[opensuse] 4GB computer slowdown

older
[opensuse] KDE kdeinit $display...

Alexandru Matei

8 May 2007 8 May '07

16:09

Hello, I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that: - although there is some occasional hard disk activity, I never see the KDE Desktop, I only get the mouse pointer and that's it - if I start a console with Ctrl+Alt+F1, it takes about a minute to see the command prompt from the moment I type my (correct) password - if after logging in I use the Midnight Commander, it's response is very slow and slurgish and navigation through folders is very slow When I still had 2GB everything was ok. I'm using openSUSE 10.2 on a Intel Core 2 Duo 1.86 Ghz, 4 MB L2 Cache, FSB 1066 Mhz / Intel DP965LT motherboard / 4 x 1 GB Corsair DDR2 675 Mhz, 4-4-4-12 / ATI Radeon EAX1300 256MB / WDC S-ATA II 250 GB HDD. Booting the XEN kernel (second option in the bootloader) works (but no ATI video driver is taken into consideration, it defaults to VESA mode), but using the default one causes the above problem. Could you please be as kind so as to give me a suggestion to fix this issue? It's really frustrating, as Windows XP and Vista work ok (but they don't see the entire 4GB, only 3.25 GB), but I would rather use Linux instead (and hopefully have all 4GB available). Thank you very much, Alex -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Show replies by date

Matthias Hopf

8 May 8 May

17:34

On May 08, 07 19:09:11 +0300, Alexandru Matei wrote:

...

Could you please be as kind so as to give me a suggestion to fix this issue? It's really frustrating, as Windows XP and Vista work ok (but they don't see the entire 4GB, only 3.25 GB), but I would rather use Linux instead (and hopefully have all 4GB available).

You won't see more than 3.5 (or 3) GB available in Linux, unless you use the 64bit version. That is inevitable. Matthias -- Matthias Hopf __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Aaron Kulkis

18:42

Alexandru Matei wrote:

...

Hello,

I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that:

- although there is some occasional hard disk activity, I never see the KDE Desktop, I only get the mouse pointer and that's it - if I start a console with Ctrl+Alt+F1, it takes about a minute to see the command prompt from the moment I type my (correct) password - if after logging in I use the Midnight Commander, it's response is very slow and slurgish and navigation through folders is very slow

When I still had 2GB everything was ok. I'm using openSUSE 10.2 on a Intel Core 2 Duo 1.86 Ghz, 4 MB L2 Cache, FSB 1066 Mhz / Intel DP965LT motherboard / 4 x 1 GB Corsair DDR2 675 Mhz, 4-4-4-12 / ATI Radeon EAX1300 256MB / WDC S-ATA II 250 GB HDD.

Booting the XEN kernel (second option in the bootloader) works (but no ATI video driver is taken into consideration, it defaults to VESA mode), but using the default one causes the above problem.

Could you please be as kind so as to give me a suggestion to fix this issue? It's really frustrating, as Windows XP and Vista work ok (but they don't see the entire 4GB, only 3.25 GB), but I would rather use Linux instead (and hopefully have all 4GB available).

The Kernel always takes up a certain amount of space, and thus you will never have *ALL* of the memory listed as "available" -- otherwise, the kernel could be overwritten with user programs and/or data....which would definitely be a BAD THING. As to why you're seeing this with 4GB of mem, but not 2 GB... perhaps because you also have swap space... and the kernel is trying to figure out how to manage 4GB + (swap partition) worth of virtual memory all WITHIN a 32-bit address space..

...

Thank you very much, Alex

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Alexandru Matei

19:08

Hello, Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens? Thank you very much, Alex On 5/8/07, Aaron Kulkis wrote:

...

Alexandru Matei wrote:

...
Hello,

I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that:

- although there is some occasional hard disk activity, I never see the KDE Desktop, I only get the mouse pointer and that's it - if I start a console with Ctrl+Alt+F1, it takes about a minute to see the command prompt from the moment I type my (correct) password - if after logging in I use the Midnight Commander, it's response is very slow and slurgish and navigation through folders is very slow

When I still had 2GB everything was ok. I'm using openSUSE 10.2 on a Intel Core 2 Duo 1.86 Ghz, 4 MB L2 Cache, FSB 1066 Mhz / Intel DP965LT motherboard / 4 x 1 GB Corsair DDR2 675 Mhz, 4-4-4-12 / ATI Radeon EAX1300 256MB / WDC S-ATA II 250 GB HDD.

Booting the XEN kernel (second option in the bootloader) works (but no ATI video driver is taken into consideration, it defaults to VESA mode), but using the default one causes the above problem.

Could you please be as kind so as to give me a suggestion to fix this issue? It's really frustrating, as Windows XP and Vista work ok (but they don't see the entire 4GB, only 3.25 GB), but I would rather use Linux instead (and hopefully have all 4GB available).

The Kernel always takes up a certain amount of space, and thus you will never have *ALL* of the memory listed as "available" -- otherwise, the kernel could be overwritten with user programs and/or data....which would definitely be a BAD THING.

As to why you're seeing this with 4GB of mem, but not 2 GB... perhaps because you also have swap space... and the kernel is trying to figure out how to manage 4GB + (swap partition) worth of virtual memory all WITHIN a 32-bit address space..

...
Thank you very much, Alex

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Aaron Kulkis

20:34

Alexandru Matei wrote:

...

Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens?

Wrong. You will never see the entire 4GB *EVER* because the OS does allow it's own memory footprint to be shown as "available memory" And SuSE has been releasing the 64-bit kernels since 10.0..maybe even 9.3.

...

Thank you very much, Alex

On 5/8/07, Aaron Kulkis wrote:

...
Alexandru Matei wrote:

...
Hello,

I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that:

- although there is some occasional hard disk activity, I never see the KDE Desktop, I only get the mouse pointer and that's it - if I start a console with Ctrl+Alt+F1, it takes about a minute to see the command prompt from the moment I type my (correct) password - if after logging in I use the Midnight Commander, it's response is very slow and slurgish and navigation through folders is very slow

When I still had 2GB everything was ok. I'm using openSUSE 10.2 on a Intel Core 2 Duo 1.86 Ghz, 4 MB L2 Cache, FSB 1066 Mhz / Intel DP965LT motherboard / 4 x 1 GB Corsair DDR2 675 Mhz, 4-4-4-12 / ATI Radeon EAX1300 256MB / WDC S-ATA II 250 GB HDD.

Booting the XEN kernel (second option in the bootloader) works (but no ATI video driver is taken into consideration, it defaults to VESA mode), but using the default one causes the above problem.

Could you please be as kind so as to give me a suggestion to fix this issue? It's really frustrating, as Windows XP and Vista work ok (but they don't see the entire 4GB, only 3.25 GB), but I would rather use Linux instead (and hopefully have all 4GB available).

The Kernel always takes up a certain amount of space, and thus you will never have *ALL* of the memory listed as "available" -- otherwise, the kernel could be overwritten with user programs and/or data....which would definitely be a BAD THING.

As to why you're seeing this with 4GB of mem, but not 2 GB... perhaps because you also have swap space... and the kernel is trying to figure out how to manage 4GB + (swap partition) worth of virtual memory all WITHIN a 32-bit address space..

...
Thank you very much, Alex

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hans Witvliet

10 May 10 May

12:06

On Wed, 2007-05-09 at 00:34 +0400, Aaron Kulkis wrote:

...

Alexandru Matei wrote:

...
Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens?

Wrong. You will never see the entire 4GB *EVER* because the OS does allow it's own memory footprint to be shown as "available memory"

And SuSE has been releasing the 64-bit kernels since 10.0..maybe even 9.3.

Some minor points, After changing/installing mem: always have memtest running for over 24 hours. Moving to 64-bit might be a step too big, afaik, no all applications/drivers were 64-safe (wasn't there something with the latest version of flash???) Thought that PAE (physical address extension) was also an solution to get beyond the 4GB-limit -- pgp-id: 926EBB12 pgp-fingerprint: BE97 1CBF FAC4 236C 4A73 F76E EDFC D032 926E BB12 Registered linux user: 75761 (http://counter.li.org) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Jonathan Ervine

11 May 11 May

05:46

Hans Witvliet wrote:

...

On Wed, 2007-05-09 at 00:34 +0400, Aaron Kulkis wrote:

...
Alexandru Matei wrote:

...
Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens? Wrong. You will never see the entire 4GB *EVER* because the OS does allow it's own memory footprint to be shown as "available memory"

And SuSE has been releasing the 64-bit kernels since 10.0..maybe even 9.3.

Some minor points, After changing/installing mem: always have memtest running for over 24 hours.

Moving to 64-bit might be a step too big, afaik, no all applications/drivers were 64-safe (wasn't there something with the latest version of flash???)

Flash is not available in native 64 bit form, so you end up having to use the 32 bit version of firefox etc. to get it to work. It's just a pain to have that requirement for things like Flash, when 64 bit firefox is fine...

...

Thought that PAE (physical address extension) was also an solution to get beyond the 4GB-limit

PAE will certainly let you see, and use, more than 4GB of memory but you can still only address 4GB at any one time. It's a bit of a cludgey (but the only one we've got) workaround to the 32 bit address restrictions. It ends up resorting to tricks similar to mapping in the memory above 4GB when required. For resource intensive applications, 64 bit is the way to go. Jon -- Jonathan Ervine Premium Support Engineer - NTS UK tel: +44 (0) 1344 326 057 mob: +44 (0) 7802 357 042 mob (HK): 9649 5745 SUSE(r) Linux Enterprise 10 Your Linux is ready(tm) http://www.novell.com/linux -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Alexandru Matei

11:18

Hello, I tried a lot of stuff these days and I'd like to share them with you. :) I'll start by saying that I finally managed to install OpenSuse x64. In my previous emails I wrote that the setup would fail at the 'Loading basic drivers...' step. In fact, I was wrong - it didn't fail, it simply slowed down a lot and it I thought the computer had hung (the initial problem I mentioned on this thread therefore repeats itself with the x64 version also). So then I tried to install with the mem=2048M switch and, suprise, it worked. I removed the switch after installing and I started openSuse. This time it slowed down when it reached starting the X server. I added the switch mem=4096M to the grub menu configuration file, and it started and I entered X => memory displayed: around 3.2 GB in the performance monitor (I added the mem in use 500 megs and the available mem of around 2.7 GB - swap was 0). I then installed the ATI proprietary drivers and rebooted and (of course) it slowed down when starting X (even though the switch was present). I then lowered the switch to mem=3072M and all is well, but it doesn't use all the 4GB. So, seeing as the same problem repeats itself and the easiest way to diagnose it (the earliest moment when it can be noticed is when starting the setup program when it slows down just before or during "Loading basic drivers" - i believe that the actual slowdown moment is before this message, though), I have only one question: - What's that special thing that setup does before 'Loading basic drivers' that slows down the computer? :) Thank you for all your help, Alex P.S.: Windows x64 sees all 4GB and it works ok, so I don't think there's anything wrong with the actual memory chips. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Petr Klíma

14:29

Well, it's quite surprising, but Linux is by order of magnitude more sensitive to hardware stability. Memory errors, overclocking and such may cause Linux to produce weird errors while Windows run smoothly. I don't recognize this as a Linux weakness though, it just requires stable hardware. Stable OS run on stable HW, it's that simple :-) I know it's not helpfull, but I have no difficulties running 64bit OpenSuse 10.2 on machines with lots of RAM. I've got several Core 2 Duo Xeons with 32GB RAM with nVidia Quadro FX4500 graphic cards and I haven't experienced any slowdown either during install nor everyday work. They are by far the fastest machines I ever used (comparing e.g. with 3.4GHz P4 with 4GB RAM). I've got 16 core Opteron with 64GB of RAM and I've got no problems as well. There's no reason for 64bit OS to do anything different with 4GB instead of 2,8 or 128GB. The slowdown is most probably caused by your hardware/BIOS setup vs. kernel clash. I don't recall exactly WHEN the "Loading drivers" takes place but I can safely guess that the there is some looooong timeout somewhere within hardware detection section. One example - I experienced such pauses on several Intel workstations - it could take 3+ minutes to detect SATA disks (well, it timed out on unoccupied SATA ports). Different kernel, updated BIOS or even simple BIOS settings change might help (or might not as well...). I would suggest to start with updating the kernel (if possible) as well as BIOS (definitely, makes miracles sometimes). Cheers, Tosuja -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Randall R Schulz

14:47

On Friday 11 May 2007 07:29, Petr Klíma wrote:

...

Well, it's quite surprising, but Linux is by order of magnitude more sensitive to hardware stability. Memory errors, overclocking and such may cause Linux to produce weird errors while Windows run smoothly. ...

I know this is the conventional wisdom, and I don't really dispute it, but I am at a loss to explain why it would be so. Does anyone know where this extra sensitivity to marginal hardware originates? Why would Windows be more tolerant? Do they do something to be more robust (e.g., catch hardware fault exceptions and attempt retries or some such tactic)?

...

...

Cheers, Tosuja

Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Aaron Kulkis

15:10

Randall R Schulz wrote:

...

On Friday 11 May 2007 07:29, Petr Klíma wrote:

...
Well, it's quite surprising, but Linux is by order of magnitude more sensitive to hardware stability. Memory errors, overclocking and such may cause Linux to produce weird errors while Windows run smoothly. ...

I know this is the conventional wisdom, and I don't really dispute it, but I am at a loss to explain why it would be so.

Does anyone know where this extra sensitivity to marginal hardware originates? Why would Windows be more tolerant? Do they do something to be more robust (e.g., catch hardware fault exceptions and attempt retries or some such tactic)?

I think it's more of an observation effect. Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Petr Klíma

18:24

Aaron Kulkis wrote:

...

I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing...

Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason. I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything. Tosuja -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Richard Bos

18:38

Op Friday 11 May 2007 20:24:49 schreef Petr Klíma:

...

Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing...

Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly. From this we learned that linux uses indeed all resources that it has available, while MS does probably not.... -- Richard Bos Without a home the journey is endless -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Aaron Kulkis

19:32

Richard Bos wrote:

...

Op Friday 11 May 2007 20:24:49 schreef Petr Klíma:

...
Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing... Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly.

And upon re-installing windows, it MAGICALLY worked properly with bad memory??? I call bullshit. NO software works properly with bad memory.

...

From this we learned that linux uses indeed all resources that it has available, while MS does probably not....

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

James Knott

19:57

Aaron Kulkis wrote:

...

Richard Bos wrote:

...
Op Friday 11 May 2007 20:24:49 schreef Petr Klíma:

...
Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing... Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly.

And upon re-installing windows, it MAGICALLY worked properly with bad memory???

I call bullshit.

NO software works properly with bad memory.

That would depend on whether parity checking is enabled. If not, you could easily have errors and not know about them. If the bad memory contains executable code, you could have a crash. If data, you might have an error and not realize it. Incidentally, parity memory introduces another failure mode, in that you don't know if the parity bit or parity check circuitry is bad. You only know you're getting errors from somewhere. There's a better system called CRC, which can not only check for errors, but can also correct many. That system should also generate an interupt, so that the OS knows there's bad memory and can notify the admin. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Petr Klíma

21:15

Aaron Kulkis wrote:

...

...
As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly.

And upon re-installing windows, it MAGICALLY worked properly with bad memory???

I call bullshit.

NO software works properly with bad memory.

Let me tell you my story. Again, Windows XP and Linux on the very same machine. Suddenly, Linux applications started crashing while Windows did show no problem. I tracked it down to bad memory too. But this time the erroneous field at one of the top addresses. And as you may know, memory errors come in different flavors and they may show up only in some circumstances. This one was one of those hideous ones. We all know that Linux uses all the memory it has got. Windows don't do that, you can frequently see quite a lot of free memory (while the system accesses swap file which is IMHO brain-dead). Realising that you can tell with quite some confidence that top addresses are accessed much more frequently in Linux than in Windows therefore there's much bigger chance to get error in Linux. Now, software MAY work properly with erroneous memory provided it simply does not access the bad area, either by purpose or just a luck.... Tosuja P.S.: We're getting off topic here. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hans Witvliet

22:18

On Fri, 2007-05-11 at 15:32 -0400, Aaron Kulkis wrote:

...

I call bullshit.

NO software works properly with bad memory.

So you can easily run M$, that never runs properly ;-)) The bugware will crash your system before some mem-fault will cause any decent O.S. to fail... -- pgp-id: 926EBB12 pgp-fingerprint: BE97 1CBF FAC4 236C 4A73 F76E EDFC D032 926E BB12 Registered linux user: 75761 (http://counter.li.org) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Mike McMullin

21:25

On Fri, 2007-05-11 at 20:38 +0200, Richard Bos wrote:

...

Op Friday 11 May 2007 20:24:49 schreef Petr Klíma:

...
Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing...

Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly.

From this we learned that linux uses indeed all resources that it has available, while MS does probably not....

From my experience, Linux has more of the robustness required to run on shoddy hardware, once you get it up and running, than XP does. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

James Knott

21:37

Mike McMullin wrote:

...

On Fri, 2007-05-11 at 20:38 +0200, Richard Bos wrote:

...
Op Friday 11 May 2007 20:24:49 schreef Petr Klíma:

...
Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing...

Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

As example: we once obtained a computer that had been running windows fine for ages. The moment we started installing linux on it, it failed. Indeed it already failed during the installation! Running a memory check tool showed that memory was bad => computer to the IT department, they stated that there was nothing wrong with the system using their tools! After talking a bit longer the faulty memory got replaced and the machine started to behave correctly.

From this we learned that linux uses indeed all resources that it has available, while MS does probably not....

From my experience, Linux has more of the robustness required to run on shoddy hardware, once you get it up and running, than XP does.

Quite so. I have both SUSE 10.2 and XP on my ThinkPad. Linux is very reliable on it, but XP frequently locks up. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

James Knott

19:46

Petr Klíma wrote:

...

Aaron Kulkis wrote:

...
I think it's more of an observation effect.

Windows is so buggy, that when hardware errors do occur, it's just background noise in the all-too-typical crashing and failing...

Well, I don't agree at all. What I experienced was correct Windows behaviour (no errors, at least none reported) while in LInux programs crashed now and then without apparent reason.

I agree with Carlos that LInux most probably uses hardware more aggressively, something like leaving less time between successive actions therefore leaving less time for the things (signal levels etc.) to settle down. That wouldn't be a problem for perfectly stable hardware whose critical operating frequencies are quite higher than the real operating ones (e.g. all transients finished soon enough). Once you got hardware which is operating at (or behind) the edge, you may get anything.

What utter nonsense. Memory read & write, I/O performance etc., are determined entirely by the hardware, with some involvement by BIOS settings. There is nothing that Linux or Windows or any other OS can do, that will change that. Further, interrupts are used by hardware, to tell the OS when it's ready for something else. Back in the DOS, Win3, Win95 & Win98 days, there were some differences in the way the hardware was accessed, i.e. polling, timing loops etc., that a *REAL* OS, such as OS/2 or Linux didn't use. A real OS wouldn't waste CPU time that way. It would use interrupts. Now, if there was a problem in the interrupt system, then it would show on a real OS, but not DOS & Windows. Software can also be written to ignore errors, so if that was done with Windows, but not Linux, then Linux is more likely to find errors and therefore more likely to be reliable. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Carlos E. R.

12 May 12 May

00:58

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Friday 2007-05-11 at 15:46 -0400, James Knott wrote:

...

What utter nonsense.

Memory read & write, I/O performance etc., are determined entirely by the hardware, with some involvement by BIOS settings. There is nothing that Linux or Windows or any other OS can do, that will change that.

That's not completely correct. For instance, if you check the old specifications for the printer parallel port you will see that you need to set up certain signal, then wait for so many milliseconds then activate other signal, or place the data on the bus, etc. If you check specifications for old video cards (I speak of old devices because those are the ones I studied) there are signals that have to be activated at certain times, waits etc, and some of them do not have a sync signal for that precise event. All devices that are no synced to the CPU clock have to take into account external timings. If you peruse the kernel code a bit you will see many places where they are waiting for things to happen. Then there are devices that can be worked in several different ways, and an OS can chose one way, and a different OS another, maybe newer or less tested. And many times they have to find out the ways because they are not documented, and then something unexpected happens. But I can not give specific details because I'm not a kernel developer - though I can guess, I did develop hardware and drivers and I know what I'm talking about. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGRRE+tTMYHG2NR9URAtZdAJ42DbB7IG7nudGN4TkTH3ebNYkLwQCeL9Je T/wfbwv/ArbEARnvNA0jEvk= =bJEc -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Carlos E. R.

11 May 11 May

16:51

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Friday 2007-05-11 at 07:47 -0700, Randall R Schulz wrote:

...

On Friday 11 May 2007 07:29, Petr Klíma wrote:

...
Well, it's quite surprising, but Linux is by order of magnitude more sensitive to hardware stability. Memory errors, overclocking and such may cause Linux to produce weird errors while Windows run smoothly. ...

I know this is the conventional wisdom, and I don't really dispute it, but I am at a loss to explain why it would be so.

Does anyone know where this extra sensitivity to marginal hardware originates? Why would Windows be more tolerant? Do they do something to be more robust (e.g., catch hardware fault exceptions and attempt retries or some such tactic)?

My guess is that linux tries to use all the capacity the hardware should have, in order to improve speed and throughput. This also means that if the hardware fails to do wht it should do, linux will also fail. Probably windows is more conservative, uses a more genereic approach - or they used more faulty harware when testing, or use conservative generic specifications. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGRJ8itTMYHG2NR9URAudoAJsG0ffHpOixRx1aOr6NpOEV5nESPQCfehXm zpF+OxwRB8uReqXTW3K0VaE= =VdZJ -----END PGP SIGNATURE-----

Randall R Schulz

17:03

On Friday 11 May 2007 09:51, Carlos E. R. wrote:

...

The Friday 2007-05-11 at 07:47 -0700, Randall R Schulz wrote:

...
...

Does anyone know where this extra sensitivity to marginal hardware originates? Why would Windows be more tolerant? Do they do something to be more robust (e.g., catch hardware fault exceptions and attempt retries or some such tactic)?

My guess is that linux tries to use all the capacity the hardware should have, in order to improve speed and throughput. This also means that if the hardware fails to do wht it should do, linux will also fail.

That doesn't really explain it, does it? If there's some subset of a machine's capabilities not used by Windows but used by Linux, why would it exactly correspond to that which tends to fail or misfire in all those cases of systems that run without overt error on Windows but cause crashes or malfunctions under Linux?

...

Probably windows is more conservative, uses a more genereic approach - or they used more faulty harware when testing, or use conservative generic specifications.

Possible, but it's still not a well-formed idea. Computers are digital devices. They don't really have any continuous properties, other than, perhaps, some issues around failure probabilities (or inter-arrival times) w.r.t. device temperature. Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Carlos E. R.

18:09

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Friday 2007-05-11 at 10:03 -0700, Randall R Schulz wrote:

...

...
My guess is that linux tries to use all the capacity the hardware should have, in order to improve speed and throughput. This also means that if the hardware fails to do wht it should do, linux will also fail.

That doesn't really explain it, does it? If there's some subset of a machine's capabilities not used by Windows but used by Linux, why would it exactly correspond to that which tends to fail or misfire in all those cases of systems that run without overt error on Windows but cause crashes or malfunctions under Linux?

For example, because as windows does not use them, errors are not reported to the manufacturer. What does it matter if Linux fails? That just a toy OS, not a serious one; they don't pay us. For example, because linux tries to get closer to the limits, uses ignored features that windows doesn't use because it can't (like 32 bits from the start).

...

...
Probably windows is more conservative, uses a more genereic approach - or they used more faulty harware when testing, or use conservative generic specifications.

Possible, but it's still not a well-formed idea. Computers are digital devices. They don't really have any continuous properties, other than, perhaps, some issues around failure probabilities (or inter-arrival times) w.r.t. device temperature.

Yes, but not exact. There are timing constraints of hardware, for instance. If you develop a driver for a piece of hardware by observing how it works, by trial and error - because you know that historically hackers had to hack because manufacturers denied giving their specifications to developers - and timings have to measured or estimated. If then you put that driver in a machine that goes a bit slower on those timings, it will fail. You could ask kernel developers to be more precise on this. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFGRLFytTMYHG2NR9URAiW/AJ4nZ46+AzIlMOizpiWwWqru3usT/ACeJbuc bujLvU5rCje8fEacgxljwbg= =eOCa -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Petr Klíma

18:48

Randall R Schulz wrote:

...

Possible, but it's still not a well-formed idea. Computers are digital devices. They don't really have any continuous properties, other than, perhaps, some issues around failure probabilities (or inter-arrival times) w.r.t. device temperature.

Well, they are presented to the end users like digital devices. But over the time operating frequencies got that high that speed of light matters. You have to take into consideration the time it takes for the impulse to reach the other end of the wire. At 3GHz the light makes only 10cm during one period. In other words, connect a piece of wire 5cm long to the 3GHz sine generator and compare the voltages on both ends. They'll differ by a half-wave... Slow light speed (it's still faster than Terry Pratchett's Discworld light) requires for all those delays in practically any communication in computer. You (well, I mean the circuitry) simply HAVE to wait for the signal to travel from transmitter to receiver. By the way, the real signals are not like square wave too much.... And people still call all THIS digital :-) It's magic to me. Tosuja -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

James Knott

17:18

Randall R Schulz wrote:

...

On Friday 11 May 2007 07:29, Petr Klíma wrote:

...
Well, it's quite surprising, but Linux is by order of magnitude more sensitive to hardware stability. Memory errors, overclocking and such may cause Linux to produce weird errors while Windows run smoothly. ...

I know this is the conventional wisdom, and I don't really dispute it, but I am at a loss to explain why it would be so.

Does anyone know where this extra sensitivity to marginal hardware originates? Why would Windows be more tolerant? Do they do something to be more robust (e.g., catch hardware fault exceptions and attempt retries or some such tactic)?

There should be no difference. There's nothing any OS can do, that works hardware more than another. What may happen though is ignoring errors. -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

eshsf

09:31

Hello, On Thu, 10 May 2007 14:06:06 +0200 Hans Witvliet wrote:

...

On Wed, 2007-05-09 at 00:34 +0400, Aaron Kulkis wrote:

...
Alexandru Matei wrote:

...
Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens?

Wrong. You will never see the entire 4GB *EVER* because the OS does allow it's own memory footprint to be shown as "available memory"

And SuSE has been releasing the 64-bit kernels since 10.0..maybe even 9.3.

Some minor points, After changing/installing mem: always have memtest running for over 24 hours.

Moving to 64-bit might be a step too big, afaik, no all applications/drivers were 64-safe (wasn't there something with the latest version of flash???)

Thought that PAE (physical address extension) was also an solution to get beyond the 4GB-limit

Well, I like to use a lot of 64bit registers(GPRs) and x86 instructions. :-) Thanks, eshsf -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

S Glasoe

8 May 8 May

21:56

On Tuesday May 8 2007 2:08:03 pm Alexandru Matei wrote:

...

Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens?

Thank you very much, Alex

On 5/8/07, Aaron Kulkis wrote:

...
Alexandru Matei wrote:

...
Hello,

I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that:

Did you replace your original 2GB with the 4GB of modules you have now? IF yes, good. If no, then you may have timing issues between the modules that may be causing your lock-ups. Are you keeping the 4GB in matched pairs? Have you swapped the pairs in their sockets? I'm not familiar with your mainboard but have you made sure the BIOS is set for default timings and not for any performance settings? Same for video settings within the BIOS. Have you checked for and applied the latest BIOS for this mainboard? You won't be able to use all 4GB because the BIOS maps out PCI, DMA and other things into the high end of your 4GB. The exact amount reserved is usually mainboard/BIOS dependent. Your swap space/file has nothing to do with this unless your system is somehow using all available RAM and immediately using swap during boot - unlikely. Next would be to downgrade any X.org updates and/or ATI driver updates. Maybe even go back to the basic 2D radeon driver to see if stability returns. -- Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Alexandru Matei

22:09

Hello, No, I've added two more DIMMs, but I've made sure they are exactly the same type as the first two (I now have 4xCorsair DDR2 667Mhz, 4-4-4-12). The BIOS settings are ok, and, since all four memories are exactly the same, I've not tried to swap them. I've also updated the system bios (my version is from 27.04.2007); it is set to default as far as the timings are concerned. I shall try with the ordinary video drivers tomorrow. BTW, the ZEN kernel works - dunno why (but I have no acceleration for the ATI video driver, which I really hate). I've also downloaded an X64 version of the install DVD, but when I try to start the setup, the system hangs at the line: 'Loading basic drivers...' It just stays there forever. I'm beginning to feel sorry I've upgraded the RAM... :( Alex On 5/9/07, S Glasoe wrote:

...

On Tuesday May 8 2007 2:08:03 pm Alexandru Matei wrote:

...
Hello,

Ok, from what I understand, I'll never see the entire 4GB unless I install a 64-bit OS. But I would still like to be able to run openSuse 10.2. :) Does anyone know why the issue below happens?

Thank you very much, Alex

On 5/8/07, Aaron Kulkis wrote:

...
Alexandru Matei wrote:

...
Hello,

I've recently upgraded my RAM to 4GB and since then, when the X server starts, the computer slows down a lot - what I mean by slows down is that:

Did you replace your original 2GB with the 4GB of modules you have now? IF yes, good. If no, then you may have timing issues between the modules that may be causing your lock-ups. Are you keeping the 4GB in matched pairs? Have you swapped the pairs in their sockets?

I'm not familiar with your mainboard but have you made sure the BIOS is set for default timings and not for any performance settings? Same for video settings within the BIOS. Have you checked for and applied the latest BIOS for this mainboard?

You won't be able to use all 4GB because the BIOS maps out PCI, DMA and other things into the high end of your 4GB. The exact amount reserved is usually mainboard/BIOS dependent. Your swap space/file has nothing to do with this unless your system is somehow using all available RAM and immediately using swap during boot - unlikely.

Next would be to downgrade any X.org updates and/or ATI driver updates. Maybe even go back to the basic 2D radeon driver to see if stability returns. -- Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Patrick Shanahan

22:18

* Alexandru Matei [05-08-07 18:10]: [...]

...

It just stays there forever. I'm beginning to feel sorry I've upgraded the RAM... :(

then remove 1/2 and do the install, then replace the ram. ps. Please don't top post. -- Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 OpenSUSE Linux http://en.opensuse.org/ Registered Linux User #207535 @ http://counter.li.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

S Glasoe

22:29

On Tuesday May 8 2007 5:09:56 pm Alexandru Matei wrote:

...

Hello,

No, I've added two more DIMMs, but I've made sure they are exactly the same type as the first two (I now have 4xCorsair DDR2 667Mhz, 4-4-4-12). The BIOS settings are ok, and, since all four memories are exactly the same, I've not tried to swap them.

That's what you'd like to believe. This is high speed memory and being produced at different times there will be differences. Since the graphics seems to be the key to your lock-ups I doubt your main memory is the issue.

...

I've also updated the system bios (my version is from 27.04.2007); it is set to default as far as the timings are concerned. I shall try with the ordinary video drivers tomorrow. BTW, the ZEN kernel works - dunno why (but I have no acceleration for the ATI video driver, which I really hate).

I'm guessing that Xen is using the base radeon 2D driver because I don't believe it supports, or maybe AMD/ATI don't support, 3D on Xen yet (host or guest.?.?).

...

I've also downloaded an X64 version of the install DVD, but when I try to start the setup, the system hangs at the line:

'Loading basic drivers...'

Try failsafe mode for the install or at least acpi=off and see what happens.

...

It just stays there forever. I'm beginning to feel sorry I've upgraded the RAM... :(

You probably are uncovering new bug reports! RAM seems fine. It sounds like the classic video driver hang from everything you are reporting. -- Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Alexandru Matei

22:46

Regarding the issue that Suse 10.2 x64 hangs, I've already tried the failsafe mode to no result, it still hangs. I'll investigate more tomorrow, since I have to get up in the morning to go to work :) Now, the ATI Driver is probably to blame, though it worked fine on the same system when I had only 2 GB of RAM. Probably he doesn't like the extra 2GB or there is a mixup with the memory addressing, since I am currently using the x32 version of suse (the x64 can't be installed, for it hangs :( ). Thanks for all your help, Alex On 5/9/07, S Glasoe wrote:

...

On Tuesday May 8 2007 5:09:56 pm Alexandru Matei wrote:

...
Hello,

No, I've added two more DIMMs, but I've made sure they are exactly the same type as the first two (I now have 4xCorsair DDR2 667Mhz, 4-4-4-12). The BIOS settings are ok, and, since all four memories are exactly the same, I've not tried to swap them.

That's what you'd like to believe. This is high speed memory and being produced at different times there will be differences. Since the graphics seems to be the key to your lock-ups I doubt your main memory is the issue.

...
I've also updated the system bios (my version is from 27.04.2007); it is set to default as far as the timings are concerned. I shall try with the ordinary video drivers tomorrow. BTW, the ZEN kernel works - dunno why (but I have no acceleration for the ATI video driver, which I really hate).

I'm guessing that Xen is using the base radeon 2D driver because I don't believe it supports, or maybe AMD/ATI don't support, 3D on Xen yet (host or guest.?.?).

...
I've also downloaded an X64 version of the install DVD, but when I try to start the setup, the system hangs at the line:

'Loading basic drivers...'

Try failsafe mode for the install or at least acpi=off and see what happens.

...
It just stays there forever. I'm beginning to feel sorry I've upgraded the RAM... :(

You probably are uncovering new bug reports! RAM seems fine. It sounds like the classic video driver hang from everything you are reporting. -- Stan -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

6200

Age (days ago)

6204

Last active (days ago)

List overview

Download

31 comments

14 participants

participants (14)

Aaron Kulkis
Alexandru Matei
Carlos E. R.
eshsf
Hans Witvliet
James Knott
Jonathan Ervine
Matthias Hopf
Mike McMullin
Patrick Shanahan
Petr Klíma
Randall R Schulz
Richard Bos
S Glasoe