I thought we might gather some useful information about the failure rates for Nvidia GPUs after all of the recent brouhaha over the known problems in Nvidia mobile GPUs and rumored problems with some of their other products by posting a poll and asking for comments from folks affected by the problem. However, it turns out that we didn’t learn much that we didn’t know.
The poll results, such as they are, indicate incredibly high failure rates for all of the Nvidia parts in question. Trouble is, I also asked people who voted about a failure to "please follow up by posting a comment explaining what failed on you and how." To date, we have 3156 votes, 438 of which indicate one or more GPU failures, and I count two posts that might be relevant to the problem: one from Hatting about a failed "8600 GTX" (wait, what?) and one from adisor19 about two failed Dell XPS 1330 laptops. Neither commenter provided any information about the nature of the failure.
That makes me think we’ve pretty conclusively proven that fanboys and corporate drones will indeed stuff the ballot box in an online poll (like I said, nothing we didn’t know), but I’m in no way convinced we’re seeing a phenomenon on the order of the GXP fiasco. Nvidia has already admitted the problem with laptop chips is a real one, and we have indeed seen some anecdotal evidence about that, but nothing we’ve seen here or elsewhere convinces me that desktop G84, G86, G92, or G94 chips are failing at abnormally high rates. Could they be? Sure. But our little exercise hasn’t exposed anything shocking.
That was pretty much to be expected, I suppose, but eh. It turned out differently in the case of the GXP.
We’ll keep watching this story for signs of life, of course, but I’m going to go ahead and close the poll. We’ll replace it with this new one, which should be just as much fun.