-you ;-)
>
> : Mouse:
> : MS IntelliMouse®
>
> no! logitech 3 button/wheel mouse (must be RETAIL pack, I'm told).
>
> : Modem:
> : 3Com® USRobotics V.90* PCI Telephony WinModem
> : for Sound
>
> the standard (in my mind) is an EXTERNAL us robotics COURIER.
>
> : CD-ROM Drive:
> : 40X Max Variable CD-ROM Drive
>
> plextor 40x. scsi and very fast/reliable. also get a plextor writer
> 8/20 - just for good measure, and have them mounted in an external
> dual-bay case (better cooling and you can use it on other systems to
> burn discs).
>
> : Sound Card:
> : Turtle Beach Montego II A3D320V Sound
> : Card,FactoryInstall
>
> I don't do soundcards so I have no idea. you don't really need one
> for a business system, do you?
>
> : Speakers:
> : Altec Lansing ACS-340 Speakers with Subwoofer
> : TV TUNER:
> : STB PCI TV/FM Tuner
>
> more flash to make Dell rich. not reasonable for a business system.
>
> if you REALLY want a killer system, get hardware raid. a pair of ibm
> 10k rpm drives and a good pci raid controller (there are a few that
> work well under linux and NT). pull a spinning drive out of its bay
> and you still have a working/running system. I run hardware raid at
> home - its very nice and gives you peace-of-mind. a low-end mylex
> (dac960-pg) can be gotton for about $600 mailorder.
>
> ethernet card should be an Intel pro100 card. is what linux vendors
> seem to be selling.
>
> speaking of such, you may want to just buy it all configured from VA
> Research or Penguin computing or some other linux system builder.
>
> my motherboard rec:
> 1st choice: asus p2bd
> 2nd choice: tyan 1832DL
>
> --
> Bryan
>
I just got a dual Celeron setup from ComputerNerds of Florida
(dudecomputer.com) for $780:
Epox KP6-BS board (5 PCI slots, no SCSI; they also sell the
Epox BXB-S board with u2w SCSI)
two Celerons on slockets modified for dual-processor use
two double-fan heatsinks
128 MB of ECC RAM.
What's nice is that it's tested, burned in, and guaranteed to run
reliably at 450 MHz in the dual configuration. These Nerds also sell a
complete box for about $3500 with Red Hat Linux and NT installed.
As for hard drives, I've been looking around for what to buy lately, and
storagereview.com has tested the new 1999 models of 7200 rpm SCSI drives
from Western Digital, IBM, and Seagate (not yet Quantum). The WD
Enterprise series comes out on top, and they are available with 2 MB
cache (standard), or 4 MB in the AV model (for continuous streaming).
The 10,000 rpm drives are slightly faster, but much hotter (fan
required) and noisier than the 7200 rpm units.
There have been extensive discussions of multiprocessing on Slashdot.
Linux makes full use of the second CPU, NT is boosted maybe 25%, and Win
9x not at all. The multithreaded design of Photoshop, however, is said
to fully exploit the second CPU.
Dual Celerons are a great bargain. I hope you can take advantage of it.
-- Jon Palmer
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
From: Christof Neuberger <[EMAIL PROTECTED]>
Subject: Up to four video cards on one system?
Date: Wed, 19 May 1999 10:56:33 +0200
What is necessary to run up to four video (graphik) cards on one single
system?
Christof Neuberger
fon: +49-(0)30-2801-1858
fax: +49-(0)30-2801-1020
email: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (jwk)
Subject: Re: IDE faster than SCSI UW?
Date: Wed, 19 May 1999 21:52:02 GMT
Reply-To: [EMAIL PROTECTED]
On Wed, 19 May 1999 11:54:21 -0400, Andy Longton <[EMAIL PROTECTED]>
wrote:
>I've got a puzzler. A new 10 GB UDMA IDE drive is showing up as faster
>than both of my 4.5 GB UW drives. Is this right?
>
>Pointers to any tuning or performance resources would be appreciated.
>
>I've searched quite a bit so far, and have come up empty handed.
>
>
>Details...
>
>Adaptec 2940 UW SCSI
>P2B system board (PII 450)
>
>/dev/hda = IBM model DTTA-351010, 10.1 GB, 5400 RPM
>/dev/sda = Quantum Viking SCA -w- UW adapter, 4.5 GB, 5400 RPM
>/dev/sdb = IBM model DDRS-34560, 4.5 GB, 7200 RPM
>
>Typical results from running hdparm -tT on an idle system...
>
># ./hdparm -tT /dev/hda /dev/sda /dev/sdb
>/dev/hda:
> Timing buffer-cache reads: 64 MB in 0.54 seconds =118.52 MB/sec
> Timing buffered disk reads: 32 MB in 2.65 seconds =12.08 MB/sec
>/dev/sda:
> Timing buffer-cache reads: 64 MB in 0.54 seconds =118.52 MB/sec
> Timing buffered disk reads: 32 MB in 3.14 seconds =10.19 MB/sec
>/dev/sdb:
> Timing buffer-cache reads: 64 MB in 0.53 seconds =120.75 MB/sec
> Timing buffered disk reads: 32 MB in 2.88 seconds =11.11 MB/sec
>
>
>Note: In no test did /dev/hdb perform better than /dev/hda for disk
>reads.
first, for real-life benchmarking, try bonnie.
second, it's quite possible this is so. Remember that the DDRS is the very
bottom end of the IBM-scsi-line, and the DTTA is a newer design.
I also have 2 DDRS'es, and I am quite pleased with them. In raid-0, I get
maximum reading speeds of about 18 mb/s, which is good enough for me.
Also, using scsi-equipment, you can have 2 harddisks, a cd-player and a
cd-writer and still use only 1 irq.
Good luck,
Jurriaan
--
Am I indecisive? Can I get back to you on that?
------------------------------
From: The Network Wizard <[EMAIL PROTECTED]>
Subject: Re: Linux compatible SMP motherboard, reviewed
Date: Wed, 19 May 1999 20:44:34 +0200
Reply-To: [EMAIL PROTECTED]
Bill Henning wrote:
> I've posted a review of the PCChips M750i motherboard - supports dual CPU's,
> Intel BX chipset, onboard i740 with 8Mb of SDRAM, onboard 3D sound
> (and it is inexpensive!) at my site, http://www.cpureview.com
>
> It works fine with Linux, I'll be posting a brief article on how to run Linux
> with an SMP kernel on it shortly. I am using two Celeron 300A's running at
> 450Mhz (this is a toy box, NOT a production machine; I don't overclock
> production machines).
Do you expect a similar performance when using Celeron 400 ?
When do you plan to post you article about SMP linux kernel?
thank you
Pol
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.setup,comp.os.linux.misc
Subject: Re: Compaq Laptop Modem
Date: Wed, 19 May 1999 20:22:48 +0200
In message <[EMAIL PROTECTED]>
[EMAIL PROTECTED] (Bob Sully) wrote:
> Hi all-
> I'm trying to set up PPP on my Compaq laptop, which has an
> integrated 56K modem. Initially, it was set to some odd memory addresses
> (it has two, now set to 2F8 and 1400h, IRQ3). I can't get Linux to operate
> it, even as root - it keeps coming back with "Modem is busy", even as root.
> Any ideas? It works with Windoze. Thanks.
>
>
This is a so called WinModem. If you would like to use a modem and get internet
access, buy a PCMCIA modem that supports also Dos programs. If it does, it will
work under Linux. I had the same problem, bought a Wysecom 56k modem and now
it's ok.
--
=====================================================
=\ Acorn RiscPC, StrongARM @202MHz - Intel Outside /=
=====================================================
... Shh! Be vewy quiet, I'm hunting wuntime errors!
------------------------------
Crossposted-To: comp.os.linux.misc
From: Chris Mauritz <[EMAIL PROTECTED]>
Subject: Re: DDS-3 DAT drive
Date: Wed, 19 May 1999 18:44:12 GMT
In comp.os.linux.misc [EMAIL PROTECTED] wrote:
> In article <[EMAIL PROTECTED]>,
> Chris Mauritz <[EMAIL PROTECTED]> wrote:
>> In comp.os.linux.misc Rod Roark <[EMAIL PROTECTED]> wrote:
>> > Chris Mauritz <[EMAIL PROTECTED]> wrote:
>> >>What's the secret to getting a DDS-3 DAT drive to work with linux?
>> >>I've tried reading/writing tapes with Redhat 5.2 and Redhat 6.0
>> >>using an HP and a Seagate drive without success.
>> >>
>> >>When I try to read/write tapes, I get:
>> >>
>> >>tape read error: Input/output error
>> >>
>> >>I've compiled SCSI tape support into the kernel (tried both 2.0.36
> and
>> >>2.2.9) and made sure the device files (/dev/nst0 /dev/st0) exist.
> Do you have 'st' (SCSI Tape) support either compiled-in or as a
> loadable module. Check it with 'lsmod'.
It is compiled in. I don't have any modules compiled (nor do I have
support for modules compiled in to the kernel).
> Also, you can use 'script' to capture output to a file then edit and
> post it. Include 'dmesg', 'lsmod', exact error message give, ... .
Will do.
C
--
Christopher Mauritz
[EMAIL PROTECTED]
------------------------------
From: **Nick Brown <[EMAIL PROTECTED]>
Subject: Re: Invalid Partition Table (help requested)
Date: Thu, 20 May 1999 13:58:25 +0200
Reply-To: [EMAIL PROTECTED]
> /dev/hda1 * 1 277 2094088+ 6 DOS 16-bit >=32M
> /dev/hda2 278 838 4241160 5 Extended
hda1 through hda4 are the first four (potential) entries in the MBR.
You have one primary partition and one extended partition. hda3 and
hda4 are always reserved, but unused here.
> /dev/hda5 555 770 1632928+ 6 DOS 16-bit >=32M
> /dev/hda6 771 838 514048+ 6 DOS 16-bit >=32M
hda5 and hda6 are the "first two logical drives that are not further
extended partitions" in the extended partition. Note that an extended
partition is a bit like a whole disk: it can have up to four
"partitions" (execpt they are usually called "logical drives"), and one
(actually, more than one, but there are DOS limits to respect) can be in
its turn an extended partition.
In this case, you appear to have an "extended extended" partition from
278 to 554.
> /dev/hda7 278 291 105777 83 Linux native
> /dev/hda8 292 427 1028128+ 83 Linux native
> /dev/hda9 428 522 718168+ 83 Linux native
> /dev/hda10 523 539 128488+ 82 Linux swap
> /dev/hda11 540 554 113368+ 83 Linux native
There are logical drives in the "extended extended" partition. But I'm
not sure how you ended up with five of them.
In any case, this kind of structure is almost certainly beyond the
capability of DOS FDISK.
The maintainer of fdisk is andries(dot)brouwer(at)cwi(dot)nl. He always
seems interested in new, exciting, different disk layouts. If you mail
him with your fdisk output, my guess is that you may get a sensible
reply.
--
===============================================================
Nick Brown, Strasbourg, France (Nick(dot)Brown(at)coe(dot)int)
Protect yourself against Word 95/97 viruses, free - check out
http://www.geocities.com/NapaValley/Vineyard/1446/atlas-t.html
===============================================================
------------------------------
From: "Johnas Cukier" <[EMAIL PROTECTED]>
Subject: MediaGX Motherboard
Date: Wed, 19 May 1999 13:35:05 -0400
I've got a MediaGX motherboard with built in video and sound. I would like
to configure the video properly at this point so that I can run X Windows.
I can't get it to start X even if I choose generic VGA for the video adapter
and monitor. I suspect I will have similar problems with my sound. At this
point, I just want the video to work.
Can anyone help? I thank you in advance for your help.
Thanks
------------------------------
From: root <[EMAIL PROTECTED]>
Subject: Re: IDE faster than SCSI UW?
Date: 20 May 99 12:45:50 GMT
OK, I've retested using Bonnie, details are inserted below...plus a few
questions about RAID 0 & spindle speed.
jwk <[EMAIL PROTECTED]> wrote:
> On Wed, 19 May 1999 11:54:21 -0400, Andy Longton <[EMAIL PROTECTED]>
> wrote:
>>I've got a puzzler. A new 10 GB UDMA IDE drive is showing up as faster
>>than both of my 4.5 GB UW drives. Is this right?
>>
>>Pointers to any tuning or performance resources would be appreciated.
>>
>>I've searched quite a bit so far, and have come up empty handed.
>>
>>Details...
>>
>>Adaptec 2940 UW SCSI
>>P2B system board (PII 450)
[ Snip! ]
/dev/sda = Quantum Viking SCA -w- UW adapter, 4.5 GB, 5400 RPM
/dev/sdb = IBM model DDRS-34560, 4.5 GB, 7200 RPM
/dev/hda = IBM model DTTA-351010, 10.1 GB, 5400 RPM
> first, for real-life benchmarking, try bonnie.
Done...
# ./bonnie -d /mnt/sda3
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
100 5987 70.2 6374 7.9 2194 5.4 4880 52.1 7814 5.0 174.2 1.4
# ./bonnie -d /mnt/sdb2
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
100 8337 97.8 15999 18.7 3964 9.4 6384 67.8 10336 8.9 605.6 4.2
# ./bonnie -d /mnt/hda3
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
100 7991 94.4 9787 10.6 3558 9.0 6237 66.9 11555 10.9 354.9 2.2
If I read this right, there doesn't seem to be much of a difference between the
SCSI CPU load and the UDMA CPU load. The DDRS is generally faster, but imposes
a greater load on the CPU.
> second, it's quite possible this is so. Remember that the DDRS is the very
> bottom end of the IBM-scsi-line, and the DTTA is a newer design.
>
> I also have 2 DDRS'es, and I am quite pleased with them. In raid-0, I get
> maximum reading speeds of about 18 mb/s, which is good enough for me.
> Also, using scsi-equipment, you can have 2 harddisks, a cd-player and a
> cd-writer and still use only 1 irq.
Well, I've used UW SCSI for about 5 years, so I have seen the benifits on my
slower machines. This machine is a smoker, though, and I'm not sure if the
benifits of SCSI are worth it anymore.
==> A higher RPM drive, regaurdless of the type of bus, seems to be the main
way to improve performance since the CPU load will be about the same per K
transfered. <==
Does this fit your experience?
Paying 2x for a drive plus $100-200 for a controller isn't cheap, though the
parts tend to be better made; example -- Toshiba 32x CDROMs - SCSI vs. EIDE:
SCSI - all metal, solid tray, heavier, quieter.
EIDE - has a paper (!) seal on top, thinner metal, tray wobbles, noisier.
Do you think switching the two SCSI drives to RAID 0 give a noticable boost?
What kind of conflicts, if any, between the Quantum Viking and IBM DDRS should
I expect -- since they don't provide the same performance? Any FAQ pointers
besides the LDP how-to?
I realize that switching to RAID 0 would most likely require backing up and
restoring the partitions after the RAID 0 conversion.
------------------------------
From: westprog <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: All the current OSes are idiotic (was Re: Is Windows for idiots?)
Date: Thu, 20 May 1999 12:27:09 GMT
In article
<[EMAIL PROTECTED]
s.net>,
[EMAIL PROTECTED] wrote:
> On Wed, 19 May 1999 11:49:29 +1200, "Stuart Fox"
> <[EMAIL PROTECTED]> wrote:
>
> >> In Windows we have a large OS, true, but we also have many of the
> >> "modules" that make it up (device drivers, services, etc.) provided
by
> >> 3rd parties. We *know* how stable this arrangement is...it's
> >> "crash-o-matic"...and you have no good way of determining which
party
> >> is guilty and has to make a fix...you just get a freezup or a BSOD
and
> >> you reboot...over and over again.
> >>
> >> On the other hand, OpenVMS is a large, monolithic OS produced by a
> >> single company. It has very few crash problems...and what it
*does*
> >> have is nearly always device-driver related...usually device
drivers
> >> produced by third parties.
> >
> >So what you're saying is when OpenVMS crashes because of a third
party
> >device driver, it's not Digitals fault, but when NT crashes because
of a
> >third party device driver, it IS Microsofts fault?
> >
> >Seems to me you're a little inconsistent there.
>
> Nope, I'm saying that privileged code (which is what device drivers
> usually are) needs to be written by folks with an intimate
> understanding of the OS internals, and who do things in a careful,
> professional way. When privileged code is written by folks who are
> not careful and professional in how they go about things you get lots
> of crashes.
> DEC (now Compaq) was a good engineering company and they produced
> solid products. What crashes I've seen in OpenVMS are mostly the
> result of 3rd party drivers.
> MS is *not* a good engineering company, and the products they produce
> are not solid, and worse, tend to cause others to make it worse by
> their design. The interfaces into NT are so weak that others have
> tended to write their code in the form of device drivers or services,
> so they can have access to the machine that NT doesn't support in a
> safe way, and this means that you've got more people, from more
> companies, writing privileged code, and the result it predictable:
> more crashes.
> What would be optimal would be an OS where it was not necessary, or
> possible, for any 3rd party to write any privileged code. Then,
> provided that the OS writers did their job properly, you should have a
> system that doesn't crash at all. Going the other direction, where
> *everyone* writes privileged code, is just asking for problems in my
> opinion.
In the new perfect OS, each component would have just enough privilege
to do its job. This still has great scope for problems - a faulty file
system for example, could cause havok - but there is no inherent reason
why, say, a Com port driver should be able to crash the system.
While Digital's approach was technically a success, it failed miserably
commercially. That is because, among other things, they failed to
support a wide range of third party hardware.
...
--
J.
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
** 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 (and comp.os.linux.hardware) via:
Internet: [EMAIL PROTECTED]
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
******************************
