Freezing trouble

[Hmse]

Whenever I go into a game such as ut2004 my comp will freeze. This has been happening for quite a long time, and I've done numerous things to get it fixed, replaced the video card, more ram, new hard drive, and even swaped out my power supply. Right now I am running Slackware 10 and absolutly love it, and kernel 2.4.
Now like I said I've had this problem for a long time, about 4 months actually, and I'm really started to think it deffianattly time to get this problem put down the drain.
As I said above, I have replaced alot of hardware to get this problem fixed, the only thing I have not replaced is my cpu. And for a long time I had no dought it was not my cpu cause I had just gotten my mother board replaced a couple weeks before, (before I knew about the problem). But I'm now begining to believe it might possibly be my cpu, partly cause it only crash's under load.

Specs:

Athlon xp 2200
hda - 120 gig
hdb - 40 gig
Nvidia GeForce FX 5700 256 mem
512 ram
300 watt power supply
SlackWare 10
Kernel 2.4.26

Like I said, I have tried for a long time to get this problem fixed, but without luck, so maybe one of you will have some ideas.

[JohnT]

Have you eliminated souncard as a suspect? Could be CPU as its a fairly intensive piece of software.

[xlilo]

is sounds a bit like your cpu but it could also be your software, before you go out and spend loads of money on a new cpu run a test and see what is going on, for example get a program that will run your cpu to the max and if it crashes its your cpu, use a program like "hot cpu", if you cpu is ok then start checking out you software for example the way you have your graphics setup or your xsever or it could even be an error in the operating system.

[paj12]

I would check the nvidia drivers first. I've found that if I don't install the drivers right, my computer will invariably freeze going into an OpenGL app.

Your problem may be different though if it only crashes under high CPU load. You might want to check your cooling too. Your Athlon 2200 and also the 2000 are notorious for crapping themselves to death under high temps.

[Hmse]

Have you eliminated souncard as a suspect?

I have an intergrated sound card.

I'm almost possitive its not my distro, cause I've tried it with multible distros. Now I'm pretty sure I have X setup right, and also my video card configured correctly.

I did a framerates test using glxgears and my rates were awesome, all the way up to 12000fps, but it also caused my comp to crash all three times I ran it.
I have very good airflow, except I did just check, and the cpu is pretty warm, what is the average tempature? How do I check the actual tempature?
Oh and on my last install of the nvidia driver I got a warning saying I have loaded the module rivafb in my kernel, and that it conflicts with the driver. Whats this?

[hyp_spec]

Kernel issue?

2.6 Kernels offer better hardware support, try one of those out for size.

[xlilo]

to test the temp of your cpu you can use lm_sensors,
i once emailed amd to ask them the what the temp should be, i got a reply saying 85c is the death temp but you want to set your temp alarm at 65c. so you dont want it any higher than 65c.

[bwkaz]

Originally posted by Hmse
Oh and on my last install of the nvidia driver I got a warning saying I have loaded the module rivafb in my kernel, and that it conflicts with the driver. Whats this? rivafb is a framebuffer console driver.

If you use an nVidia video card, you MUST not use rivafb if you want to use the nVidia X driver. This is noted in the nVidia driver README (and, I suspect, in the rivafb documentation also). You NEVER want two independent incompatible drivers trying to drive the same piece of hardware, and rivafb is DEFINITELY not compatible with the nVidia X driver.

Use vesafb instead. When compiling your kernel, compile in "VESA VGA graphics support", and make sure you DEselect "nVidia Riva support" (don't even make it a module).

xlilo -- your temperatures are only valid with SOME AMD chips, not all of them. The Barton core seems to be able to run a lot cooler than other AMD chips -- I believe the absolute maximum temperature is still the same, but if you have a Barton core, then I'd start to get worried if you're running above 45-55c (I run at 50c when under full load, when the room is hot also -- if the room cools down or I don't run some high-load programs, then it cools down to about 44c).

My old Palomino core chip ran at about 60-65c when under full load, and its absolute maximum was the same 85c.

[Hmse]

Just checked my tempature, after my comp had been running for over 12 hours, and it was good, 55'C.
I'm working on compiling a 2.6.8.1 kernel right now, but I've been having some problems, I did a make && make modules_install, and it all appeared to go fine except for when it came to modules_install, no I didn't get errors, but it did only take about 2 seconds, and looked nothing like what it should. But I went ahead and moved the bzImage and edited my lilo.conf, then rebooted. It looked fine until it said something like "a non fatal correctable inccident has occured on cpu 1" and then it just sat there.
Did I compile the kernel correct?

[Hmse]

Ok I compiled it again, except before I did make menuconfig I did make mrproper and make clean, and then everything compiled fine, it booted up fine too, but it did give me a warning in the boot up process, it said that the dma on my hard drive is turned off, what is this?
And as for it fixing the freezing trouble I not sure yet, it did let me play for about 20 minutes or so before it exited the game, it gave me the reason, but it looked to be something with the game, not the comp. I'm going do to some more tests and then I'll get back to you all.

[Hmse]

Ok well it didn't fix the freezing. But I'm still wondering if possibly it it still a configuration problem. Is there any other config items that could be causing this?

[Hmse]

I just got the 2.4.26 kernel recompiled without the module that conflicts with nvidia, but have a small (large) problem, when I boot up it says "modprobe: cannot find module net-pf-1, except it doesn't just say that once, it says it like hundreds of time, while booting, and it even continues printing it after I'm booted.

[bwkaz]

net-pf-1 is the alias for the networking module for protocol family 1.

A Google search for net-pf-1 reveals that it's the module that handles Unix filesystem sockets (they're like named pipes, but they're sockets instead of pipes -- they do 2-way communication).

Ensure you have that option either turned on in your kernel or set up as a module. Turning it on would probably be better, since you'll never be able to remove the module anyway (your syslog daemon creates a Unix socket, /dev/log, for other processes to connect to, so the module won't be removable as long as syslog is running).

I don't know where in the 2.4.26 kernel configuration the option is at, but it's named CONFIG_UNIX in 2.6.8. Try looking for that in Configure.help (in the Documentation subdirectory), and then map that section to the area in "make menuconfig" (or whatever you use) that corresponds to it.

[Hmse]

Ok got it fixed. Well not the freezing thing.
Any other clues to the freezing?

[bwkaz]

No, not really (not if you're not using rivafb but it's still happening, that is)... :(

[hard candy]

What motherboard and chipset? "lspci" will give you some info. And going into the bios will give you some more.
Do you boot up windows with this computer? If so, does it freeze with games in windows?

Also, open the case and make sure all the fans are running when you boot up. Sometimes the controller/bridge has a heatsink and overheating there can mess things up.

[ghostwalker]

That is a bit much hardware for a 300w PSU. Try uping it to about 420w PSU.

CPU + 2 HD + VIDCard + MB

Walt

[Hmse]

lspci results:

00:00.0 Host bridge: Silicon Integrated Systems [SiS] 735 Host (rev 01)
00:01.0 PCI bridge: Silicon Integrated Systems [SiS] Virtual PCI-to-PCI bridge (AGP)
00:02.0 ISA bridge: Silicon Integrated Systems [SiS] SiS85C503/5513 (LPC Bridge)
00:02.2 USB Controller: Silicon Integrated Systems [SiS] USB 1.0 Controller (rev 07)
00:02.3 USB Controller: Silicon Integrated Systems [SiS] USB 1.0 Controller (rev 07)
00:02.5 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] (rev d0)
00:02.7 Multimedia audio controller: Silicon Integrated Systems [SiS] Sound Controller (rev a0)
00:03.0 Ethernet controller: Silicon Integrated Systems [SiS] SiS900 PCI Fast Ethernet (rev 90)
00:13.0 USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller] (rev 50)
00:13.1 USB Controller: VIA Technologies, Inc. VT6202 [USB 2.0 controller] (rev 50)
00:13.2 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 51)
01:00.0 VGA compatible controller: nVidia Corporation: Unknown device 0342 (rev a1)
Do you boot up windows with this computer? If so, does it freeze with games in windows?

No I don't duel boot with windows, I'm purely a mac and linux guy, but I have a friend who's gonna give me a copy of windows 98 to try.

Also, open the case and make sure all the fans are running when you boot up. Sometimes the controller/bridge has a heatsink and overheating there can mess things up.

All my fans are running.

That is a bit much hardware for a 300w PSU. Try uping it to about 420w PSU.

CPU + 2 HD + VIDCard + MB

Walt

I am planning on upgrading it, but I don't think thats my problem, a friend of mine has a lan room and he has almost the exact same specs as I, and so does another friend, and neither of them have trouble with freezing.

I just changed a few options in my kernel and am gonna try ut2004 again, will get back to you guys when I get the result.

[Hmse]

Ok ut2004 worked beautiful for about 10 minutes, then quit and totally screwed up X. But it did give me an error messege:

Xlib: extension "XiG-SUNDRY-NONSTANDARD" missing on display ":0.0".
center point Xlib :extension "XiG-SUNDRY-NONSTANARD" missing on display ":0.0"
Signal: SIGFPE [floating point exception]

Aborting

I'm gonna do some google searchs on that and get back.

[Icarus]

Check your ~/.ut2004/System/UT2004.log
There maybe hints in there as to why the system hangs, also /var/log/messages may have some clues also...

[bwkaz]

The XiG warnings are just that, warnings (they're telling you you aren't running an X server from Xi Graphics). The thing that causes the game to not work is the SIGFPE error.

SIGFPE signals are sent when a program performs an illegal floating-point operation (an unmasked floating point exception, basically). Common culprits are using SSE2 when your processor doesn't support SSE2 (e.g. the Athlon XP), or using MMX on a vanilla Pentium which didn't support it (all the "enhanced instruction set"s like MMX, SSE, SSE2, 3DNow!, 3DNow2!, etc., are built on top of the floating point hardware). Or a misbehaving processor, or an over-optimized set of libraries.

[ghostwalker]

There might be simularities between you setup and the other lan computer setup. But is the demand of the software on the hardware the same? That is where I think the issue is.

Walt

[Hmse]

There might be simularities between you setup and the other lan computer setup. But is the demand of the software on the hardware the same? That is where I think the issue is.


At the current I do have more hardware demand, but before when this first started happening the demand was exactly the same.

[Hmse]

I think its fixed. Not entirely sure of what the problem was, but I have a couple guesses.
I unplugged one of my harddrives, and also set my cpu speed down to 100/100, I know it should be able to handle 133/133 but I dunno if it one time got too hot or something. I might do a few more test's and see which it was.

Thanks for the help guys, I may post back here after I do the test's.