Hi folks,
i've upgraded from an AMD Athlon 650 to this system :
AMD Athlon 1,3 GHz, 266 FSB
2*128 MB Kingston SDRAM PC133
Epox 8KTA3+ (VIA KT133A, via686b Southbridge)
Linux 2.4.5-ac27
(Powersupply is 400 Watt, Video is handled by GeForce2 GTS 32 MB)
Compiling the kernel works great, fast and stable. Whenever there is high load
on the systembus (e.g. compiling some source) the noise of my powersupply fans increase.
(Not under Windows, so i think linux makes much more use of the power of AMD's CPUs)
I ran memtest86 for hours without any problems, i couldn't crash the system or any process..
until i start X 4.0.3.
I've tested it with KDE2 and Gnome 1.2. After a (mostly) successful start of one of the
windowmanagers, most applications crush whenever i open them.
Running console-applications in xterm or compiling the kernel in xterm works still stable, but
working with X11 apps is no more fun ... crashes all over ... (segfaults, sig 11)
This message is meant to be a better approach in solving the problem(s) with AMD Athlons/Durons.
I didn't overclock anything. Compiling the kernel with K6-2 Support or PIII fixes the problems
(and the noise of the power-fan stays constant, so i think the cpu-power isn't used to the max :) )
I've monitored my system using lm_sensors and didn't get any unusual high or low values
(cpu is about 42?)
Seems to be the problem with the AMD optimazion in the kernel.
Thomas
> Seems to be the problem with the AMD optimazion in the kernel.
We believe its a VIA chipset problem. Don't expect anything to happen on this
front
Thomas Foerster wrote:
> Seems to be the problem with the AMD optimazion in the kernel.
Funny, I have only had one minor problem with my setup. It's the same
processor, only with one 512 meg PC133 block, and the ASUS A7V133
motherboard (which is equipped with the same chipset). My videocard is also
the same (ASUS V-7700), but my PSU is only 300Mhz.
The only instability I've experienced is when I was running KDM on both vt7
and vt8, and then logging out from X. Sometimes (entirely random) it would
freeze solid (something like once a week). I've had this setup since 2.4.1,
and the kernel has always been compiled with Athlon optimizations, and with
VIA 82CXXX chipset support.
As a sidenote, Linux runs blazingly fast! I was a little worried about the
recent mucking about VIA-chipsets when I ordered my hardware, but those
worries has been put to shame... Or maybe I'm just lucky ;-)
Oh, yeah, my computer is turned on and using cycles some 5-10 hours per
day, right now running 2.4.6.
If there's anything I can help with regarding this VIA-problem, please let
me know. I'm probably not worth anything as a kernel hacker, but I'm
willing to test kernel-patches to see if they cause/cure instability.
Best Regards
Kenneth Vestergaard Schmidt
On Thu, 12 Jul 2001, Thomas Foerster wrote:
> Seems to be the problem with the AMD optimazion in the kernel.
This sounds like a hardware problem -- none of the Athlons I'm using
experience such problems, but then I'm using boards with AMD, ALi and SiS
chipsets, but no VIA. Try updating your BIOS and making your memory
settings more conservative. If that doesn't help, I suggest replacing
your motherboard.
-ben
Hi,
>> Seems to be the problem with the AMD optimazion in the kernel.
> This sounds like a hardware problem -- none of the Athlons I'm using
> experience such problems, but then I'm using boards with AMD, ALi and SiS
> chipsets, but no VIA. Try updating your BIOS and making your memory
> settings more conservative. If that doesn't help, I suggest replacing
> your motherboard.
It's my second Athlon!
(AMD 650, ASUS K7M, 256 MB PC133 => AMD 1300, EPOX 8KTA3ue+, 256MB PC133)
It still doesn't run solid with any of the two machines! (400 Watt Powersupply!)
My BIOS is the latest release.
I've just phoned with Epox here in Germany and they told me, that their boards
are testet with linux and they are working.
I've already tried to slow down the bus, cpu and memory
(254 MHz FSB, 123 SDRAM, 1234 MHz) but this didn't help me out.
NOTE : Things are ONLY crashing when being NOT root!!
If i log in as root i can't get KDE/Gnome apps to crash, only when i'm a
"normal" user! Opening xterm as normal user, su-ing to root and starting
applications works too!
I'm very, very, very confused!
Please help me :-)
Thomas
Hi!
Am Donnerstag, 12. Juli 2001 09:15 schrieb Thomas Foerster:
> AMD Athlon 1,3 GHz, 266 FSB
> 2*128 MB Kingston SDRAM PC133
> Epox 8KTA3+ (VIA KT133A, via686b Southbridge)
> Linux 2.4.5-ac27
> (Powersupply is 400 Watt, Video is handled by GeForce2 GTS 32 MB)
> ...
> I've monitored my system using lm_sensors and didn't get any unusual high
> or low values (cpu is about 42?)
You mentioned, you've got a 400W power supply. But this does not mean you
have enough (available) current on the 3.3V and 5V lines. Or perhaps your
Core voltage is dropping very hard every time your "monster" graphics card
draws a lot of current. Perhaps you should check that (partly possible via
lm_sensors).
I also use a 400W supply with a TB1.4 GHz and I can see a drop on the 3.3V
lines every time I start X on a GeForce 2MX...
Salut, J?rg
> My BIOS is the latest release.
> I've just phoned with Epox here in Germany and they told me, that their boards
> are testet with linux and they are working.
They dont test with Athlon optimisations on . ;)
> NOTE : Things are ONLY crashing when being NOT root!!
Thats important.
> If i log in as root i can't get KDE/Gnome apps to crash, only when i'm a
> "normal" user! Opening xterm as normal user, su-ing to root and starting
> applications works too!
Do you get random crashes or actual logged kernel oopses. Also what X server
> I'm very, very, very confused!
The kernel isnt known for a tendancy to oops according to user id, so me too
Hi,
>> My BIOS is the latest release.
>> I've just phoned with Epox here in Germany and they told me, that their boards
>> are testet with linux and they are working.
> They dont test with Athlon optimisations on . ;)
Seems to be so :)
>> NOTE : Things are ONLY crashing when being NOT root!!
> Thats important.
>> If i log in as root i can't get KDE/Gnome apps to crash, only when i'm a
>> "normal" user! Opening xterm as normal user, su-ing to root and starting
>> applications works too!
> Do you get random crashes or actual logged kernel oopses. Also what X server
I got only one oops in inode.c (forget the actual line number)
The rest are random application crashes on XFree 4.0.3 (GeForce2 GTS, nVidi DRI (older version))
The System NEVER hangs, only applications crash!
(i strace'd konqueror and tried to get some clue why it's crashing ... for me it seems to crash
when bulding list or something like that in memory, if you want, i can attach a strace output)
>> I'm very, very, very confused!
> The kernel isnt known for a tendancy to oops according to user id, so me too
That's what i belivieve in :)
But it's real .. being root (login in as root or suing to root in xterm) prevents applications
from crashing. (doing exactly the same as if i were non-root)
Thomas
Kenneth Vestergaard Schmidt wrote:
>
> Thomas Foerster wrote:
> > Seems to be the problem with the AMD optimazion in the kernel.
>
> Funny, I have only had one minor problem with my setup. It's the same
> processor, only with one 512 meg PC133 block, and the ASUS A7V133
> motherboard (which is equipped with the same chipset). My videocard is also
> the same (ASUS V-7700), but my PSU is only 300Mhz.
>
> The only instability I've experienced is when I was running KDM on both vt7
> and vt8, and then logging out from X. Sometimes (entirely random) it would
> freeze solid (something like once a week). I've had this setup since 2.4.1,
> and the kernel has always been compiled with Athlon optimizations, and with
> VIA 82CXXX chipset support.
I am wondering if you are using the NVidia binary driver for X. They
seem to cause some "funny" things like SIGSEGVs and random hangs. Even
without K7 optimizations.
bye,
Wilfried
Hi,
> Kenneth Vestergaard Schmidt wrote:
>>
>> Thomas Foerster wrote:
>> > Seems to be the problem with the AMD optimazion in the kernel.
>>
>> Funny, I have only had one minor problem with my setup. It's the same
>> processor, only with one 512 meg PC133 block, and the ASUS A7V133
>> motherboard (which is equipped with the same chipset). My videocard is also
>> the same (ASUS V-7700), but my PSU is only 300Mhz.
>>
> I am wondering if you are using the NVidia binary driver for X. They
> seem to cause some "funny" things like SIGSEGVs and random hangs. Even
> without K7 optimizations.
I do. I use the version coming with redhat 7.1
Do they behave different when being non root?
Thomas
> The rest are random application crashes on XFree 4.0.3 (GeForce2 GTS, nVidi DRI (older version))
> The System NEVER hangs, only applications crash!
nvidia dri.
Talk to nvidia, I have so many people reporting problems only with their binary
only stuff its not funny.
On Friday 13 July 2001 13:12, Thomas Foerster wrote:
> I got only one oops in inode.c (forget the actual line number)
> The rest are random application crashes on XFree 4.0.3 (GeForce2 GTS, nVidi
> DRI (older version)) The System NEVER hangs, only applications crash!
So this is off-topic, but anyway here is my experience:
The binary-only Nvidia driver consistently deep-froze my system,
i.e. screen and keyboard were dead. The system log showed, however,
that the system was still alive. Just forget the Nvidia driver...
My 2c,
Stefan
--
Stefan R. Jaschke <[email protected]>
http://www.jaschke-net.de
Thomas, try recompiling your kernel with the 3dnow optimization line
commented out in arch/i386/config.in .
This has solved the crashes for me, but in my case, they are more severe.
I believe Alan has already commented on VIA problems and the problems with
their chipsets.. ahh, found it
"We believe its a VIA chipset problem. Don't expect anything to happen on
this front"
This bums me out. AS i am using ABIT kt7a-raid with the kt133a chipset, and
3dnow kernel optimizations and i oops right as i boot (sometimes before i
complete any init-scripts).
Anyways... I am confirming a problem with my via chipset and 3dnow
optimizations. VIA82CXXX in kernel support not affecting outcome.
Gabriel Friedmann
Let me know if i don't make sense.
> This bums me out. AS i am using ABIT kt7a-raid with the kt133a chipset, and
> 3dnow kernel optimizations and i oops right as i boot (sometimes before i
> complete any init-scripts).
Yep. Its not an ideal situation. People are btw also reporting similar
photoshop oopses in windows. I suspect related problems
> Anyways... I am confirming a problem with my via chipset and 3dnow
> optimizations. VIA82CXXX in kernel support not affecting outcome.
Nope - you want the via82cxxx ide support, that is a big performance win as
you get UDMA IDE. The prefetch/movntq optimisations are less of an impact,
so an Athlon with a non athlon optimised kernel is still a killer devel box
On Sat, 14 Jul 2001, Gabriel Friedmann wrote:
>
> This bums me out. AS i am using ABIT kt7a-raid with the kt133a chipset, and
> 3dnow kernel optimizations and i oops right as i boot (sometimes before i
> complete any init-scripts).
>
> Anyways... I am confirming a problem with my via chipset and 3dnow
> optimizations. VIA82CXXX in kernel support not affecting outcome.
>
mmm, i have also one Athlon (my workstation) with the same MB and
everything works with AThlon optimizzations and 3dnow enabled...
When i got an oops it was because i configured the bios to erogate just
1.7 volt to the processor instead of default 1.75 volts (too hot here,
with 31 degrees it is difficoult to keep cold a processor like this).
I think that FSB speed is the key, at less for some of the VIA chipset. I
have there an Athlon 1300
Mhz with 13x100 setting and FSB of 200 Mhz.
Are you using a processor with 266 Mhz FSB?
(All 266 Mhz FSB Athlon I administer are running on AMD chipset).
bests
Luigi Genoni
Hello all,
On Friday 13 July 2001 13:12, Thomas Foerster wrote:
> I got only one oops in inode.c (forget the actual line number)
> The rest are random application crashes on XFree 4.0.3 (GeForce2 GTS, nVidi
> DRI (older version)) The System NEVER hangs, only applications crash!
Some more experiences with AMD and X-related crashes:
-> I've an Athlon 800 on a Epox EP7KXA Mobo (686A) with 512MB RAM and no
nVIDIA graphics card, but ATI XPERT 2000.
I've got similar problems with 2.4.x-kernels and I would be very glad,
if the cause could be found. The problem is more than half of a year
old!
I posted it in this list some time ago (more than once) - nobody seemed
to be interested. I posted it 2 times to the X bug list - no interest.
But as I can see now, the problem seems to be greater than I thoght.
Therefore I write here some of my latest experiences:
# X 4.1 (if I remember right) and 2.4.6 -> a lot off solid crashes (even
pinging the machine didn't work) while starting X, no matter if DRI was
turned on or off; no matter if agp was loaded or not. Unuseable.
# X-CVS and 2.4.6ac2 eg. is working fine - with DRI turned on - I didn't
test it without DRI.
# Before, I tried to run X 4.1 and ac-Kernels. They have been crashing
too as described above. That's why I'm using X CVS.
# If I try to run X CVS and vanilla 2.4.6, I'm getting blinking screens
after some restarts of X or after long working with X. It's unuseable,
too.
For me, it seems to be not a nVIDIA related problem.
The question for me is:
where are the differences between X CVS / X 4.1 and Vanilla / ac-patches
- and there combinations.
Why is the combination 2.4.6ac2 or the ac-patches before and X CVS
(about 4 weeks old; not DRI-CVS; but they have been merged as far as I
know) working for me without any problems?
Did you try this combination too? It would bee interersting if it would
work for you too!
Regards,
Andreas Hartmann
Hello,
I have a data point that may be of some value. I have 3 athlon systems
running a variety of FreeBSD/NetBSD/Linux and I experience rock solid
stability, even with K7 optimized 2.4 kernels. The things that seem to
be different for me are that:
1) I always use scsi disk subsystems (adaptec or symbios).
2) I do not use nvidia graphics cards.
My boards are Abit (kt7raid, kt7) and Asus (a7v). I have taken no special
power supply precautions beyond using AMD certified units. I'm very curious
as to why I have had such an excellent experience (I'm unlikely to purchase
an intel cpu at this point), even though I use the dreaded via chipsets.
I hope that someone finds this useful. I'd be happy to supply configs or
other details if desired.
skd
On Sun, Jul 15, 2001 at 09:09:20AM +0200, Andreas Hartmann wrote:
> Hello all,
>
> On Friday 13 July 2001 13:12, Thomas Foerster wrote:
>
> > I got only one oops in inode.c (forget the actual line number)
> > The rest are random application crashes on XFree 4.0.3 (GeForce2 GTS, nVidi
> > DRI (older version)) The System NEVER hangs, only applications crash!
>
> Some more experiences with AMD and X-related crashes:
> -> I've an Athlon 800 on a Epox EP7KXA Mobo (686A) with 512MB RAM and no
> nVIDIA graphics card, but ATI XPERT 2000.
>
> I've got similar problems with 2.4.x-kernels and I would be very glad,
> if the cause could be found. The problem is more than half of a year
> old!
> I posted it in this list some time ago (more than once) - nobody seemed
> to be interested. I posted it 2 times to the X bug list - no interest.
>
> But as I can see now, the problem seems to be greater than I thoght.
> Therefore I write here some of my latest experiences:
>
>
> # X 4.1 (if I remember right) and 2.4.6 -> a lot off solid crashes (even
> pinging the machine didn't work) while starting X, no matter if DRI was
> turned on or off; no matter if agp was loaded or not. Unuseable.
>
> # X-CVS and 2.4.6ac2 eg. is working fine - with DRI turned on - I didn't
> test it without DRI.
>
> # Before, I tried to run X 4.1 and ac-Kernels. They have been crashing
> too as described above. That's why I'm using X CVS.
>
> # If I try to run X CVS and vanilla 2.4.6, I'm getting blinking screens
> after some restarts of X or after long working with X. It's unuseable,
> too.
>
>
> For me, it seems to be not a nVIDIA related problem.
>
> The question for me is:
> where are the differences between X CVS / X 4.1 and Vanilla / ac-patches
> - and there combinations.
> Why is the combination 2.4.6ac2 or the ac-patches before and X CVS
> (about 4 weeks old; not DRI-CVS; but they have been merged as far as I
> know) working for me without any problems?
>
> Did you try this combination too? It would bee interersting if it would
> work for you too!
>
>
> Regards,
> Andreas Hartmann
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
Thomas Foerster ([email protected]) wrote :
> Hi,
>
> > Kenneth Vestergaard Schmidt wrote:
> >>
> >> Thomas Foerster wrote:
> >> > Seems to be the problem with the AMD optimazion in the kernel.
> >>
> >> Funny, I have only had one minor problem with my setup. It's the same
> >> processor, only with one 512 meg PC133 block, and the ASUS A7V133
> >> motherboard (which is equipped with the same chipset). My videocard is also
> >> the same (ASUS V-7700), but my PSU is only 300Mhz.
> >>
>
> > I am wondering if you are using the NVidia binary driver for X. They
> > seem to cause some "funny" things like SIGSEGVs and random hangs. Even
> > without K7 optimizations.
>
> I do. I use the version coming with redhat 7.1
Just to clarify : The nVidia drivers shipped with redhat 7.1 ( and any
other version ) are the free ones from xfree86.org and not the proprietary
ones from nVidia.
To avoid confusion maybe people should say 'proprietary' or 'closed source'
instead of 'binary' as free software is 'binary' too.
> Do they behave different when being non root?
>
> Thomas
--
David Balazic
--------------
"Be excellent to each other." - Bill & Ted
- - - - - - - - - - - - - - - - - - - - - -