Q: NVDA(OpenGL): Channel exception!

Thanks for the tip. You generally think about max temperatures and the harm they can do. I never thought about the other extreeme being equally hazardous! Someone tell me if I am out of my mind with regard to what I'm about say next (be kind now), When I think about a component that has become faulty or intermittent due to high temp exposure, I generally assume it begins manifesting symtoms when some threshold temperature is reached. Then either the temperature needs to be brought back below the threshold or component needs to be allowed to cool completely before it will function again (assuming it's not completely burned out.) Now, is it possible, for a component to be impacted in a way that causes it to only function when the operating in a narrow or narrowing range of temps?

The reason I ask, in part, is because I wonder if this isn't a compound problem. I have repeated the same routine over and over on a thermally stable Mac Mini - reboot - open Safari with home page set to a static html (no flash or other dynamic content), and within a few seconds to 2 minutes, I see an NVDA channel error(s) followed by channel timeout & freeze. I repeat this maybe 10 times or more over a half an hour or more, then all of a sudden, I reboot, again with out allowing things to cool & presto! I'm able to run Safari visit YouTube plus Itunes, plus virtually anything I want for several hours before the next freeze. If this is soley a driver issue why isn't it more consistent in the way it manifests? Or is it being consistent, and I am just blind to the consistency?

My box is less than 3 months old. Warrenty repair/replacement is available to me for a hardware issue - but if I can't use this thing reliably for more than a few hours at a time due to a driver issue what options do I have?

How often is "Occasionally" ?

There is no set pattern that I can see. However, with my machine, it's much more infrequent than frequent.

Dec 31 13:30:03 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Dec 31 13:30:24 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Dec 31 13:33:53 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Dec 31 13:34:23 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Mar 31 11:09:38 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Apr 14 12:06:34 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Jul 23 14:23:54 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Jul 29 16:14:51 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Jul 29 16:20:27 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
Aug 9 16:38:34 kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error



As you can see, I have the issue, but not nearly as much as other people in this thread.

Also, lately, it's been triggered by simply doing a Get Info on a file in Finder.

-Kevin

Message was edited by: netnothing

steepleton wrote:
fyi- best keep an eye on your hd temp if you're manually adjusting the fan, if it drops towards 30 you're endangering it's longevity (the lubricants are less efficient at lower temps)

There is almost no direct surface-to-surface contact in modern HD's. Most use a high tech magnetic or similar zero-friction bearing for the platter motor's spindle, & that is the only high speed part that would be prone to wear from friction.

R C-R wrote:

steepleton wrote:
fyi- best keep an eye on your hd temp if you're manually adjusting the fan, if it drops towards 30 you're endangering it's longevity (the lubricants are less efficient at lower temps)

There is almost no direct surface-to-surface contact in modern HD's. Most use a high tech magnetic or similar zero-friction bearing for the platter motor's spindle, & that is the only high speed part that would be prone to wear from friction.

just going by the google drive survey "drives that are cooled excessively actually fail more often than those running a little hot"
http://labs.google.com/papers/disk_failures.pdf

Bump.

I really hope 10.6.5 fixes this problem.

I do not find that quote anywhere in the Google study. The closest I see is "In fact, there is a clear trend showing that lower temperatures are associated with higher failure rates." However, I also see "The distributions are in sync with Figure 4 showing a mostly flat failure rate at mid-range temperatures and a _modest increase_ at the low end of the temperature distribution." This suggests the effect is not strong, & the data shows the greatest effects are at substantially lower temperatures than 30° C.

In any case, there is no indication that lubricants play any part in this effect.

It's not temperature related neither it is a hardware failure sympton.

I've checked the airflow, the fan speeds, and even tested six different graphics cards (advantages of having a Mac Pro: 8800GS, 7800GTX, 7900GT, 7300GT, 9800GT, GTS250) from different manufacturers (Apple, Asus, Gigabyte and Club3D).

On 10.6.3 I saw that problem once a week, with 10.6.4 pre-graphics-update I saw it twice each day, graphics update makes no change.

I went from a completely stable system with 2 x 24" monitor to one which regularly fails with some flavor of these errors by switching one 24" monitor for a 30". As long as the 30" is connected I have a chance of these failures, it increases greatly if I add a 24" as well. Happens with all combinations i've tried of an 8800GT and the stock 7300 (both or 30" on one card, one on each card in either order) in both available card slots. Sometimes it fails on or shortly after boot, sometimes it fails several minutes in, usually if I can make it an hour, it'll be solid until I shut down.

Temperature does not appear to be a (heatsinks are clean, 8800 fan works, temperatures dont get out of control, if it starts out stable it tends to stay stable)

Hardware failure is doubtful here either given both cards exhibit the problem.

It feels like a driver issue, something not being initialized correctly.

Happens on a brand-new 2.53 Ghz MacBookPro6,2(Intel HD - built-in, Nvidia GEForce GT 330M - x16 PCIe). I occasionally use a 24-inch Dell FP monitor at work, but this happens with or without that monitor being connected at the time of the freeze.

/var/log/kernel.log spews these exceptions by the dozens per second. The machine is unusable, perhaps because of the error logging which just hogs all the resources on the machine.

One option is to disable such logging completely for a given window of time and see if the machine hangs. If it does, then logging these messages is a secondary issue, indicating that the GPU (or firmware/drivers) are not handling conditions well and instead of giving up and crashing the driver, just unconditionally consume all available resources (bad idea).

Since the specific machine in question has two graphic cards, one crashing is not catastrophic.

NVidia - in general - has a bad track record in my book. For example, http://support.apple.com/kb/TS2377 (this is still an on-going issue - 3+ years as of today). I have already had the logic-board replaced once for a different issue - which resulted in a new/refurbished GPU.

Since this thread has originated in Jan 2009, I suspect, this issue might just take as long to be acknowledged, and dealt with - as the 8600 issue mentioned in the support article.

Found an interesting application - perhaps not applicable on all machines.

http://codykrieger.com/gfxCardStatus/

Will update this thread after some tests.

People, do not get iLife 11! It seems iPhoto (the only thing I got iLife for) triggers the channel exception error like crazy. I can't use my Mac when this new iPhoto is running. So I hope I can downgrade or something. Effing bummer.

Check this, after a looooong freeze:

23/10/10 2:09:12 AM kernel NVDA(OpenGL): Channel timeout!
23/10/10 2:09:12 AM kernel NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
23/10/10 2:09:12 AM kernel 00000069
23/10/10 2:09:32 AM kernel NVDA(OpenGL): Channel timeout!
23/10/10 2:09:52 AM kernel NVDA(OpenGL): Channel timeout!
23/10/10 2:09:57 AM iPhoto[435] An instance 0x3858e00 of class MWController was deallocated while key value observers were still registered with it. Observation info was leaked, and may even become mistakenly attached to some other object. Set a breakpoint on NSKVODeallocateBreak to stop here in the debugger. Here's the current observation info:
<NSKeyValueObservationInfo 0x2e63950> (
<NSKeyValueObservance 0x2e638d0: Observer: 0x2e58710, Key path: currentAlbum, Options: <New: YES, Old: YES, Prior: NO> Context: 0x0, Property: 0x2e5ef50>
)

Yup.....same here. iPhoto '11 has caused it at least 4 times so far.

-Kevin

Remember the days when Apple made quality software that even improved performance on their older models? Good times. Past times.

Yeah huh. I have a 2006 Mac Pro with 11GB of RAM and iLife '11 runs noticeably slower than iLife '09 did just a day ago. C'mon Apple....how about progress software so that each version is faster and leaner....you're not Adobe after all!

-Kevin