Strange crashing problem on new computer

Associate
Joined
16 Sep 2006
Posts
15
Hi,

I built the following machine about a month ago:
£114.95 x 1 - Intel Core 2 DUO E6300 "LGA775 Alendalle" 1.86GHz (1066FSB) - OEM (CP-135-IN)
£89.95 x 1 - Gigabyte GA_965P_DS3 (Socket 775) PCI-Express DDR2 Motherboard (MB-061-GI)
£129.95 x 1 - Connect3D ATI Radeon X1800 XT 256MB GDDR3 AVIVO TV-Out/Dual DVI (PCI-Express) - Retail (GX-044-CO)
£62.95 x 1 - Seagate Barracuda 7200.10 320GB ST3320620AS SATA-II 16MB Cache - OEM (HD-078-SE)
£79.95 x 1 - Antec P180 Advanced Super Midi Tower Case - No PSU (Black) (CA-049-AN)
£28.95 x 1 - Zalman CNPS9500-AT Aero Flower (Socket 775) CPU Cooler (HS-020-ZA)
£59.95 x 1 - Enermax Liberty 500W ELT500AWT ATX2.2 Modular SLI Compliant PSU (CA-025-EN)
£124.95 x 1 - Corsair 2GB DDR2 XMS2-6400C5 TwinX (2x1GB) (MY-092-CS)

The machine worked fine for a few weeks, however then I began to notice frequent yet intermittent crashes when playing some games (Battlefields 2 and Half-Life 2 are examples). These were generally VPU recovery errors, where the GPU seemed to be able to crash independently of the rest of the system.

I immediately assumed this was a problem with the graphics card, and replaced it with the following:
£129.95 x 1 - HIS ATI Radeon X1900 GT ICEQ 3 SILENT Heatpipe 256MB GDDR3 AVIVO TV-Out/Dual DVI (PCI-Express) - Retail (GX-057-HT)

This didn't resolve the problem, it carried on exactly as described. I've also replaced the motherboard with a different unit of the same model, tested both ram sticks individually, and tried a different power supply. None of these steps helped.

I'd tried both the 6.8 and 6.9 release of the Catalyst drivers, but it had never occured to me to try the 6.7 release (which I had been using originally when the crashes didn't occur). After installing the 6.7 release last night the problem seems to have become either much reduced or gone altogether (I need to do more extensive testing to be sure). I find this extremely suprising that this might resolve my problem, since searching on the internet doesn't reveal anyone else actually experiencing it.

I've tried various chipset drivers, BIOS revisions, reinstalling Windows, disabling sound altogether in the BIOS, I can't really think of what else to do.

Does anybody know why this is the case, whether the card is faulty (it's the second card I've tried though...), and if the card isn't faulty whether ATI will fix the drivers so I can use the latest release?

Many thanks for any help you can give, this has been driving me crazy for weeks!

Rob.
 
Last edited:
I have run memtest86 for a few passes, on the sticks together and individually but this didn't reveal any errors.

I have left all the timing at their default values, however have increased the voltage from the default 1.8v (of the motherboard) to 2.1v.

Do you think changing the timings from their default settings might help, and what would I change them to?
 
Thanks for the ideas, I think I'll have a look into manually setting my memory timings.

It's strange if this is a driver issue, since a Google search doesn't reveal any people with the same problem. With my old nVidia card I had there were a few such issues, however searching on Google normally found many many people with the same problem as myself!

Cheers,

Rob
 
Admiral Huddy said:
You havn't said if you cleaned the drivers and CAB before installing the new ATI drivers.

Just un-installing the NVIDIA drivers then installing the new ATI drivers isn't really recommended and this could be the problem.. You could have a few conflicting left overs... possibly..
Sorry I don't think I made myself clear, I was just talking about my experiences with a previous system I built many years ago, the system has never had an nVidia card in it.

As for cleaning drivers etc., I've reformatted and reinstalled Windows a couple of times which is effectively the same, though I will try your suggestion anyway for the sake of thoroughness!

Cheers!
 
It has been some time since my original post, however my issue still remains.

My system seems stable when using the Catalyst 6.7 driver release, but on trying 6.8, 6.9, 6.10, 6.11, 6.12, 7.1, 7.2, 7.3, 7.4 I get frequent VPU recovery events on many of my games (e.g. ‘Half-Life 2’ and other 'Source' based games, ‘FarCry’, ‘Battlefield 2’ and ‘Supreme Commander’ to name a few.)

It seems strange that I can't use ANY of these newer driver releases, and I don't seem to be able to find anyone else experiencing the same issues as me (I search for this frequently!).

Does anyway have any ideas on how to fix this?

Many thanks.

Rob
 
Thank you for your reply, this problem is starting to drive me a little crazy!

I've always assumed it wasn't power related, but I've just tried Supreme Commander which seems to quite quickly produce these VPU recovery events, while logging both temperatures and voltages.

At the time of the crash there are no obvious spikes on any of the readings, here are the maximum temperature readings for about a minute before and after the VPU recovery event (none of these changed by more than 2C in this time):

Motherboard ~43C
CPU ~42C
CPU Cores #1,#2 ~45C
GPU ~69C
GPU Ambient ~53C
GPU VRM ~66C

Also voltages:
CPU Core: 1.33V, occasionally fluctuating to 1.31V
+3.3V: 3.31-3.33V
+12V: 12.19V, occasionally fluctuating to 12.12V
+5V Standby: 5.11V-5.19V
DIMM: 1.9-1.92V

I’m not sure what I’d be looking for, but to me all that looks pretty normal.

What I have done:-
1. Switched a Radeon 1800XT to a 1900GT.
2. Replaced my Gigabyte DS3 motherboard (with the same).
3. Ran memtest86+ on both ram sticks together, and individually (no errors). I have also tried games with both sticks individually, but VPU recovery events still occur.
4. Tried a variety of motherboard drivers, sounds drivers and bios revisions, including the latest of all of these.
5. Ensured all cables are connected properly, re-seated the graphics card a number of times.

None of the above has made any difference to my problem.

Originally I said that no errors occurred with the 6.7 driver release, however using these over the last 6 months has produced a few VPU recovery events in Windows (strangely in Adobe Acrobat), while 3D games appear stable.
Using 6.8 onwards produces frequent VPU recovery events in most popular 3D games, of varying frequency though generally after 5-10 minutes of play and then much more often (around every 2 minutes).
I read the release notes of each new Catalyst release, though have seen nothing that obviously relates to the issue I’m experiencing. Any suggestions would be much appreciated!

Cheers,

Rob
 
Thank you for your reply.

Temperatures and voltages in my last post were from the log file Everest can be set to produce, whilst I was playing Supreme Commander which was producing a VPU recovery error. There doesn't appear to be any deviation in the voltages at the time of the crash.

I've stress tested the CPU using Orthos for a number of hours (with priority level 9) and the system seems stable. The system is often used for a variety of video encoding tasks which stress the CPU and these never crash (though the software I use only uses one core, but Orthos stresses both anyway).

Unfortunately I don't have access to a multimeter or a PSU compatible with my system.

Using the 6.7 release of Catalyst seems to be my only option at the moment, but it would be nice to use newer releases as there are fixes in the newer releases which are of interest to me. Like you said above, I'm at a loss as to why I seem to be the only person experiencing this problem (I've searched for hours on Google, at least a few times a month...), since my problems are *only* with release 6.8 onwards I'd have thought someone else might be having the same problem if it were a driver issue.

Is there anything else I can try to eliminate without trying a different PSU?

Cheers,

Rob
 
I mean I'm pretty sure the CPU is fine, which only leaves the power supply (perhaps the card is drawing more power with the newer driver releases)?

The 500W power supply should be enough for my system; I'm only using one hard disk. I'm not sure how a fault with the power supply would manifest itself, though issues do only occur when the graphics card seems to be under full load (and I assume drawing the most power, however would a different driver release be enough to tip my system over the edge?).

Rob
 
Yeah both connectors are connected to the motherboard.

I've replaced my motherboard the same model already, so I am assuming this isn't the cause.

I've tried every BIOS revision as it has been released, these
don't seem to effect the issue. I've tried all new versions of the motherboard/chipset drivers also. Release notes for both of these don't mention any issues such as mine.

I've been using the chipset drivers off the Gigabyte website for my motherboard, I might download them directly from Intel and see if that makes a difference, but I think I might have already tried this a while ago.

Cheers,

Rob
 
After further testing, it would appear after the first boot on installing the latest drivers and then running the game Supreme Commander resulted in some strange graphical corruption. This was at a similar time to when a VPU recovery event would have occured, I get the feeling that this is what happens if the 'feature' fails to kick in:

http://robaj.sent.com/untitled.JPG
http://robaj.sent.com/untitled1.JPG

Upon rebooting the system, it fell back into its usual ways of frequent VPU recovery errors, but sometimes all out system crashes with repeating sounds in place of these.

I should probably have mentioned this earlier but I have experienced this problem before, again on the first reboot after installing a new driver release, but with the game Oblivion. I get the feeling corruption such as this could not be caused by the power supply, I mean surely it should just point at the graphics card but I've already replaced that!

I was considering upgrading to Windows Vista, however I'm not sure what this issue will mean if I make that change (i.e. will there be *any* drivers which I can use, or will all of them give me this problem).

Does anyone have any ideas given the above screenshots?

Many thanks.

Rob
 
I bought this all from overclockers.co.uk in August 2006, so it might be a little late for me to RMA it all...

That said, I've already RMA'd the motherboard (for the same model) and the graphics card (for a different model), and at the end of it all I'm not even certain either were faulty (though they both were tested as faulty).

Upon finding that the 6.7 drivers resolved my issue, I immediately assumed that this was in fact a driver issue, believing that it would be fixed in one of the newer releases. Now over 6 months has passed and this is looking unlikely.

Connectors are all connected properly (to my knowledge), and I'm using the newest chipset, BIOS, sound and the newest video drivers. I have also tried various combinations of the above, however the *only* thing which causes the problem is the use of Catalyst release 6.8+.

I've stress tested the CPU fairly extensively, as for running memtest86+ it was a while ago when I did this. I believe I let it run for at least 2 cycles with both sticks together, and then the same for both individually also.

I haven't messed much with memory timings, currently it's being set by SPD. I have however changed a setting in the BIOS to ensure that the memory is receiving the full required 1.9V.

Currently I'm living in Durham (at university), however in about a month I'll be returning home to York.

Cheers,

Rob
 
The power supply is modular, and they've kindly colour coded the cables for the PCI-E connectors so I believe they are connected properly. I've tried a different cable in case it was a dodgy one, no difference however.

I'm hopefully building a system for a friend in around a month or two so this will give me the opportunity to swap the power supply and graphics card, and hopefully find out what's going on once and for all.

Many thanks for all your suggestions, this is the strangest problem I have ever encountered. I've built a number of systems and never experienced anything such as this.

Now it's a race for me to test the components in question and RMA whatever the weak link happens to be; if I still can. I might even revise for my upcoming exams also.

Cheers,

Rob
 
Last edited:
I hope that's not the case, but I'll give it a try!

Cheers for all the help you've offered, just one last thought; is it possible for me to update the BIOS on my videocard, might this have an effect. That said I wouldn't know where to obtain a newer release from, or the procedure involved for flashing it.

Rob
 
Back
Top Bottom