Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
So are all MBP owners with the 8600m GT cards screwed?
- Thread starter jjahshik32
- Start date
- Sort by reaction score
Yes, but we have different hardware than other models. The memory leak could be in a driver that is only used on MBPs, or in some common high in application that that is only run on MBPs. It could also be a software issue that can only be triggered on the mobile version of the hardware, as a different code path is taken.
To keep it short, we have different hardware, that means different code paths at some point are being taken. My guess is that the memory leak would be in a driver lvl, and that would be why killing all of the apps does not seem to free the memory.
To keep it even shortly (and hopefully not rude) your logic is flawed and yes, a software issue can be isolated to a single hardware platform.
-Zeek
If you're right then it must be an OS rather than app issue to be affecting so many people.
It should also be easy to see, which means that as you have a machine where this triggers, you could watch memory use and catch the faulty process.
Of course, your explanation leaves the mystery of why Apple haven't figured this out after months of replacing both components and whole machines when this fault is reported, nor does it explain the MBP users who don't seem to have the problem.
It especially doesn't explain people whose machines run just peachy for a while then fail, nor does it explain machines that remain faulty after a cold boot.
It could be that I am seeing some thing entirely different. And that we have two different problems. But with my experience with kernel level driver programing (linux) I could see how a code path could be triggered that could eat up all of the memory in a matter of seconds.
Again, my guess is that this would be in a driver some place, and not a application.
Today I am going to write an app that just eats up as much memory as I can to see what happens. And if that fails to reproduce the problem I will write a driver for OS X that eats up as much memory as at can to see if I can reproduce what I am seeing on my end.
Since its been official that all 8600m GT cards are defective that we can only prolong the gpu failing..
Ok so this makes me feel like I've just gotten kicked in the crotch for the 2nd time. So this means the 3 year apple care I bought might be a lottery pick.
For the people who hardly ever games and reaches a low temperature of their gpu cards would take much longer for the gpu to die and what if your gpu dies a year after the 3 year warranty~
This has ruined my day, no week, no month.. no year~!!
I just came across this. I have a 2.2 MBP 8600 128MB and have not seen any of the problems discussed here. I have played hours of games (RTS, Company or Heroes, AOE III, etc (under Windows Vista in Bootcamp) and run Parallels all day (which keeps the fans on a good portion of the time) and have never had a lockup or any graphic glitches.
Is there any way to determine if this problem is likely to occur on any given system or is it just a time bomb waiting to go off as other suggest?
Cheers,
Then to test your theory, you would agree that one of the MBP owners here who has experienced this problem in OSX to install windows and/or linux in boot camp and see if the problems occurs also. If it occurs, then its a hardware fault. It it doesn't occur, then we know its a possible software fault.
Yes and no. Mine MBP "video card" failed after watching OTA HDTV for half an hour. It's pretty heavy load. Logic board was replaced and no problem since. I might suggest for people in doubt about video card, whose computer still under warranty, to watch Olympics for the rest of the week on computer screen
Yes and no. Mine MBP "video card" failed after watching OTA HDTV for half an hour. It's pretty heavy load. Logic board was replaced and no problem since. I might suggest for people in doubt about video card, whose computer still under warranty, to watch Olympics for the rest of the week on computer screen
hhaha lol I'll give that a try just for kicks
I'm giving up on this generation. Last week I took back a brand new MBP becasue it started displaying the dark vertical lines. This one just turned off abruptly, then showed a scrambled screen when I turned it on again. Back to the slow old iBook G4 for me... 'til they bring out an MBP with a different graphics card
Well that's the thing, when you push the gpu harder it results it getting much hotter then what do most of us do when we stop using our notebooks? We put it to sleep and from I can confirm that after you put the mbp asleep for a while the machine gets cold like it's off and I'm sure you heat cycle like this 2-3 times a day.
I almost never put my system to sleep. I do a complete shutdown. And when I leave work I put it in the sleeve and in my bag. It's still pretty warm when I do so. I been doing this over a year. I have not experienced this problem. If it is a materials issue with the GPU, isn't it possible that some systems are fine due to a better than average part being in that computer?
Cheers,
Cheers,
I'm not sure what you mean by "better than average part". a GPU is a GPU is a GPU, that is, an nvidia 8600M is an nvidia 8600M, and so on.
Probably just wishful thinking on my part. Quality Control being such that it is, not all parts are manufactured equally. Certainly there exists the possibility that some of these GPUs were manufactured with a better set of materials.
One can hope, yes?
...In a laptop I paid 2100$ for.
A mistake like this in inexcusable. I expect this laptop to last a good 15-20 years even though I may buy a new one in that time frame.
Yeah I was afraid of that. It was weird too, the screens picture was "swimming" with random colored blocks and random images of what was supposed to be on the screen before it froze completely. Though after a restart everything appears to be working normally again.
Bah I guess i'll be calling apple care tomorrow
inexusable or unavoidable
Given the complexity of computers, things like video card failures or power supply failures or this or that part failing is virtually assured. Unfortunately this appears to be a very wide-spread problem. nVidia has denied this but without a definitive serial number list along with an abnormally large number of reports by macbookpro users, the denial rings hollow. Some people's usage will ensure their machine will work fine for well beyond a year. Others have failures relatively quickly. I am personally ready to buy two new macbookpros but without firm confirmation that the issue has been resolved it would be stupid to do so.
my best advice is to buy applecare if you haven't already. if your machine is over a year old and you don't have applecare, sell it. if you haven't bought one, don't, and if you're thinking of buying a used macbookpro, don't. humorous contradiction there, eh?
If nVidia would backup their claims that the problems are limited we could have a lot more confidence but it would not explain the large number of failures reported in supposedly "OK" machines.
Given the complexity of computers, things like video card failures or power supply failures or this or that part failing is virtually assured. Unfortunately this appears to be a very wide-spread problem. nVidia has denied this but without a definitive serial number list along with an abnormally large number of reports by macbookpro users, the denial rings hollow. Some people's usage will ensure their machine will work fine for well beyond a year. Others have failures relatively quickly. I am personally ready to buy two new macbookpros but without firm confirmation that the issue has been resolved it would be stupid to do so.
my best advice is to buy applecare if you haven't already. if your machine is over a year old and you don't have applecare, sell it. if you haven't bought one, don't, and if you're thinking of buying a used macbookpro, don't. humorous contradiction there, eh?
If nVidia would backup their claims that the problems are limited we could have a lot more confidence but it would not explain the large number of failures reported in supposedly "OK" machines.
that's the whole deal right now that there were no different set batches of the material that's the blame for defective materials used in all the g84 and g86 gpu.
I would agree though that even though all the g84 and g86 gpus have the exact same parts that failure % can still vary, like some cards will die after a 6-12 months others probably 2-4 years.
I guess it depends on what ambient room condition, your usage patterns as well.