Linux-Setup Digest #305, Volume #19 Wed, 2 Aug 00 18:13:17 EDT
Contents:
Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!! (James Knowles)
Re: Why is Athon 650 slower than P-II/400? (James Knowles)
Re: Coppermine SLOW PERFORMANCE... (Vincent Fox)
Re: fwd: SuSE Linux 7.0 released (blowfish)
Re: Newbie IPCHAINS quest (nate)
XFree86 4.0 install error ("Matt & Penny Edmonds")
Re: Coppermine SLOW PERFORMANCE... (Johan Kullstam)
Re: Red Hat 6.2 & sound support: bummer! ([EMAIL PROTECTED])
Re: shared internet ("lekker lekker")
how to cut down /swap to increase / partition please ("Ian Turnbui")
Re: Coppermine SLOW PERFORMANCE... (EKK)
Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!! (Johan Kullstam)
Thanks everyone, BUT > make command not found!!! ("Ian Turnbui")
Can Redhat 6.2 run on Umsdos filesystems(FAT32)?Answer to my E-mail PLEASE.
("Aleksandar Antok")
linux memory limits? ("Matthew N. White")
Re: FWD: Red Hat's CFO abandoning ship.
Re: Coppermine SLOW PERFORMANCE... (Vincent Fox)
Re: Why is Athon 650 slower than P-II/400? (David C.)
Re: Coppermine SLOW PERFORMANCE... (Kent Perrier)
Re: Windows 2000 and RedHat 6.2 (Frank E Harrell Jr)
Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!! (David C.)
Re: Coppermine SLOW PERFORMANCE... (EKK)
----------------------------------------------------------------------------
From: James Knowles <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!!
Date: Wed, 02 Aug 2000 14:31:27 -0600
> > I still think the dual processor setup would be faster than a single
>
> Depends on what you're testing.
The only thing I dare generalize on is that an SMP system can handle a
heavy load from mupliple active processes better. For single-user
systems, there's generally no point.
--
People who reach their potential spend more time asking, "What am I
doing well?" rather than "What am I doing wrong?"
- John C. Maxewll
------------------------------
From: James Knowles <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Why is Athon 650 slower than P-II/400?
Date: Wed, 02 Aug 2000 14:32:36 -0600
> Since you were posting C sources, could you have posted it in a readable
> format?
Bugger! Upgrading the BIOS solved the problem, but just for giggles:
#include <stdio.h>
#include <stdlib.h>
#define MSIZE (96*1024*1024)
#define NITERS 16
int main(void)
{
int i,j;
char* c;
char* p;
c = (char*) malloc( MSIZE );
for( i=0; i<NITERS; i++ )
{
p = c;
for( j=0; j<MSIZE; j++ ) *p++ = 0;
}
free(c);
exit(0);
}
--
Do or do not, there is no try.
- Yoda
------------------------------
From: [EMAIL PROTECTED] (Vincent Fox)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: 2 Aug 2000 20:32:05 GMT
In <[EMAIL PROTECTED]> EKK <[EMAIL PROTECTED]> writes:
>I HAVE SAID THIS BEFORE, BUT DIDN'T GET AS MUCH RESPONSE
>AS I THOUGH I WOULD.
SHOUTING IN ALL CAPS WON'T HELP.
>HAS ANYONE EXPERIENCED BELOW-EXPECTATIONS PERFORMANCE FROM
>COPPERMINE PENTIUM III CHIPS?
Could be, depends on whether the app will fit in the 256K cache.
Maybe you need a Xeon. Maybe you need a smarter compiler and/or
some hand-optimization of your code. Not enough data from you.
--
"Who needs horror movies when we have Microsoft"?
-- Christine Comaford, PC Week, 27/9/95
------------------------------
From: blowfish <[EMAIL PROTECTED]>
Reply-To: ..
Crossposted-To: alt.os.linux.suse,comp.os.linux.misc
Subject: Re: fwd: SuSE Linux 7.0 released
Date: Wed, 02 Aug 2000 13:42:07 -0700
[EMAIL PROTECTED] wrote:
>
> In article <[EMAIL PROTECTED]>,
> .. wrote:
> > [EMAIL PROTECTED] wrote:
> > >
> > > ... and it's bad news! They split the distribution in two versions:
> > > personal (for private, desktop users) and professional (server
> > > related). They are also more expensive, don't have any major release
> > > (neither kernel 2.4, KDE 2.0, etc.) because they're too soon.
> There's
> > > is no reason to upgrade to 7.0. Wait til kernel 2.4 is released ...
> > >
> >
> > From what I've read all over the net today. There're enough
> improvement
> > to justify the upgrade to SuSE 7.0.
> There might be enough improvement (just) to justify 7.0 but not to
> justify the
> huge price increase. The professional edition will be bascically the
> same as the current 6.4 in terms of content i.e 1500+ applications on 6
> CDs but from what I can gather is going to cost more than twice the
> price thats very bad news.
Inflation...
But not really. The supports now has been extented, the "Personal"
version is now cheaper.
Plus SuSE usually gives some real improvement in the extras as well.
I've been using *nix for a decade now. SuSE is the only Linux I have
confident in. It's pretty solid engineered, well polished stuff.
Get the DVD. It contains more than the 6 CD set have.
If you can wait. You can get it at a discounted rate from other vendors.
I'm running SuSE 6.4 right now. I'm not interested in the 2.4 kernel, or
any of the Linux kernel du jour at all. But I'll update the FFS from
their ftp site, or the new SMP patch, if any. I even dumped ipchains
already.
-snipped-
--
- Alex / blowfish.
--
- If Vi is God's editor. Then, God must have too much free time on his
hands,
lives a very dull and unproductive life; so he needs Vi to waste his
time.
But Vi was still too fast. So God created EMACS on the 8th day - which
takes
Eight Months to load, And Counting Still...
KISS rules. That's why I use Easy Edit (ee). Small. Simple and fast.
:-)
- The UN-GEEK CODE:(?What is a
geek?)-#!?+++??++++|$????+++++?????+++!!!!???+++---
geek + vi | ~/emacs
==>ZZZZZZZZZZZZZZZZZZzzzzzzzzzzzzzzz!!!!!!!!!!.......:P~
newbies + Windoz | C:\LOOKOUT
EXPRESS==>_the_horrors_the_horrrrrrrroOOOOORRRRRRRRRSSSSsssss!!! :-|
- My SAS (Sing-A-Song)Fingerprint -v.i007bond: Doe1(-a deer a female
deer.) RaY2(- a drop of golden sun.)
Me3(- A name, I call myself.) FAr4(- A long, long way to run.) Sew5(-A
needle pulling thread.)
lA6(-A note to follow sew.) TeA7(-A drink with jam and bread.) That
will bring us back to DOe-oh-oh-oh...
------------------------------
From: nate <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,alt.os.linux.caldera
Subject: Re: Newbie IPCHAINS quest
Date: Wed, 02 Aug 2000 20:37:38 GMT
An IPChains based soultion would work well for you. I'd check
http://www.linuxdoc.org , and http://www.linuxnewbie.org to get all the info
there for IPChains. You'll also wanna look into ipmasqadm portfw for masqing
packets inside (I'm assuming that you wanna put up something like a web server,
mail server, FTP server, etc, and keep them inside your firewall). If this box
is meant to be just a firewall / router, I agree with the previous post
mentioning that an i486 or low end Pentium (~100MHz) with about 128M RAM would
suffice. Good luck. btw, you may wanna look into Portsentry, and bastille if
this is gonna be a firewall.
nathan
[EMAIL PROTECTED]
Trent Cook wrote:
> What do you mean by, "more power is dangerous for attacks?"
>
> I would like to use masq. Have all internal clients with "fake" ips. Run
> it on a linux box with ipchains to do it all.
>
> Is this not a good solution?
>
> Thanks guys.
>
> Trent
>
> "Leonard Lin" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Aurelien Marchand wrote:
> > >
> > > >Can anyone tell me how many users a linux box with IPCHAINS can handle
> > > >without bogging down too much?
> > > >
> > > >We have about 350 clients.
> > > >
> > >
> > > Hi,
> > > 350 Clients: I suggest using a good i486 and a switch.
> > > I guess even a small-class pentium would make it: Pentium 100 w/ 128Mb
> Ram.
> > > Of course, a linux-based OS. :)
> > >
> > > Salut
> > > Aurelien
> >
> > I agree with you!
> > More power is just dangerous for attacks.
> > But if there is no servers/proxies installed on that engine I don't
> > think that
> > 128 Mb is necessary.
> >
> > sincerely
> >
> > Leo
> >
> > --
> > Leonard Lin
> > M�llerwis 21
> > CH - 8606 Greifensee
> > Switzerland
> >
> > Phone: +41 (0)1 941 40 53
> > Work: +41 (0)1 430 54 50
> > mobile: +41 (0)79 416 01 07
> >
> > Email: [EMAIL PROTECTED]
------------------------------
From: "Matt & Penny Edmonds" <[EMAIL PROTECTED]>
Subject: XFree86 4.0 install error
Date: Thu, 3 Aug 2000 06:40:24 +1000
Hi
I'm running Suse 6.4 and want to install KDE2 so I went to install XF86 4.0.
There were some error messages towards the end of the install but I missed
tham as they scrolled past.
whenI went to startx after that I got the following message ....... can
anyone please help.
Thanks in advance
Matt
XFree86 Version 4.0 / X Window System
(protocol Version 11, revision 0, vendor release 6400)
Release Date: 8 March 2000
If the server is older than 6-12 months, or if your card is newer
than the above date, look for a newer version before reporting
problems. (see http://www.XFree86.Org/FAQ)
Operating System: Linux 2.3.46 i686 [ELF]
Module Loader present
(==) Log file: "/var/log/XFree86.0.log", Time: Wed Aug 2 19:12:27 2000
(==) Using config file: "/etc/XF86Config"
Parse error on line 31 of section XInput in file /etc/XF86Config
"XInput" is not a valid section name.
(EE) Problem parsing the config file
(EE) Error from xf86HandleConfigFile()
Fatal server error:
no screens found
When reporting a problem related to a server crash, please send
the full server output, not just the last messages.
This can be found in the log file "/var/log/XFree86.0.log".
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: 02 Aug 2000 16:27:13 -0400
EKK <[EMAIL PROTECTED]> writes:
> OK,
>
> I HAVE SAID THIS BEFORE, BUT DIDN'T GET AS MUCH RESPONSE
> AS I THOUGH I WOULD.
>
> HAS ANYONE EXPERIENCED BELOW-EXPECTATIONS PERFORMANCE FROM
> COPPERMINE PENTIUM III CHIPS?
> MY NEW PIII850, PIII650 PERFORM ONLY MARGINALLY BETTER THAN
> MY OLD PII450.
ok. what's marginally in numbers?
> CACHE!!!!
>
> IS THIS OR IS THIS NOT AN ISSUE?
perhaps.
> SUPPOSEDLY THE NEW 256KB ON-DIE CACHE IS MORE EFFICIENT, BUT
> PERHAPS ONLY FOR MUNDANE WINDOWS TASKS. IF I AM RUNNING A
> MEMORY-INTENSIVE LARGE PROBLEM THAT IS MOSTLY FLOATING POINT
> OPERATIONS, AM I BETTER OFF WITH THE LARGER CACHE.
> IT SEEMS TO BE THE CASE WITH OTHER PROCESSORS, LIKE MIPS OR
> ALPHA. FOR EXAMPLE THE ALPHA 667MHZ (DP264) HAS A FAT 4MB
> CACHE AND IT IS TWICE AS FAST AS A PIII500(512KB CACHE).
> ALSO, THE MIPS PROCESSORS FREQUENCY IS BELOW PENTIUM FREQ.
> BUT THE LARGER CACHE USUALLY SEEMS TO MAKE UP IN OVERALL
> SPEED.
it's not just the cache, but a mips or alpha have many more registers
than a pentium. while the pentium shuffles its registers around and
spills all over the place, the risc type cpus with many registers keep
on cranking. the ia32 arch just plain sucks.
alpha creams the pentium and athlon but unfortunately it costs so much
more.
btw the amd athlon and especially the athlon thunderbird beat intels
pentia in floating point at the same clock frequency. factor in the
lower cost of the athlon compared to a pentium of similar frequency.
the athlon has a huge floating point performance over price advatange
versus the pentium.
> NOW. I KNOW THE ALPHA IS THE FASTEST OUT THERE AND I AM
> VERY HAPPY WITH IT, BUT I THOUGHT THAT A PIII850 WOULD AT
> LEAST BE 1.5 TIMES FASTER THAN A PII450.
> WHAT IS GOING ON?????
is your memory 1.5 times faster?
is your front side bus 1.5 times faster?
is your hard disk 1.5 times faster?
what is your mobo chipset? BX chipsets are very efficient. i820 with
MTH is horrible. rambus may or may not be a bust compared to sdram.
find your bottleneck. what is your application like? does it process
large randomly accessed arrays or does it crunch on small data sets?
are you use a lot of disk action?
while cpu speeds have been growing by leaps and bounds, the rest of
the system hasn't been growing nearly as fast. if you are ram bound
and have pc100 sdram everywhere, you won't see any increase.
what benchmarks have you run? are they consistent with what you'd
expect? are these benchmarks consistent with your application
results?
> SHOULD I JUST RETURN THESE NEW PROCESSORS AND HUNT FOR AN
> EXTINCT PIII600MHZ WITH THE OLD-STYLE 512KB CACHE????????
that might make sense. it may also be futile.
> PERPLEXED,
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
sysengr
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: linux.redhat.install
Subject: Re: Red Hat 6.2 & sound support: bummer!
Date: 2 Aug 2000 20:51:41 GMT
Redhat works fine with PCI. Some sound cards haven't made it into the
shipped kernel. Whether or not your card is one of them may require a little
research.
You can also try the ALSA sound driver which people seem to have an easier
time with for some cards. see:
http://www.alsa-project.org/
I saw a couple of cards on their supported list with the CMI8738 chipset,
using the cmipci.o module that comes with ALSA.
hth,
Chris
In linux.redhat.install Fausto Arinos de A. Barbuto <[EMAIL PROTECTED]> wrote:
: Hello,
: I have recently installed Red Hat 6.2 on my
: computer. Then it came the time to setup the
: sound devices. sndconfig did a good job as to
: detecting the correct hardware, a CMI 8738
: [/C3DX] PCI audio device. However, no sound
: has ever been heard from the sound boxes.
: sndconfig said it would perform a sound test
: but no sound ever echoed from the multimedia
: speakers. What's wrong? What to do? Can't RH
: 6.2 work with PCI devices?
: Yet another question: during the install, I
: selected a US keyboard with "deadkeys", as
: I wanted accent support for my texts in
: Portuguese. There was a text box to test the
: accents, and they worked fine. But now that
: RH is installed, no one of the text editors I
: have yields neither accents on the vowels (or
: any other character) nor cedils. I tried to
: select a US-Latin keyboard from kbdconfig,
: but that choice was of no use. What to do in
: this case?
: Any help will be much appreciated. Thank
: you!
: ---Fausto
: Sent via Deja.com http://www.deja.com/
: Before you buy.
------------------------------
From: "lekker lekker" <[EMAIL PROTECTED]>
Subject: Re: shared internet
Date: Wed, 02 Aug 2000 20:49:19 GMT
"lekker lekker" <[EMAIL PROTECTED]> schreef in bericht news:...
> hi
>
> I have:
> 2X windows
> 1x corel linux
>
> in a network.
>
> its al linked bij a switch
>
> and I have a internet conection by cable modem, wat is connected bij de
> switch....
>
> I can connect all computers one by one to the internet, but how can I
>
> make linux server, ftp, internet shared bij this computer.
>
> I would like to here tips or how to do,.....
>
> thanx for you time
>
> risj
>
>
------------------------------
From: "Ian Turnbui" <[EMAIL PROTECTED]>
Subject: how to cut down /swap to increase / partition please
Date: Wed, 2 Aug 2000 22:04:35 +0100
Is there an easy way to cut down the swap space to increase the main
partiton?
Ian Turnbull
0961 931941
mailto:[EMAIL PROTECTED]
Web : www.turnbui.freeserve.co.uk
------------------------------
From: EKK <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: Wed, 02 Aug 2000 07:42:22 -0700
Vincent Fox wrote:
>
> In <[EMAIL PROTECTED]> EKK <[EMAIL PROTECTED]> writes:
>
> >I HAVE SAID THIS BEFORE, BUT DIDN'T GET AS MUCH RESPONSE
> >AS I THOUGH I WOULD.
>
> SHOUTING IN ALL CAPS WON'T HELP.
>
> >HAS ANYONE EXPERIENCED BELOW-EXPECTATIONS PERFORMANCE FROM
> >COPPERMINE PENTIUM III CHIPS?
>
> Could be, depends on whether the app will fit in the 256K cache.
> Maybe you need a Xeon. Maybe you need a smarter compiler and/or
> some hand-optimization of your code. Not enough data from you.
>
> --
> "Who needs horror movies when we have Microsoft"?
> -- Christine Comaford, PC Week, 27/9/95
The app. definitely won't fit on the entire cache.
The app. takes up about 205MB RAM. Right?
It is finite element analysis code, so a large matrix and lots
of integration-type FLOATING POINT operations being performed.
>while cpu speeds have been growing by leaps and bounds, the rest of
>the system hasn't been growing nearly as fast. if you are ram bound
>and have pc100 sdram everywhere, you won't see any increase.
PC100 sdram is slowing EVERYTHING down??
I don't know about that. Similar systems (i.e. same mboard, memory,
hard drive) compare pretty linearly in performance increase, amongst
PIIs and 512KB-cache-PIIIs at least. So I have seen a linear increase
in performance when going from PPro200 to PII350 to PII450, but not
once I get into the PIIICoppermines.
AG
--
Alessandro Giachino, Software Engineer
EKK Inc.
2065 West Maple C309 tel. 248-624-9957
Walled Lake MI 48390 fax. 248-624-7158
_____________________________________________
http://www.ekkinc.com
------------------------------
From: Johan Kullstam <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!!
Date: 02 Aug 2000 16:47:22 -0400
James Knowles <[EMAIL PROTECTED]> writes:
> > > I still think the dual processor setup would be faster than a single
> >
> > Depends on what you're testing.
>
> The only thing I dare generalize on is that an SMP system can handle a
> heavy load from mupliple active processes better. For single-user
> systems, there's generally no point.
it depends on the user. i run a lot of simulations and it is nice to
light up both cpus.
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
sysengr
------------------------------
From: "Ian Turnbui" <[EMAIL PROTECTED]>
Subject: Thanks everyone, BUT > make command not found!!!
Date: Wed, 2 Aug 2000 22:27:04 +0100
In the instructions from Techworks and after a 15MB download and installing
binutils & gcc rpm's the next instruction was to type
> make
but the Konsole reports that command not found ???
is this an environment thing or do I need to be in a particular directory or
....
Any help appreciated - I'm nearly there now (he said hopefully, little
realizing the future trials and tribulations - just one more step then I can
really do something useful like Samba ??) Is samba so I
can see shared files / drives on a Windows98 / NT box??? and vice-versa? Is
this going to be a bit of a hassle too? Well, at least I'm learning which if
I didn't have these problems I wouldn't so, Ian less bitching and more
questions (and of course reading the HOWTO's etc)
Thanks to everyone for your help, may it continue.
Ian Turnbull
0961 931941
mailto:[EMAIL PROTECTED]
Web : www.turnbui.freeserve.co.uk
------------------------------
From: "Aleksandar Antok" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux,comp.os.linux,comp.os.linux.hardware
Subject: Can Redhat 6.2 run on Umsdos filesystems(FAT32)?Answer to my E-mail PLEASE.
Date: Tue, 1 Aug 2000 21:17:25 +0200
Can Redhat 6.2 run on Umsdos filesystems(FAT32)?
My hard-drive crashed when I maked Ext2 partition, so this time I want to go
much easier way by installing Redhat 6.2 on FAT32 partition.
Please answer to [EMAIL PROTECTED], not to newsgroup.
Thank You.
------------------------------
From: "Matthew N. White" <[EMAIL PROTECTED]>
Subject: linux memory limits?
Date: Wed, 02 Aug 2000 17:33:39 -0400
Hi,
I have redhat 6.2 running on some machines with
2G memory. For some reason, any job over around
1G is terminated before any more memory can be
allocated. I found the problem occurs in a simple c
program repeatedly calling malloc. Is there some
inherent kernel limit to memory size for a single
job in linux? If so, can it be changed? I'd like
to run a job that can take up the majority of the 2G
memory.
Matt
------------------------------
From: [EMAIL PROTECTED] ()
Crossposted-To: comp.os.linux.misc,gnu.misc.discuss
Subject: Re: FWD: Red Hat's CFO abandoning ship.
Date: Wed, 02 Aug 2000 21:34:46 GMT
On Wed, 02 Aug 2000 12:33:02 -0700, blowfish <[EMAIL PROTECTED]> wrote:
>John Hasler wrote:
>>
>> Christopher Browne writes:
>> > ...as well as under a license whereby Be Software pays me $500,000, and
>> > then is allowed to include a GnomoVision DVD in their boxed sets of BeOS.
>>
>> On the other hand, had Christopher released GnomoVision under the BSD
>> license, Be would already be allowed to include GnomoVision in their boxed
>> sets of BeOS under a proprietary license without paying him a penny or
>> revealing the source to their enhancements to him or anyone else.
>>
>> The fact that the GPL does not permit this seems to be what enrages the BSD
>> trolls: they are evidently offended that we are not willing to make them a
>> gift of unlimited rights to our code.
>> --
>> John Hasler
>> [EMAIL PROTECTED]
>> Dancing Horse Hill
>> Elmwood, Wisconsin
>
>I think the DVD under Linux issue will soon be a moot point.
>
>I've read from The Register from the U.K. that SiS will have
>native/hardware DVD support for Linux in its up-coming video chip set.
This from the manufacturer who's product is pretty much
a 'quick run for your life' signal to Linux running users...
[deletia]
Vapour I say! Vapour...
There was an announcement from InterVideo months ago and still
nothing has come of it...
--
Finding an alternative should not be like seeking out the holy grail.
That is the whole damn point of capitalism.
|||
/ | \
------------------------------
From: [EMAIL PROTECTED] (Vincent Fox)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: 2 Aug 2000 21:33:34 GMT
In <[EMAIL PROTECTED]> EKK <[EMAIL PROTECTED]> writes:
>> Could be, depends on whether the app will fit in the 256K cache.
>> Maybe you need a Xeon. Maybe you need a smarter compiler and/or
>> some hand-optimization of your code. Not enough data from you.
*snip*
>The app. definitely won't fit on the entire cache.
>The app. takes up about 205MB RAM. Right?
>It is finite element analysis code, so a large matrix and lots
>of integration-type FLOATING POINT operations being performed.
Let me explain. Most programs spend a lot of time in a small
amount of actual code. It doesn't matter how much data space
in total you are talking about, most programs can use a 256kb
or 512kb cache and run very well. You miss sometimes and then
you do some cache loading. But overall, cache is a win.
Ideally you would have cache = RAM. 2 problems though:
1) Too deuced expensive and complex to have giganto cache
2) Usually not neccessary. Let me give an example.
I had some CFD (Computational Fluid Dynamics) code that I ran on
a Celeron 366 overclocked to 550. It ran about 8% faster on a
Xeon 550 system but consider the cost diff this hardly seemed
a worthwhile improvement.
>>while cpu speeds have been growing by leaps and bounds, the rest of
>>the system hasn't been growing nearly as fast. if you are ram bound
>>and have pc100 sdram everywhere, you won't see any increase.
>PC100 sdram is slowing EVERYTHING down??
>I don't know about that. Similar systems (i.e. same mboard, memory,
>hard drive) compare pretty linearly in performance increase, amongst
>PIIs and 512KB-cache-PIIIs at least. So I have seen a linear increase
>in performance when going from PPro200 to PII350 to PII450, but not
>once I get into the PIIICoppermines.
You misunderstand. We don't know what is going on with your systems.
You don't give enough details about your hardware or your applications.
Lacking that, many will tend to assume this a system configuration issue.
It might be, it might not.
In my experience a PIII 650 at 866 really is about twice as
fast on for example a large Matlab run than my previous PII-450.
Same results when making big runs of Fluent, a commercial CFD
application. So I'd say the CPU can deliver in terms of running large
matrix calculations using fat apps like Matlab. What is up with your
particular situation no one can say without close scrutiny.
The CPU would be my last suspect.
--
"Who needs horror movies when we have Microsoft"?
-- Christine Comaford, PC Week, 27/9/95
------------------------------
From: [EMAIL PROTECTED] (David C.)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Why is Athon 650 slower than P-II/400?
Date: 02 Aug 2000 17:40:17 -0400
James Knowles <[EMAIL PROTECTED]> writes:
>>
>> Since you were posting C sources, could you have posted it in a readable
>> format?
>
> Bugger! Upgrading the BIOS solved the problem, but just for giggles:
>
> #include <stdio.h>
> #include <stdlib.h>
>
> #define MSIZE (96*1024*1024)
> #define NITERS 16
>
> int main(void)
> {
> int i,j;
> char* c;
> char* p;
>
> c = (char*) malloc( MSIZE );
>
> for( i=0; i<NITERS; i++ )
> {
> p = c;
> for( j=0; j<MSIZE; j++ ) *p++ = 0;
> }
>
> free(c);
>
> exit(0);
> }
Thanks.
-- david
------------------------------
From: Kent Perrier <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: 02 Aug 2000 16:34:28 -0500
EKK <[EMAIL PROTECTED]> writes:
> OK,
>
> I HAVE SAID THIS BEFORE, BUT DIDN'T GET AS MUCH RESPONSE
> AS I THOUGH I WOULD.
<SNIP>
Now my ears hurt :)
Kent
--
They were killing machines. I wanted to be them so bad.
-- Jamie Zawinski
http://www.jwz.org/gruntle/grave.html
------------------------------
From: Frank E Harrell Jr <[EMAIL PROTECTED]>
Subject: Re: Windows 2000 and RedHat 6.2
Date: Wed, 02 Aug 2000 21:42:34 GMT
Daniel Waites wrote:
>
> I've just upgraded to Windows 2000 and now I can only boot Linux from a
> floppy. These are the previous and current setups:
>
> Previously:
> Windows 98 is installed on the first hard drive.
> RedHat 6.2 is installed on the second hard drive. LILO is installed in the
> superblock of the root partiton, not in the MBR.
> PQ Bootmagic is installed on a small primary partition on the first hard
> drive, which selects which OS to load.
>
> Currently:
> Windows 2000 is installed on the first hard drive.
> As far as I know, Bootmagic is still there, but its partition is no longer
> marked active in the partition table.
> RedHat 6.2 is still on the second hard drive.
>
> I am using Bootmagic from the 4.0 release of PartitionMagic. As I
> understand, there was no Windows 2000 support in that release. To make
> matters worse, PowerQuest no longer will support 4.0.
>
> How do I get Bootmagic working again? Or is there an alternate way to get
> Windows 2000 and Redhat 6.2 to dual-boot? I can't use LILO in the MBR,
> since it's incompatible with the Win2000 boot loader and it can't mount a
> boot partition from the second hard drive.
>
> Thanks is advance.
> Daniel Waites
> [EMAIL PROTECTED]
My only solution, before I found that Linux did everything I needed
and I deleted Win2k last week, was to boot Linux from floppy and
Win2k from the hard disk. I tried everything else, including
handing WIn2k the linux boot sector to boot.ini - nothing worked.
If you had installed Win2k's file system as FAT32 bootmagic would
have worked. I found that out the hard way.
--
Frank E Harrell Jr Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine http://hesweb1.med.virginia.edu/biostat
------------------------------
From: [EMAIL PROTECTED] (David C.)
Crossposted-To: comp.os.linux.hardware
Subject: Re: Why is Athon 650 slower than P-II/400? !!!SOLVED!!!
Date: 02 Aug 2000 17:48:42 -0400
James Knowles <[EMAIL PROTECTED]> writes:
>>>
>>> I still think the dual processor setup would be faster than a single
>>
>> Depends on what you're testing.
>
> The only thing I dare generalize on is that an SMP system can handle a
> heavy load from mupliple active processes better. For single-user
> systems, there's generally no point.
A few years ago, I'd say that.
Today, however, most people I know run multiple processes at once.
On my home system (dual PPro/200), I routinely have running:
- TiK (an instant-messenger client)
- Netscape (for web and mail)
- trn (for reading news)
- Emacs
- At least one shell window
- RealPlayer (if my web browsing takes me to a video clip)
- XFree86 (plus a window manager)
- A bunch of daemons (like sendmail)
Most of the time, the second processor probably doesn't contribute much,
but when one of them decides to start using a lot of CPU, having the
other one around keeps the system responsive.
-- David
------------------------------
From: EKK <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Coppermine SLOW PERFORMANCE...
Date: Wed, 02 Aug 2000 08:22:33 -0700
Kent Perrier wrote:
>
> EKK <[EMAIL PROTECTED]> writes:
>
> > OK,
> >
> > I HAVE SAID THIS BEFORE, BUT DIDN'T GET AS MUCH RESPONSE
> > AS I THOUGH I WOULD.
>
> <SNIP>
>
> Now my ears hurt :)
>
> Kent
> --
> They were killing machines. I wanted to be them so bad.
>
> -- Jamie Zawinski
> http://www.jwz.org/gruntle/grave.html
What???
AG
--
Alessandro Giachino, Software Engineer
EKK Inc.
2065 West Maple C309 tel. 248-624-9957
Walled Lake MI 48390 fax. 248-624-7158
_____________________________________________
http://www.ekkinc.com
------------------------------
** 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.setup) 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-Setup Digest
******************************
