Linux-Hardware Digest #119, Volume #11 Sat, 28 Aug 99 06:13:27 EDT
Contents:
Sound problem ("Paolo Barolat-Romana")
Re: Webserver specs needed...thanks (chris pitzel)
Re: win 98 and nt crashing but linux is not on o/c'ed computer? what gives?
("xrongor")
Re: CDR (Michael Ravits)
Re: Finding LAN Bottleneck (Greg Leblanc)
Re: Dual Pentium II shows as Dual Celeron... (Greg Leblanc)
Re: Adaptec 78xx Hardware RAID and RH6 (Greg Leblanc)
Re: cost of leased line in notting hill, London England ("Andrew Smith")
Aktien zu verschenken!!!! Free Stock Programm!!!! ("erhard.m")
Re: Linux and Ultra DMA/ATA-66 drives ("Jeff Towers")
NEW_VER-NewsControlSystem2.1.0_EMailForDemoKey(FreeNEWSservers) ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: "Paolo Barolat-Romana" <[EMAIL PROTECTED]>
Subject: Sound problem
Date: Sat, 28 Aug 1999 06:29:16 GMT
I recently got redhat linux and I've been trying to get my sound to work. I
have a card with an ess1888 sound chip. The linux sound how-to says that
this sound ship is supported. When I run sndconfig or setup, however, the
computer does not offer this as an option. I have tried other ess type
cards, but they don't seem to work: I keep getting some sort of message that
my sound card is busy.
So, now I want to install the driver "by hand". However, I don't know where
to find the driver for this sound chip. Do you know where I can get it?
Then, I figure that I have to recompile my kernel with this new driver. Is
this right? If so, any help with this would be greatly appreciated.
Thanks,
Paolo
------------------------------
From: chris pitzel <[EMAIL PROTECTED]>
Crossposted-To: comp.infosystems.www.servers.misc,comp.infosystems.www.servers.unix
Subject: Re: Webserver specs needed...thanks
Date: 28 Aug 1999 07:35:25 GMT
>other than NT. Anyway, what I have is the following (see below) but need
>a question answered about the CPU, which really comes down to, "is it
>enough"?
>Pent 166 with 256 megs RAM (going to upgrade to 1gig), 1 ultra wide SCSI
>HD for OS, and a RAID5 config. with 3 10Krpm Seagate Cheetah HD's.
The main issue you will experience with the older Pentium-166 board will
be an issue of cost.
Quite simply put, equipping a mobo with a Pentium 166 likely will require
SIMMS. SIMMS, to achieve a density of 256mb likely will have to be at
least 64mb parts, and if you want to upgrade further, you will need 128mb
SIMMS.
If you've ever looked at prices on high-density SIMM RAM, you'll quickly
realize that you could buy a Pentium II processor *and* motherboard *and*
your desired amount of RAM (in the form of DIMMS) for the same money.
>spend the extra "jack" on a new CPU and MotherBoard...or with the 1 gig
>RAM and RAID config, is this enough? ANY opinions or different
You really don't have any choice. You'll never achieve 1 gigabyte on a
socket 7 or SS7 board. Only a few of the SS7 chipsets claim to support
1gb, but nobody, as of yet, has ever successfully built a 1gb system on a
SS7 board.
>configuration ideas would be GREATLY appreciated ...peace
ASUS P2B-D with 4 x IBM 256mb SDRAM ECC registered DIMMS works just fine
for a 1gbyte configuration. Make sure you get Pentium-2-350 or better
processors for 1gb support.
------------------------------
From: "xrongor" <[EMAIL PROTECTED]>
Crossposted-To: alt.comp.hardware.overclocking
Subject: Re: win 98 and nt crashing but linux is not on o/c'ed computer? what gives?
Date: Sat, 28 Aug 1999 00:53:47 -0700
this is a guess, but its probably the ram that's failing. windows uses a
different method to access the ram than linux does and i know from
experience that bsod is often attributable to bad ram.
i also had one oc'd computer that i could not get 98 to install on, so i
slowed it down.. after it was installed it worked fine when oc'd.. go
figure.
i'd try and borrow some different ram and see if it solves it.
hope this helps
randy
Gaiko Kyofusho <[EMAIL PROTECTED]> wrote in message
news:7q7p2k$lmm$[EMAIL PROTECTED]...
> Hi, i recently started having problems with the windows parts of my
> tri-boot computer (win nt for most apps, win98 for other apps and linux
> for when i don't need some win app) recently. it started with my
> yelling at my computer and cursing the "blue screen of death" <which is
> becomming more and more the norm><hence the cursing> and next win98 flat
> out locking at the bootup screen. I was having to run scan disk more
> than i ran anything else (i have a funky setup with 1 fat partition, and
> 2 fat32 partition which nt uses some bought fat32 drivers... i have had
> problems with it in the past <which i expected> but *never* like this).
> All this time i have not had any problems with linux (well that is not
> true but no related problem that i can tell, certainly not with
> stability). This is not some linux testimonial cause i have my fair
> shar of problems with linux but the linux problems i don't prevent me
> from using apps etc.
>
> why did i post to a o/c group? well i have been running my p2 300 at
> 504 for the past 7 months and have 3 fans running a large cpu fan with
> monster fins and cpu goo to ensure proper convection. According to my
> mother board thermometer (bh6) i have not run over 95 degrees F <though
> i realize that i not always quite acurate>. I have had little to no
> problems with having a o/c'd computer but after 2 reinstallation of each
> windows OS and 2 "re-failures" of each windows OS i <in a moment of
> desparation> tried "slowing" down my computer to 450 mhz. Now it seems
> to be running a bit better though only for 1 day so far (not much of a
> test). My slowing down the clock on my comptuer and then having windows
> seem to would seems to indicat that o/c'ing was giving me a problem (a
> conclusion that i do not want to have to admit) but if linux still
> worked ok could that possibly point to my windows problems *not* being
> o/c'ed related?
>
> any commentary/help with this would be greatly appricated.
>
>
> -Gaiko
>
> Gaikokujin Kyofusho
>
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.
------------------------------
From: Michael Ravits <[EMAIL PROTECTED]>
Subject: Re: CDR
Date: Sat, 28 Aug 1999 10:44:57 +0200
Thanks.
That's what I wanted to hear. :)
Michael
Hans J=F8rgensen wrote:
> =
> Michael Ravits <[EMAIL PROTECTED]> wrote:
> >What would be the best choice CDR to buy? (and a SCSI controller)
> =
> Plextor something and Tekram DC-310U(if your not going to use it for ha=
rddisks)
> I use a Plextor 4220Ti (20x/4x/2x) and the Tekram-controller on a old
> P90 linux-box with 32meg's of ram.
> Works like a dream :O)
>
------------------------------
From: Greg Leblanc <[EMAIL PROTECTED]>
Subject: Re: Finding LAN Bottleneck
Date: Sat, 28 Aug 1999 08:40:35 GMT
In article <[EMAIL PROTECTED]>,
"Andrew J. Norman" <[EMAIL PROTECTED]> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>
> The situation is as follows:
>
> I'm trying to find the bottleneck in a small (half dozen) analysis
node
> cluster. The prime importance of this is to decide which LAN topology
and
> hardware configuration will give the best response for data analysis.
In
> particular it is important to measure the effects of saturating the
LAN in
> communications with diskless processor nodes. With this said let me
> detail the situation and ask for appropriate information....
Sounds pretty slick... will this be running full or half duplex? If
running half duplex, one way to eliminate much of the overhead
assiciated with FTP would be to do a TFTP transfer. FTP uses TCP, which
sends acknolegements when it recieves data. This can greatly slow
traffic on a half duplex connection (only one direction can transmit at
a time). TFTP uses UDP, which does not acknowledge packets, so they
just keep getting dumped to the network. This might be a problem if
your recieving machine isn't as fast as your transmitting machine.
>
> Abstract/Question
> =====================
> For station to station data transfer over a 100BaseTX LAN what is the
> maximum sustainable (station memory to station memory) transfer rate
using
> the FTP protocol?
>
> How can this rate be demonstrated?
>
> Can sustained full saturation of the LAN be demonstrated in station to
> station communications? Detail the method used to show this.
>
> Details
> =====================
> Section 1: The LAN
>
> The test LAN is a standard Fast Ethernet (100BaseTX) running though a
> small (5 port) 100BaseTX Hub with maximum station to station distances
of
> 4 meters (to be reduced eventually). All cabling is CAT5 and cables
have
> been verified for correct operation.
That's always a good way to start off.
>
> Section 2: The NICs
>
> The NICs which I'm trying to analyze for speed bottlenecks are as
> follows:
>
> Linksys LNE100TX (Rev 1)
> Linksys LNE100TX (Rev 2)
> Netgear FA310TX
> SohoWare SFA110A
I haven't used any of these myself, but the tulip driver is very mature
for most os's.
>
> You will note that all of the above cards are DEC Tulip based and
either
> supported by the current tulip driver, or by modifications from the
> manufacture. Additionally the cards are all inexpensive ($10-$30
retail)
>
> Section 3: Test Procedure
>
> The ideal test procedure would demonstrate LAN saturation in station
to
> station communications (and be sustainable) So far the best solution
to
> the challenge that I have been able to demonstrate is a station to
station
> file transfer using standard FTP.
Again, consider TFTP.
>
> Example: consider a disk file "test.img" which has a size on the order
of
> 1gig (remember I want sustained transfer and saturation) residing on
> machine A. The file is transfered to disk on machine B. Rates are
> calculated.
How much ram do these machines have? If you can put 1 gig of ram in
them, this is a good test. The BEST way to eliminate disk or i/o
"interference" is to eliminate disk transfers. This could be
accomplished by having a file that would fit completely in the memory of
the machine. Say you had 512 MB ram in your "server", and 12 was taken
by OS and programs (Not unreasonable, as long as you're not running an
MS os). You could take a 500MB file, transfer it from one machine (the
server) to the other machine, which would put it in the bit bucket
(/dev/null). The first time you ran the test, you'd get bogus results,
due to disk activity, but if you transfered the file again, you should
be testing the memory bandwidth, and the network bandwidth. The memory
bandwidth should be a heck of a lot faster than the network bandwidth,
unless it's a pretty pathetic arcitecture. Your "client" machine would
only need to have enough ram to buffer the file transfer.
The other solution that I can think of (if that is unreasonable) is to
get yourself several UW or U2W scsi disks, and hopefully hardware RAID
(although software MIGHT be fast enough) and create a stripe set of
those drives. I can't explain a stripe set right now, since it's
1:30AM, but perhaps somebody else can, or I'll do that when I'm awake
again. This would be the best alternative to removing the disk all
together. The more spindles you have (disks) the faster you'll be able
to transfer data, up to a point (saturation of the disk bus). I think
that two disks would likely be sufficient, but you might need three,
depending on the capability of the drives.
>
> Using this method the highest station to station rate I can hit is
> 3.5Megabytes/sec or only 28% of theoretic. The method is clearly
flawed
> since the bottle neck on each side is the I/O associated with disk
access
> (the initiating machine utilizes an FastWide SCSI bus but the
recieving
> machine contains a standard PIO Mode 4 IDE drive. While each bus is
> capable of delivering the required 12megabytes/sec, it is not clear
> whether either device can produce can produce a sustained transfer
rate
> this high) In a more extreme example using the same initiating
machine
> but talking with a machine equipped with a UDMA 33 controller (and a
driver
> that should produce the requisite rate) a maximum throughput of only
> 2.4Megabytes/sec (19% of theory) was recorded.
>
> To remove disk I/O on one end the same procedure was used but with the
> destination replaced by the "bitbucket" -> /dev/null.
>
> Using the same initiating station a sustained 5.3 Megabytes (42% of
> theoretic) could be achieved to all stations regardless of disk
> configuration etc....
>
> The problem still is that the disk I/O associated with the initiating
> machine is still folded into the test.
>
> Section 4: Question/Suggestions
>
> The first question that should be asked is "what is the overhead
> associated with the FTP protocol as a percentage of the actual
> throughput?"
>
> Second how can I facilitate a mem to mem transfer over this protocol
(I
> did the obvious thing of creating a FIFO attached to a bit generator
but
> the protocol demands a regular file for it's transfers and thus this
will
> not work) or is there another method/protocol which could be used for
> testing.
>
> As an alternative does anyone know of a good way to deconvolve the
effect
> of the I/O operations so as to really test the network lines and
hardware?
>
> Andrew J. Norman
> ______________________________________________________________
> Dept. of Physics Phone: 757-221-3571
> College of William & Mary [EMAIL PROTECTED]
>
> "It is only with the heart that one can see rightly;
> what is essential is invisible to the eye" -The Little Prince
> ______________________________________________________________
>
> -----BEGIN PGP SIGNATURE-----
> Version: 2.6.3i
> Charset: noconv
>
> iQCVAwUBN8bziyRqxyk0Hh01AQHGzQQAxDb0meaTCu4sowBqARvqNmctIEPQKFFY
> bM6x5PkbI0TSS8A3sJpgQfj6AORHXw6BiO7Doc0/AsrTzwjiJE5nK7EyMFOsKv4g
> tV8mz2DeCWubq/xGONi/vA1kwO9iHmCwRf2P/+ybF0hbibq8CdS2Q7Jh4yau3hM5
> su4FZqPAig8=
> =hME1
> -----END PGP SIGNATURE-----
>
>
If you've scrolled down all the way to here, I'm impressed. Sorry for
doing all that rambling, but it's late, and I'm doing this in between
Cold Fusion development. Oh, BTW, the best OS for this would probably
be something like weight and unix like...
Greg
Almost a guru.
--
It's pronounced "sexy" not "scuzzy"!
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Greg Leblanc <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Dual Pentium II shows as Dual Celeron...
Date: Sat, 28 Aug 1999 08:54:29 GMT
I don't know if you've already asked and got some replies, but the first
thing to check is whether or not CPU EXTERNAL CACHE is enabled or
disabled in the BIOS. If it's disabled, that could cause CPU
misdetection. I can't think of anything else off hand that would even
have a chance of causing it, but I promise to think on it again when I'm
awake.
Greg
In article <[EMAIL PROTECTED]>,
fprintf <[EMAIL PROTECTED]> wrote:
> My brand new dual pentium II system shows as Dual Celeron (Covington)
in
> my /proc/cpuinfo. As requested I am enclosing my cpuinfo and dmesg
output
> in hopes that someone can help identify a) why my CPUs are being
> incorrectly identified and b) why it is showing as having no cache
> available.
>
> I have kernel 2.2.7 from the SuSE 6.1 distribution and I have enabled
SMP
> via a kernel recompile.
>
> Thanks,
> Stuart
> Cheshire, CT USA
>
[snipped to save the net]
--
It's pronounced "sexy" not "scuzzy"!
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Greg Leblanc <[EMAIL PROTECTED]>
Crossposted-To: redhat.hardware.arch.intel
Subject: Re: Adaptec 78xx Hardware RAID and RH6
Date: Sat, 28 Aug 1999 08:48:41 GMT
Well, redhat has been know to often be full of crap. If you configure
the RAID before you start the install, I 'suspect' that you won't have
any problems. I just did an "un-supported" install today. I got a new
DPT smartraid V for a new web server, and I installed RH6 on the RAID
drive without any trouble. The only problem is that XFree has poor
support for the Cirrus 5446 (or something similar) which is the chipset
on most "server" computers from HP, Dell, and Compaq. If anybody
happens to read this with some pointers on how to fix X on that, let me
know.
In article <[EMAIL PROTECTED]>,
Alex Ibrado <[EMAIL PROTECTED]> wrote:
> Hello,
>
> I know that the Red Hat HW compatibility list says:
>
> "We currently do not support hardware raid in any of the Adaptec
AIC78xx
> controllers."
>
> and
>
> "Currently you cannot install Red Hat Linux's root partition onto a
raid
> device."
>
> Unfortunately, I have to work with such hardware and I was wondering
if
> someone out there has gotten HW RAID up and running on the 78xx
series,
> preferrably booting off the RAID as well. Any pointers would be of
help as
> the HOWTOs seem obsolete or not applicable. If not, a "got it going on
> Software RAID" note would also be appreciated. Thanks!
>
> ------------------ Posted via CNET Linux Help ------------------
> http://www.searchlinux.com
>
--
It's pronounced "sexy" not "scuzzy"!
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "Andrew Smith" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: cost of leased line in notting hill, London England
Date: Sat, 28 Aug 1999 09:58:20 +0100
hi
STOP RIGHT THERE! don't go and get any of these lines in a hurry. ever heard
of ADSL - Asymetric Digital Subscriber line. It's a new service from BT,
wherewby it allows VERY high speed data access down a normal phone line so
long as you use a special modem, and your telephone exchange is equipped
with the right hardware. cost is expected to be about £40 per month, with NO
CALL CHARGES.
for more info goto http://www.bt.com/ADSL. The service won't be available
for quite a few months as BT are still equipping the first 400 exchanges,
but if you're in London, you may have had your exhange already converted.
hope this helps,
Andrew
Peter Eddy <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> colin wrote:
> >
> > Could someone very kindly tell me if this seems a reasonable price for a
> > leased line in Notting Hill, London England. Prices are in pounds
sterling
> >
> > £2000 set up
> > £8700 128k
> > £12200 256k
> > £17900 512k
> >
>
> I've heard that European connectivity prices were higher than in the
> U.S. but these prices seem criminal.
------------------------------
Date: Sat, 28 Aug 1999 10:45:38 +0200
From: "erhard.m" <[EMAIL PROTECTED]>
Subject: Aktien zu verschenken!!!! Free Stock Programm!!!!
Crossposted-To:
comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.development.system,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.os.linux.powerpc,comp.os.linux.setup,comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.nt.misc,comp.mail.sendmail
Hallo zusammen,
Die Firma Tradehall (www.tradehall.com) vergibt umsonst Aktien als
Promotion-Aktion. Sie wollen damit Interesse an Ihrem neuen Unternehmen
generieren. (Die letzte Firma, die das gemacht hat, war Yahoo!!!)
Wenn ihr die Seite http://www.tradehall.com/ besucht, müßt Ihr auf
'free stock program' klicken und den Bedingungen zustimmen, oder der
direkte Link:
http://www.tradehall.com/cgi-bin/trader/ms.cgi?run=show_svc&fl=1
dann die persönlichen Infos eintragen (Name, Adresse, ... müssen
angegeben werden; Alter, Ausbildung... sind freiwillig. Login und
Passwort ausdenken und folgende Referenznummer angeben:
10747866
Mit der Anmeldung bekommt Ihr eure Gratisaktie - und eine eigene
Referenznummer! (Für jede Person, die sich über diese Referenznummer
anmeldet, bekommt Ihr noch mal eine halbe Aktie dazu!!! - insgesamt
könnt
ihr bis maximal 100 Aktien sammeln) Anschließend erhaltet Ihr eine
E-Mail. Dort erste URL anklicken und Eure Login, Passwort und
Referenznummer eingeben und bestätigen.
Es ist GRATIS! (der neue Kundenaccount natürlich auch) - solange es sie
noch gibt - das ganze läuft nämlich nur solange, bis tradehall den 1
Millionsten Kunden hat - und davon sind sie nicht mehr allzu weit
entfernt!!
Yahoo ist auch so gestartet. Die Aktie ist inzwischen ca. 140 Euro wert!
Also, auf gehts. Umsonst ist umsonst.
Bis dann
Michi
------------------------------
From: "Jeff Towers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help
Subject: Re: Linux and Ultra DMA/ATA-66 drives
Date: Sat, 28 Aug 1999 05:40:34 -0400
I read that there is a patch and the newest development kernels
now support ATA66. I also read that the "burst" from 33 to 66 may not
be a great improvement since the "burst" is not the bottleneck.
I've decided to move a second hard disk to the ATA33 port and use
that for linux. Unfortunately, it seems the ATA66 hasn't added much.
------------------------------
From: [EMAIL PROTECTED]
Subject: NEW_VER-NewsControlSystem2.1.0_EMailForDemoKey(FreeNEWSservers)
Date: 28 Aug 1999 09:57:39 GMT
(ATT:WEBMASTER)
1.NO WAREZ OR CRACKS ON MY PAGES
2.NO SPAM COMING FROM MY SITES
3.Because Someone Doesnt like this
doesnt justifity Deleting any Page
4.There is no ILLEGAL product or Activity on my site
WhiTeRaBBiT,s News Control System
What Does Do???
This program does a few things:
* Allows you to forge posts (i.e. make other people look as if they
posted something)
* Allows you to make 100% anonymous posts (when used with certain
NNTP servers; installed with system)
* Allows you to post to moderated groups without the approval of
some asshole moderator
* Allows you to cancel others posts
* Allows you to send your own "control" messages (e.g. you can remove
groups or make your own groups)
*Includes large list of free news servers
(search groups in free servers)Never Hunt for another Server
*Read news on one server post or cancel using another
*Mail to News posting using built in EMail Sender
(large list of mail2news gateways)
some of the changes from 2.0.1 in 2.1.0
Added Mail2News List to Mail2News
Added NEW NEWS GROUP SEARCH
Added Change in Code For Better Close
click this ling to download
http://rabbitpro.hypermart.net/
http://wowsites2.com/rabbitpro/
Email me at mailto:[EMAIL PROTECTED]
FOR 10 DAYFREE TRIAL RUN or questions
if you are already a registered owner
download and install over Ver2.0.0
the program will check and install
new serial if it matches the code
For Full Registration and UpGrades and Improvements
If you find this program usefull send
$20.00 Money Order
Leave Pay To Blank
Include Your Registration Name:
Include Your E-Mail Address
You will Recieve Your Full Registration
By E-Mail
Send To
RocksU.com
P.O. Box 468
Zionsville,In
46077
Thank You For Useing A
WhiTeRaBBiT
RaBBiTproDucTionS ProDuct
(ATT:WEBMASTER)
1.NO WAREZ OR CRACKS ON MY PAGES
2.NO SPAM COMING FROM MY SITES
3.Because Someone Doesnt like this
doesnt justifity Deleting any Page
4.There is no ILLEGAL product Activity on my site
------------------------------
** 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
******************************
