AMD thunderbird oops

AMD thunderbird oops

[joeja]

I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.  

It is oppsing left and right.  I recompiled the kernel with Athelon as the CPU but keep getting these oopses..

I also get these same problems while trying to install RH 7.1

Anyone know is this a supported processor / MB and has anyone had these problems?

Joe please cc me as I am not on this list.

Re: AMD thunderbird oops

[Alan Cox]

> I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.  
> 
> It is oppsing left and right.  I recompiled the kernel with Athelon as the CPU but keep getting these oopses..
> 
> I also get these same problems while trying to install RH 7.1
> 
> Anyone know is this a supported processor / MB and has anyone had these problems?

Random oopses normally indicate faulty board cpu or ram (and the fault may 
even just be overheating or dimms not in the sockets cleanly). I doubt its
the board design or model that is the problem, you probably jut have a faulty
component somewhere if its oopsing randomly even during installs and stuff

memtest86, and heatsink compound may be your best friends

Re: AMD thunderbird oops

[Alexander V. Bilichenko]

Athlon TB CPUs >= 1000mhz potentially not stable at all because of great
heat-production, without correct hardware maintence - try to check whether
Your cpu overheating or cpu cooler make mighty vibration.

Best regards,
Alexander         mailto:dmor@7ka.mipt.ru
----- Original Message -----
From: "Alan Cox" <alan@lxorguk.ukuu.org.uk>
To: <joeja@mindspring.com>
Cc: <linux-kernel@vger.kernel.org>
Sent: Monday, June 25, 2001 10:55 AM
Subject: Re: AMD thunderbird oops


> > I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz
FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.
> >
> > It is oppsing left and right.  I recompiled the kernel with Athelon as
the CPU but keep getting these oopses..
> >
> > I also get these same problems while trying to install RH 7.1
> >
> > Anyone know is this a supported processor / MB and has anyone had these
problems?
>
> Random oopses normally indicate faulty board cpu or ram (and the fault may
> even just be overheating or dimms not in the sockets cleanly). I doubt its
> the board design or model that is the problem, you probably jut have a
faulty
> component somewhere if its oopsing randomly even during installs and stuff
>
> memtest86, and heatsink compound may be your best friends
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: AMD thunderbird oops

[Dan Hollis]

On Mon, 25 Jun 2001, Alan Cox wrote:
> Random oopses normally indicate faulty board cpu or ram (and the fault may
> even just be overheating or dimms not in the sockets cleanly). I doubt its
> the board design or model that is the problem, you probably jut have a faulty
> component somewhere if its oopsing randomly even during installs and stuff

Ive found a number of problems can be traced to power supply problems, the
board providing a bit of under voltage. Manually adjusting CPU voltage a
notch or two up has fixed many a board for me.

Also manually adjusting SDRAM drive strength can help, BIOS sometimes gets
this wrong.

-Dan

Re: AMD thunderbird oops

[Thomas Kotzian]

as i've said before - i have the same problem too

the memory is OK! - have run memtest86 for hours ... - no errors ... -
heatsink - CPU@45°C - Case @ 25°C

after changing the kernel compile to PentiumII (nearly) everything worked
fine ....

2.4.6pre5 works as PentiumII but not as Athlon
2.4.5ac16 doesn't work at all - random oopses ......

ThomasK.

----- Original Message -----
From: "Alan Cox" <alan@lxorguk.ukuu.org.uk>
To: <joeja@mindspring.com>
Cc: <linux-kernel@vger.kernel.org>
Sent: Monday, June 25, 2001 8:55 AM
Subject: Re: AMD thunderbird oops


> > I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz
FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.
> >
> > It is oppsing left and right.  I recompiled the kernel with Athelon as
the CPU but keep getting these oopses..
> >
> > I also get these same problems while trying to install RH 7.1
> >
> > Anyone know is this a supported processor / MB and has anyone had these
problems?
>
> Random oopses normally indicate faulty board cpu or ram (and the fault may
> even just be overheating or dimms not in the sockets cleanly). I doubt its
> the board design or model that is the problem, you probably jut have a
faulty
> component somewhere if its oopsing randomly even during installs and stuff
>
> memtest86, and heatsink compound may be your best friends
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>
>

Re: AMD thunderbird oops

[Luigi Genoni]

Works well for me, Athlon 1300 (200Mhz FSB).
Did you check your ram?

Luigi

On Sun, 24 Jun 2001 joeja@mindspring.com wrote:

> I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.
>
> It is oppsing left and right.  I recompiled the kernel with Athelon as the CPU but keep getting these oopses..
>
> I also get these same problems while trying to install RH 7.1
>
> Anyone know is this a supported processor / MB and has anyone had these problems?
>
> Joe please cc me as I am not on this list.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: Re: AMD thunderbird oops

[joeja]

Thanks, 
    I have a heat sink and it is huge about 2 inches, plus fan. Plus another 4" fan in the case. (real nice case). 

    I think it is the memory, as yesterday my gcc was bombing with 'internel compiler error', which is usually a good mem tester.  So I started setting mem=64m and things worked better and the install went all the way through.  I think I need to slow my drams down a bit or add some delay in the bios settings.  

   The oops says something like 'kernel null pointer at address 0x000000'.  How do I 'catch' the output of an oops when the filesystem goes and I get ext2fs errors and am forced to reboot and manually run e2fsck?  

    Lastly with the mem=64M or mem=128M when I do a make dep, I get an error message that says Error 'missing seperator'.  What does that mean?  It stops in the drivers/net dir when I get this message?

Thanks
Joe

Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> > I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz FSB) processor  / 512Meg of RAM, and an Asus kt7a motherboard.  > 
> It is oppsing left and right.  I recompiled the kernel with Athelon as the CPU but keep getting these oopses..
> 
> I also get these same problems while trying to install RH 7.1
> 
> Anyone know is this a supported processor / MB and has anyone had these problems?

Random oopses normally indicate faulty board cpu or ram (and the fault may 
even just be overheating or dimms not in the sockets cleanly). I doubt its
the board design or model that is the problem, you probably jut have a faulty
component somewhere if its oopsing randomly even during installs and stuff

memtest86, and heatsink compound may be your best friends

Re: AMD thunderbird oops

[Keith Owens]

On Mon, 25 Jun 2001 12:24:11 -0400, 
joeja@mindspring.com wrote:
>   The oops says something like 'kernel null pointer at address
>0x000000'.  How do I 'catch' the output of an oops when the filesystem
>goes and I get ext2fs errors and am forced to reboot and manually run
>e2fsck?

Read linux/Documentation/oops-tracing.txt.

Re: AMD thunderbird oops

[Thomas Foerster]

Hi,

> as i've said before - i have the same problem too

Me too

> the memory is OK! - have run memtest86 for hours ... - no errors ... -
> heatsink - CPU@45°C - Case @ 25°C

> after changing the kernel compile to PentiumII (nearly) everything worked
> fine ....

I HAD a Asus K7M with a 650 MHz Athlon, 256 MB RAM (Infineon, Ram is OK).
Using 2.4.2 AND 2.4.4 didn't work.
Whenever i started KDE2, the crashes startet, even oopses in ext2-code appeared.

The strange thing is : If i startet KDE2 as root, the crashes didn't happen!

I don't know why, my distribution is RedHat 7.1

Now i have a Epox 8kta3+ with an 1,333 MHz Athlon, FSB266, 256 MB RAM (Infineon)
and the crashes still appear.

Changing the Kernel from Athlon to Pentium-II/III makes the system stable again.

Thomas

Re: AMD thunderbird oops

[Luigi Genoni]

I went trought 8 Athlon, (latest 1300 Mhz 200Mhz FSB).

Usually those stability problems are related to power supply (it should
be at less 300 Watt).
If the power supply does not give enought Ampere, and the energy
is fluttuant, the Athlon cpu really suffers.

I usually do a little overvolt of the vcore to 3.4, and this makes things
more stable. For what i know, the Athlon optimizzation does stress
the memory I/O at best, and does use the Athlon FSB with all its power.

Using the pentiumII/III stuff, you loose this memory performance boost.

If a little overvolt of the VCORE does not solve the things,
(someone also does a little overvolt of the CPU core to 1.78)
then check if your power supply if working well, and the signal strenght
setting inside of your bios (should be 2).

This is just my experience with Athlon processors.

Luigi


On Tue, 26 Jun 2001, Thomas Foerster wrote:

> Hi,
>
> > as i've said before - i have the same problem too
>
> Me too
>
> > the memory is OK! - have run memtest86 for hours ... - no errors ... -
> > heatsink - CPU@45°C - Case @ 25°C
>
> > after changing the kernel compile to PentiumII (nearly) everything worked
> > fine ....
>
> I HAD a Asus K7M with a 650 MHz Athlon, 256 MB RAM (Infineon, Ram is OK).
> Using 2.4.2 AND 2.4.4 didn't work.
> Whenever i started KDE2, the crashes startet, even oopses in ext2-code appeared.
>
> The strange thing is : If i startet KDE2 as root, the crashes didn't happen!
>
> I don't know why, my distribution is RedHat 7.1
>
> Now i have a Epox 8kta3+ with an 1,333 MHz Athlon, FSB266, 256 MB RAM (Infineon)
> and the crashes still appear.
>
> Changing the Kernel from Athlon to Pentium-II/III makes the system stable again.
>
> Thomas
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: AMD thunderbird oops

[Thomas Foerster]

Hi,

> I went trought 8 Athlon, (latest 1300 Mhz 200Mhz FSB).

> Usually those stability problems are related to power supply (it should
> be at less 300 Watt).
> If the power supply does not give enought Ampere, and the energy
> is fluttuant, the Athlon cpu really suffers.

I have a 400 Watt power supply.
With 2.2.x Kernel my System runs stable, but when i'm using whatever 2.4.x kernel i like,
i get the oopses and processcrashes (but only when i'm NOT root!)

Thomas

Re: AMD thunderbird oops

[Alan Cox]

> With 2.2.x Kernel my System runs stable, but when i'm using whatever 2.4.x kernel i like,
> i get the oopses and processcrashes (but only when i'm NOT root!)

That bit is fascinating. Do you do different things as non root ?

Re: AMD thunderbird oops

[Alex Deucher]

I get oopses too when I use kernels compiled for athlon on my redhat
7.1, athlon 850 system.  runs rock solid when I compile for i686.  I
assumed the athlon optimizations in the kernel were broken, or gcc's
athlon optimization was, as I seem to recall some discussion of this a
while back on the LKML.

athlon 850 (200 FSB)
512 MB RAM
IWILL KK266

Alex

----------------------------

I just upgradede my system to an 1200Mhz AMD Athlon Thundirbird (266Mhz
FSB) processor / 512Meg of RAM, and an
Asus kt7a motherboard. 

It is oppsing left and right. I recompiled the kernel with Athelon as
the CPU but keep getting these oopses.. 

I also get these same problems while trying to install RH 7.1 

Anyone know is this a supported processor / MB and has anyone had these
problems? 

Joe please cc me as I am not on this list. 
- 
To unsubscribe from this list: send the line "unsubscribe linux-kernel"
in 
the body of a message to majordomo@vger.kernel.org 
More majordomo info at http://vger.kernel.org/majordomo-info.html 
Please read the FAQ at http://www.tux.org/lkml/

Re: AMD thunderbird oops

[Alan Cox]

> I get oopses too when I use kernels compiled for athlon on my redhat
> 7.1, athlon 850 system.  runs rock solid when I compile for i686.  I
> assumed the athlon optimizations in the kernel were broken, or gcc's
> athlon optimization was, as I seem to recall some discussion of this a
> while back on the LKML.

Most IWILL K266 people report this. Those who swapped them for other boards
mostly report the problem then going away. Yes it could be a bug we trigger
that by chance the IWILL boards show up more than others but I'm sceptical

Alan

Re: AMD thunderbird oops

[Alex Deucher]

What's weird though is that it is rock solid as long as I don't use
athlon optimizations.  I'm not sure how much of a speed improvement they
provide, but everything's fine with i686, so I can't complain, besides I
doubt I can return the board at this point anyway.  BTW, which would be
better with an athlon, k6 or i686 optimization?  I've heard i686 is
faster, but I've never really looked into it too much myself.

Thanks,

Alex

Alan Cox wrote:
> 
> > I get oopses too when I use kernels compiled for athlon on my redhat
> > 7.1, athlon 850 system.  runs rock solid when I compile for i686.  I
> > assumed the athlon optimizations in the kernel were broken, or gcc's
> > athlon optimization was, as I seem to recall some discussion of this a
> > while back on the LKML.
> 
> Most IWILL K266 people report this. Those who swapped them for other boards
> mostly report the problem then going away. Yes it could be a bug we trigger
> that by chance the IWILL boards show up more than others but I'm sceptical
> 
> Alan

Re: AMD thunderbird oops

[Alan Cox]

> What's weird though is that it is rock solid as long as I don't use
> athlon optimizations.  I'm not sure how much of a speed improvement they

That fits the pattern beautifully. 

> provide, but everything's fine with i686, so I can't complain, besides I
> doubt I can return the board at this point anyway.  BTW, which would be
> better with an athlon, k6 or i686 optimization?  I've heard i686 is
> faster, but I've never really looked into it too much myself.

The optimisations are worth several %age points on some benchmarks - the
prefetching memory copier it uses basically hits the full bus bandwidth on
memory copies which rep movs will not do.

My current speculation is that the sdram setup on some of these boards can't
actually take the full CPU spec caused by these hand tuned routines. There is
some evidence to support that as several other boards only work with Athlon
optimisation if you set the BIOS options to 'conservative' not 'optimised'

Equally we don't see the problem on AMD chipset boards and we don't know if
that indicates a bug in the kernel not tripped on such boards or a chipset
problem (even BIOS setuo maybe) on the VIA.

Alan

Re: AMD thunderbird oops

[Dave Jones]

On Tue, 26 Jun 2001, Alan Cox wrote:

> My current speculation is that the sdram setup on some of these boards can't
> actually take the full CPU spec caused by these hand tuned routines. There is
> some evidence to support that as several other boards only work with Athlon
> optimisation if you set the BIOS options to 'conservative' not 'optimised'

Interesting, and plausable theory. It would be more interesting to see
register dumps of the memory timing registers on both good and bad
systems, to see if this is the case.

Unfortunatly afair the register level specs of all the affected chipsets
are not available.

regards,

Dave.

-- 
| Dave Jones.        http://www.suse.de/~davej
| SuSE Labs

Re: AMD thunderbird oops

[Pierre Etchemaite]

Le 26-Jun-2001, Alex Deucher écrivait :
> What's weird though is that it is rock solid as long as I don't use
> athlon optimizations.

Some ASUS boards (mostly P3B-F) would either freeze or self reboot when using
PhotoShop 5. Everything else would run perfectly.

Disabling MMX optimizations in this software would "solve" the problem. Another
solution found on the web (sorry, I don't have the URL at hand) is to add two
or three additionnal capacitors on the back of the board, to solve the electric
instabilities that cause the reboots.

So yes, I can believe that that sort of "corner case" problems exist.

Best regards,
Pierre.


-- 
We are the dot in 0.2 Kb/s

Re: AMD thunderbird oops

[Tim Moore]

> Some ASUS boards (mostly P3B-F) would either freeze or self reboot when using
> PhotoShop 5. Everything else would run perfectly.
> 
> Disabling MMX optimizations in this software would "solve" the problem. Another
> solution found on the web (sorry, I don't have the URL at hand) is to add two
> or three additionnal capacitors on the back of the board, to solve the electric
> instabilities that cause the reboots.

This is incorrect information.  Abit BP6 early revs suffered under load
from a 100uF cap (EC10, between the CPU sockets) that should have been
1500uF.  This was compounded by a weak or otherwise inadequate power
supply.

Having run literally 7 P3F-Fs and 6 of their P2B-F predecessors, not a
single one had any problems.  They were the premiere overclocking boards
of their day.

rgds,
Tim.

--

Re: AMD thunderbird oops

[Pierre Etchemaite]

On 28-Jun-2001 Tim Moore wrote:
>> (I wrote)
>> Some ASUS boards (mostly P3B-F) would either freeze or self reboot
>> when using PhotoShop 5. Everything else would run perfectly.
>> 
>> Disabling MMX optimizations in this software would "solve" the
>> problem.  Another solution found on the web (sorry, I don't have
>> the URL at hand) is to add two or three additionnal capacitors on
>> the back of the board, to solve the electric instabilities that
>> cause the reboots.
> 
> This is incorrect information.  Abit BP6 early revs suffered under load
> from a 100uF cap (EC10, between the CPU sockets) that should have been
> 1500uF.  This was compounded by a weak or otherwise inadequate power
> supply.
> 
> Having run literally 7 P3F-Fs and 6 of their P2B-F predecessors, not a
> single one had any problems.  They were the premiere overclocking boards
> of their day.

Don't get me wrong, I like ASUS boards, and BX chipset boards were a
very safe choice at the time, that's why this problem was very
disturbing when I experienced it on 2 or 3 boards (out of more than 100,
but I guess many never ran Photoshop since then).

I digged my archives.

Adobe Knowledgebase article about the problem
(http://www.adobe.com/support/techdocs/2256a.htm) mentions spontaneous
reboots with Dell Optiplex GX1, ASUS P2B-F and ASUS P3B-F.

I could not find anything about the problem on ASUS websites at the
time.

(unofficial) P2B-(L)S fixing is documented here:
http://www.turbotech.ch/articles2000/000815-p2bls_rework-01.html

(unofficial) CUBX fixing (some of those are also affected):
http://members.ams.chello.nl/mgherard/html/photoshop.html

Best regards,
Pierre.

Re: Re: AMD thunderbird oops

[joeja]

Well considering the other night the power supply went dead, I think that is part of the problem.  It is brand new, and I am being sent another one (free of course).  

I also had my mb loaded at the time (scsi cd-rw, cdrom, internal zip, floppy, 1 hd, Sound card, video, modem, NIC, scsi card) but my last tyan was fine with that load it may be a kt7a thing. 

Several people said that random (keyword here) oopses are more often a hardware thing.  I wonder if the kt7a is going to be able to perform  fully loaded.. 

is anyone running one fully loaded? 4 ide drives, 2 floppy, (5 pci and 1 isa) or 6pci, agp, 512MEG+ RAM?

Joe 

Dave Jones <davej@suse.de> wrote:
> On Tue, 26 Jun 2001, Alan Cox wrote:
> My current speculation is that the sdram setup on some of these boards can't
> actually take the full CPU spec caused by these hand tuned routines. There is
> some evidence to support that as several other boards only work with Athlon
> optimisation if you set the BIOS options to 'conservative' not 'optimised'

Interesting, and plausable theory. It would be more interesting to see
register dumps of the memory timing registers on both good and bad
systems, to see if this is the case.

Unfortunatly afair the register level specs of all the affected chipsets
are not available.

regards,

Dave.

-- 
| Dave Jones.        http://www.suse.de/~davej
| SuSE Labs

Re: AMD thunderbird oops

[Tim Moore]

joeja@mindspring.com wrote:
> 
> Well considering the other night the power supply went dead, I think that is part of the problem.  It is brand new, and I am being sent another one (free of course).
> 
> I also had my mb loaded at the time (scsi cd-rw, cdrom, internal zip, floppy, 1 hd, Sound card, video, modem, NIC, scsi card) but my last tyan was fine with that load it may be a kt7a thing.
> 
> Several people said that random (keyword here) oopses are more often a hardware thing.  I wonder if the kt7a is going to be able to perform  fully loaded..
> 
> is anyone running one fully loaded? 4 ide drives, 2 floppy, (5 pci and 1 isa) or 6pci, agp, 512MEG+ RAM?
> 
> Joe

Similar board (KA7) had non-heat related lockups with 133MHz FSB (1756
BogoM).  100MHz FSB + 4 way interleave has been fast and stable (1690
BogoM).

/dev/hda:
 Timing buffer-cache reads:   128 MB in  1.00 seconds =128.00 MB/sec

Abit KA7 (VT82C686a, rev 22), Athlon 850, 2x256MB PC133@CL2, 2 ide,
CR-RW, Colorado Travan, linux 2.2.20p6+ide.2.2.19.04092001, SPI 300W
power, PCI: Firewire, Netgear FA310TX, Turtle Beach Santa Cruz audio,
AGP: 32MB TNT2.

CONFIG_M686=y
CONFIG_X86_WP_WORKS_OK=y
CONFIG_X86_INVLPG=y
CONFIG_X86_BSWAP=y
CONFIG_X86_POPAD_OK=y
CONFIG_X86_TSC=y
CONFIG_X86_GOOD_APIC=y
CONFIG_1GB=y
CONFIG_MTRR=y

--