Tracking Down Verde 2xx Freeze Up on M17x

Okay, so for many of us (but not all, oddly enough), the nVidia Verde drivers have been giving mystery lock-ups (graphics freeze, sometimes crash to desktop or driver failure notification but usually a hard lock with no BSoD). This seems strange as not all notebook users on the Verde drivers are seeing this issue.... in fact, it appears not all M17x users with nVidia cards are seeing the issue!

So, in the interest of tracking this down to either get fixed or find the work around, I want to try and get as much info on what happens and what is running when the freezes happen.

For simplicity's sake, use the latest Verde beta (260.63) ( 32-bit / 64-bit)with a clean install (don't use the installer's clean install method, uninstall and clean out nVidia drivers using Drive Sweeper or similar; when you reboot Windows should be installing default VGA drivers) with PhysX installed and 3D Stereo omitted.

So, here's the format I would like to use....

I just got done playing Civ 5 for about 2 hours with no CTD, going to try testing with WoW as that appears to almost always freeze up with the Verde drivers (too bad, too, as with Verde I can have everything turned up (except shadows, those at -1 max) with 1200p resolution and 4x AA and still get 40+ fps in Dalaran).

**EDIT** I uninstalled AVG Free 9 after reading somewhere AVG can sometimes interfere with some games/drivers. Using Microsoft Security Essentials now, which has resulted in a much faster start-up
(45s v 1m).

**EDIT 2** Updated with steps I had forgotten.

**EDIT 3** WoW test failed... played 35m (did two heroic runs and was on WG) when it froze; oddly, though, i was able to recover normally and get back in which hasn't happened before (played for another half hour before quitting). System seems stable... nothing in crash log and ThrottleStop doesn't show any throttling occurring.

 

Computer Overview
Model: M17x-R1
CPU: Core 2 Quad Q9000 @ 2.26 Ghz (W/OC)
RAM: 4GB (2x 2GB DDR3)
HD1: 500 GB 7200rpm
HD2: 160 GB 5400rpm (no raid)
BIOS: A03
SLI: YES 280m (hybrid mode)

Software Overview
OS: Windows 7
AV: Windows Defender
Connection: WiFi (802.11g / WPA2)
Other: none
Drivers: All latest
Freeze Up Description
random freeze doing nothing, or just surfing in web, or playin for a while, i can't find any pattern on this.
Steps Taken So Far
unninstall crapy drivers and roll back to 197.83 (really stables + latency fix)

 

Point is we shouldn't have to roll back as the drivers work for most people, already have the latency fix, and are necessary for newer DirectX 11 games that are coming out (or already out, in the case of Civ5). There's something small and stupid here causing these lock ups, just trying to pin it down as these drivers seem to be working for most M17x-R1 users.

 

I was fine with Civ 5 for about an hour the first night. Next time it froze within 10 minutes. Wow plays fine for me for a 3 hr 25 raid as long as I had NO add-ons enabled. Any add-ons caused it to fail the driver then stay at a black screen. With Civ the game just errors out after the video card driver resets.

EDIT: I try to play games in windowed mode. There didn't seem to be any difference between going from full screen to windowed full screen.

EDIT 2: This was with a fresh windows install + 260.63. Uninstalled and went back to 197.83 after.

EDIT 3:
Computer Overview
Model: M17x-R1
CPU: Core 2 Quad Q9000 @ 2.00 Ghz
RAM: 4GB (2x 2GB DDR3)
HD1: 320 GB 7200rpm
HD2: none
BIOS: A05
SLI: YES 280m (no hybrid mode)

Software Overview
OS: Windows 7 Ult 64
AV: None
Connection: Wired
Other: none
Drivers: 197.83 since 260.63 doesn't allow anything but Minecraft to be played.
Freeze Up Description:Random Freezes in Wow(with add ons enabled) and Civ 5 would black screen, flash, nvlddmkm.sys restarts and Win7 states that Civ 5 crash for an unknown reason.

 

Uggh. I wish this thread would started a week or two ago. I would have been happy to help out with information from the same issue with our M17x that is currently listed as wife's computer in my signature. Now I have a big red brick sitting on the coffee table waiting for a system exchange because the on-site tech was doing a motherboard replacement & broke the ribbon cable clasp from the motherboard to the alienfx daughter board & apparently something else because he fried the majority of the system too (wonderful smell ). I'm just glad I had already backed up my data when doing the multiple OS reinstalls & hard drive replacements before that service.

In all honesty, I kind of feel that he did me a favor by damaging it. It sped up the system exchange process since I was fairly certain that a motherboard replacement would not have corrected the issue. After multiple OS reinstalls (1 clean install on a spare hard drive I had before I called for support, 2 factory restores & 1 clean install on the original hard drive, then 2 clean installs on the replacement hard drive) I still feel this issue has some relation to the combination of the BIOS, nVidia chipset, nVidia graphics card & nVidia drivers for each (& this is coming from someone that usually prefers nVidia over ATI).

 

Well, on a whim, I used ThrottleStop to force the CPU clock to stay put and undervolt'ed it to 1.05

Since then, I have been playing WoW for 2 hours no crashes or slow downs or lock ups (aside from a slight lag when Pitbull switched to 25-man mode) (did WG, 25 man VoA, 10 man Voa... FPS was awesome) and Civ 5 for about 2 hours with no problems (aside from a CTD when switching from regular map to strategic and back again... but this is being complained about by many people). System up-time is now at 5 hours 10 minutes.

This may be a voltage issue... very unsurprising. All I did was undervolt the CPU (while testing with Prime95 on FFT) and all of a sudden I have a stable rig using the 260.63 drivers. Wish I could confirm undervolting actually worked and not something random done that worked magically after a clean restart.....

** EDIT ** Quick peek at ThrottleStop log shows no throttling of the CPU... both before undervolting (when I was still locking up) and after. May not have been a CPU throttling issue but looking very much like a voltage issue which is not quite a surprise for any owner of these systems...

** EDIT 2 ** Fixed typo, not 11.05v, just 1.05 (value in VID in ThrottleStop). I can not take credit for this solution as unclewebb suggests it for eliminating clock modulating (which causes our favorite old issue, DPC latency)

 

Interesting, You need a confirmation... I'll try it tomorrow. To late right now. =P

 

Well a guy here in the forum said that all issues seems to appear when the SLI mode is enable. that would explain why some R1 user (with no SLI) no have problems and why other (with SLI) have problems. I just can't believe Dell actually set as official in their website one of those 2xx.xx drivers for the M17x.

 

Our R1 did not have SLI, but had the identical issue.

 

Okay, I've had the laptop running since my last post with no issues. I went to bed while it did some video encoding (MediaCoder NT CUDA); woke up and played Civ 5, then did something stupid....

I left Furmark running in stability mode while I went to the car dealership to get some work done. I came back, remembered, and fully expected to see Furmark frozen or the PC recovered from a BSoD.

Furmark was still running after 2+ hours. Temps were average for game playing (60C/58C) and system responded fine. Fired up WoW for giggles, played for a quick half-hour with no problems.

I'm not saying this is the solution, but here is what I've done for those that want to confirm:

ThrottleStop runs on start-up; options are set to start minimized, not reset FID/VID on exit, nvidia GPU, and intel core 2 CPU. ThrottleStop has Clock Modulation enabled at 100, Multiplier enabled at 10.5, and Voltage ID at 1.05; Turbo is also disabled.

The multiplier and voltage may vary depending on CPU; I used 10.5 as google suggested that is the proper multiplier for a T9600 and 1.05 was at unclewebb's suggestion (to another user, about a DPC of all things... funny to think this still applies here).

@Tazalanche - This is the first I've seen a non-SLI user complain of the lock up issue, however, you may have been able to benefit from the fix as well... it is quite likely nVidia changed something in the 2xx drivers that makes the cards real power hogs so it becomes more evident in SLI setups as opposed to single card. Of course, this is all speculation and it could be I just changed the magic number of things to make this all work.

 

Update: Played WoW a bit before work today... crashed during 10-man weekly (recovered, then hard lock); crashed again while doing solo of BRD (about 45m in... I got lost, leave me alone). Funny thing is after restarting, then doing a proper restart, system was stable again. Finished another BRD run, 2 quick 5man heroics, and a WG battle with no crashes, then played Darksiders for about an hour with no crash. Very strange.

Turned M17x off, turned back on when I got home and had no desktop (no task bar, no Task Manager, nothing). I had to hard restart which resulted in Windows disabling both 260m's; re-enabled and been fine since. Not sure if this is a sign the undervolting does not fix the problem or what.

 

I have hybrid enabled and always play with SLI enabled and have never encountered this lockup issue. This theory is more than likely not valid.

 

Where can I get 197.83? I have 197.16

 

I'm not sure they exist, Dell released 197.84 beta drivers though.

 

laptopvideo2go.com is where I got mine. I also had a set from when they first came out. =P

 

Are you running the Verde drivers? If so, the Dell release or the latest nVidia beta (260.63)?

I ask only because if one checks the various Verde threads around the Alienware forums here at NBR, there are usually quite a few cases of, "Crap, they don't work because of [lock ups]; reverting back to Dell's last 197 release." My understanding (perhaps flawed) was that this issue is semi-prevalent for SLI users on the M17x-R1.

Even if there are not as many cases agreeing that they have lock ups, there are still more than enough to stop it from being dismissed as a simple fluke for one or two people...

 

I usually try every new driver that's out so I have tested a bunch I can't tell you exactly sorry...

but from nvidia betas and WHQL as well as all the Dell betas (right now i have the 260.63, installed them the day they came out on laptopvideo2go ) and no lockup or black screen issue.

but i have to say i use the dedicated cards only when gaming or crunching cuda for boinc, so i guess maybe i use the integrated quite a lot compared to others.

 

I've tried every version of the 2xx.xx drivers (clean installs) and they've all had issues of black screen freezes on my M17x R1 280M SLI'd system...with or without SLI enabled.

Always have come back to the 197.84 drivers which are stable...

 

Hey.... I had a R1 with a single 4870. Also i had the dpc latency problems. Music, movies, applications and games freezes with a nasty sound that follows. I compained alot to dell and finaly they agreed to give me a new system. But i had fixed the nasty freezings and lockups before i returned it. What i did was downloaded the app from dells download section. Cant remember the name right now but what it does is uninstalling the hdd handeling. It seemed to be atleast for my problem if windows handles the hdds the dpc latency program registrates almost zero high spikes. If you havnt tried this yet i can check the name on the program when i come home. Im on the iphone right now.