
UNSOLVED
Aurora R3 Lockups & nVidia Driver problems
I received my R3 about 2 weeks a go. Worked great for about a week. Then nVidia driver has recovered from a failure message, followed by black screen lockups, blue screen of death, random reboots.
My specs are i7 960 2600k 3.4 overclocked to 4.1
16 Gb DDR3 Ram, dual nVidia GTS 450 (SLI). 2 Hybrid SATA 500Gb drives (Raid 0)
SLI option in nVidia control panel appears and disappears (along with other options) at random. Dell tech support has been totally useless (I used to be a tech support manager for a software company in UK and built most of the departments servers & desktops).
If you are running dual graphics cards, you are going to experience a discrepancy in GPU temps between the 2 cards. The ventilation system, in my humble opinion is flawed. There is a 22 degree (F) difference between upper and lower cards. Running CUDA apps (almost identical) primarily using the GPU's, the load on the hotter GPU is 98%, the cooler one is 83%. Now I have a second PC running an overclocked EVGA branded GTS 450. Running equivalent CUDA app, load is 98% but GPU temp is nearly 28 degrees (F) less than cooler one in the Aurora R3.
Electrical resistance is inversely proportional to temperature. The hotter components get, the higher the resistance. If you are running dual GPU's, you need to have both cards running at about the same temp. If you are gaming and SLI is enabled, the 2 cards are not strictly performing identically therefore you are going to get video problems, crashes, lockups, blue screens etc.
I have sent 3 messages to Dell/Alienware about the internal design and airflow within the chassis and, surprise surprise I have heard absolutely nothing. So much for their Resolution Center's "we'll get back to in 48hours" promise.
Responses (74)
Solutions (0)
One thing to also note is that the upper card won't have much 'air' compared to the lower card due to the lower card literally hindering air flow. As recommended above, you may want to consider altering the Active Thermal for the fans to manual curve to force more airflow for the PCI fan.
As they used to say, let your fingers do your walking, dial tech support. If I was as worried as you seem I'd be on the phone since chances are anyhow you aren't going to get it checked out by email. I have always found talking to a human always works the best. Since getting my Area 51 ALX I have called twice and never call Monday morning or in this case this week being a Holiday Tuesday have so far not waited more than 5 minutes on hold if that long. I make it a point if possible to do all checking out of what I feel they will ask me to do. This site and one person has really helped me and when talking to tech support told him all Ihad done and what was told me on this site and put on hold for a couple minutes and came back and said they would set me upwith new part and someone toinstall. Very painless.
I have spent 5 hours with tech support on the phone already. They have been totally useless so far as fixing the problem or offering any solutions.
They also couldn't work out why my card reader wasn't working until I asked if there was a size limit on the cards it could read. 10 minutes later the guy came back and said yes there was a problem, they were not sure whether UDMA cards would work.. They have not offered a solution so I spent $20 and got an Atech Pro Card Reader that recognizes up to date technology. (My R3 is 2 weeks old!)
Thanks Tesla 1856,
Actually the CPU is not overclocked, Dell/Alienware overclock the turbo boost. Their specs are really misleading on this one. I've run an Intel diagnostic on processor and its performing well within limits. The drivers are a clean install. I tried the latest nVidia driver and that made matters worse, so I'm going back to an earlier (pre Dell) driver and see if that stops driver errors. Currently SLI options in nVidia control panel has disappeared again. SLI bridge ribbon cable defective? I have, under Tech Supports "guidance", basically stripped down the internal components and re-seated everything.
Your suggestion about manually controlling fans had occurred to me. I have been running all the fans I can control at 100% for the last hour and the temperature differential between the 2 cards has dropped 5 degrees F, as has the temp of the upper card. Incidentally I also have a HEPA filtered 42" vertical fan blowing over the unit as well.
There are so many posts about the freezing/lockup issue, especially with the R3. Memory Manufacturer comes hign on the list. My R3 has Micron 4x4Gb.
I've only had 1 error message today, with one black screen.
FYI I'm a photographer and the applications I use are not really that demanding apart from DxO Optics Pro. The R3 rendering time for images is 3 times faster than my Dell XPS Studio 8100 (2.8 Gig Quad Core i7).
Actually, I had similar driver 'recovered' errors along with the Texture graphics issue that follows it later. I am not using the same system as you are, mine is an area 51 with nVidia GTX 560 Ti card that came with the system. After spending 3 weeks with trying to trouble shoot it with nVidia techs and with Alienware, I found putting my card on Performance for Power instead of Adaptive solved my problem with the driver issue. I am planning on talking with nVidia to see if this is normal on their developer side. I do know that Alienware will state their systems are not designed for 24+ hour running, as it is not server hardware, so I am believing that the OEM manufacture of the card might be the culprit. I would try setting the power management for the cards to Performance and see if it stabilizes, assuming your card does Variable clockrates for the GPU/Memory/Shader.
Sorry I have been away from forum for a few days. Here is an update.
1. Replaced the unbranded card reader with an Atech unit. This has corrected Renasus USB driver problems. Now able to read UDMA cards over 8 Gb. Shane I had to pay for it.
2. Upgraded nVidia drivers to latest version (280.26 from nVidia) and still had occassional driver not responding/recovered mesages. SLI option has not, as yet, disappeared
3. Uninstalled Alienware Respawn software and I have now gone almost 3 days without error messages or black/blue screen problems, or, random reboots.
I still think ventilation flow to graphics cards is poor. I'll let you you know if there any recurrences.
Thanks for everyone input...really appreciated.

Tesla1856
10 Wizard
•
17985 Posts
•
71600 Points
2362
0
Posted September 4th, 2011 23:00
Have you tried removing the CPU OverClock?
Have you tried setting the PCIe fan the Manual-Curve to get more air in the front? This is with the fans on the cards themselves on Auto ... next step is Manual.
Be sure you have a nice clean install of drivers. Maybe not the latest, but something stable and official for 450.
Ya, an Area-51 would be better for dual-cards, but 450's are not real powerfull, so it doesn't seem like they should be running hot.
Dell Rockstar
Microsoft Windows and Apple iOS Developer (Retired)
- Like many of you, I can appreciate a good game-engine.
- I answer questions here, but I'm not a Dell employee.
- Consider giving posts you like a "thumbs-up"
- Posting models-numbers and software versions speeds trouble-shooting.
- Click "Mark as Accepted Answer" on any post that answers your question best.