I usually like this guy’s videos, but he lose a bit of credibility in the beginning of this one with his “apple has a timer that run out when apple care is over and breaks these machines” -_-
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.
MBP 1,1-10,2 Force 2011 MacBook Pro 8,2 with failed AMD GPU to ALWAYS use Intel integrated GPU (EFI variable fix)
- Thread starter AppleMacFinder
- WikiPost WikiPost
- Start date
-
- Tags
- a1286 amd black screen blue and black screen discrete graphics dmg dmg files gfxcardstatus glitching gpu failure gpu panic gpu switch graphic card hardware fault hardware issue logic board failure macbook pro "15 2011 macbook pro (15-inch early 2011) macbook pro 2011 radeongate restart loop white screen
- Sort by reaction score
- Status
I usually like this guy’s videos, but he lose a bit of credibility in the beginning of this one with his “apple has a timer that run out when apple care is over and breaks these machines” -_-
Well, plainly said, it was... but in a sense he is right. That is also why i mentioned that companies are aware of the crackling and breaking issue and use acceptable raw material. That does not mean that they use perfect material or they would be using pure gold cuz of its good electric abilities (and heat emission is a great identifier for any electronic product cuz it is an indicator of how much the product will live, and if you pay close attention, any electronic device that i heating more, is living less). All these things are machines and they can be programmed, not nessessarily with timers. Have you ever heard how warranties are calculated? A rough estimate would be that knowing properties of any material and average to high usage, mixing them with some testings and statistics, you will come to an s^2 estimate which, in combination with economics and some jibrish marketing stuff, will show you what will be your gain out of the life expectancy of an "item". I am sorry that i do not get deeper into that as it is auite complex, my english is not enough for this, and as always googling is pretty easy nowadays (back in highschool, google was dialed through 28kb/s modems). So please dont criticize someone as quickly. I work in IT and fix things as well and I always deal with new problems. Never before have i dealt with Macs but i decided i wanna know more about it. In general, "timers" can easily be interpreted as to the life expectancy of an electronic part (in this thread, the dgpu chips). And nowadays, if you observe it and exclude the misuses of a product by consumers, electronic products break down just about when their warranty is over. That means that the research on the product was successful, with very little deviation from the initial estimate, which is called hardware failure. In the past, companies used more durable raw material and products were not breaking down frequently or soon enough. That would have an immediate effect on the company's economy as the consumers were not consuming anymore. I dont wanna get more into it cuz it is a damn long conversation and out of topic. You can always look it up over the internet.
Last edited:
In one word, yes. When conjunctions and solderings within a chip melt, they lose material to heat and surroundings. It will be fixxed for sometime, but will become thinner, thus easier to break again. If you oven it enough times, there will be no material left to "melt" together in order for the circuits to be closed and working.
And what time frame (approximately) are we talking about till it gives me same issue?
There are 4 chips protected by shields on the logic board. I have taken the shields for CPU and GPU off and cleaned all the thermal paste. The other chips (I'm not sure what they are)... Beforeng it in oven, should I take off those shields and clean them? I'm not a professional in reapplying thermal paste!
Almost always haf of the life it had. The first will give you at most a year of life. The second will give you about half of that. Leave only the black color mirroring chip on the green (most probably) pcb. Dont forget to cover EVERYTHING in aluminum foil of tripple or more layers apart from that black chip. Search some videos there are on youtube.
I only made reference to the 4 minutes and 30 seconds of the video.
The problem is on the AMD chip and not the solder. If someone does a hardware repair, then you have to change the AMD chip as well. Only soldering service would not work for a long time.
I was overheating Macbook Pro just to install Mac OS. Actually the main video card remains the AMD after overheating.
I had to do this before. After the discovery of AppleMacfinder, now I no longer need to overheat.
That overheat is maintained by a few reboots or shutdowns. Then the AMD video chip problem comes back.
If it was not AppleMacfinder's discovery, maybe I'd have to spend money buying a new AMD video chip or even buying a newer Macbook Pro model.
Imagine having to overheat several times a week to be able to use Macbook Pro. The useful life of the Macbook Pro would fall far short.
Update:
My overheating is a home-made overheating. Macbook Pro turn on and frozen on the gray screen and superheat the AMD video chip for 20 or 30 minutes.
If it was an Apple technician (only soldering), then the technician would open the Macbook Pro and apply heat directly to the area of the AMD video chip and the heat would be higher.
So overheating done at home is short-lived.
Now I do not need to overheat.
I prefer to stay only with the Intel video card than having to make several overheating during the week in order to use the AMD video card.
Simply use:
sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00
[doublepost=1501875243][/doublepost]
But the part of the overheating I verified in practice.
I did several overheating before the discovery of AppleMacFinder. And overheating works. Too bad that for a short time (home made overheating).
Last edited:
Previous Shutdown Cause: 3
SMC: smcInitHelper ERROR: MMIO regMap == NULL - fall back to old SMC mode
This is normal. It depends on the version of Mac OS you are using. I got this message on Yosemite with normal boot and running all right.
Doing real test again, i saw that all these lines show up on a successful boot but the mbp does not go white screen, but boots into os normally.
The white screen you refer would be the logo screen of Apple with the progress bar?
In my case, the boot starts in verbose mode but at the end appears a white / gray screen with the progress bar and the Apple logo. Then the login screen appears.
When the AMDRadeonX3000.kext was still in /System/Library/Extensions, then sometimes the system crashed into en2: promiscuous mode. But in the above lines was the famous IOConsoleUsers ScreenLockState 3.
Even the MacBook Pro Retina 15 "Mid 2012 has had this video problem.
https://discussions.apple.com/thread/6769227?start=0&tstart=0
I was able to decrease a few degrees Celsius with these additional settings.
Now stands at 30 degrees Celsius in idle mode. Only open Active Monitor and Terminal.
Google Chrome also with open pages is around 34C / 45C.
When the video chip AMD was still working properly, I noticed that the Mac OS self put gpu-policy=00% or gpu-policy=01% depending on the GPU case in operation.
sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-active=%01%00%00%00
sudo nvram gpu-policy=%01
https://github.com/ah-/switcher/issues/2
https://github.com/ah-/gmux-scripts/issues/1#issuecomment-68113930
Now stands at 30 degrees Celsius in idle mode. Only open Active Monitor and Terminal.
Google Chrome also with open pages is around 34C / 45C.
When the video chip AMD was still working properly, I noticed that the Mac OS self put gpu-policy=00% or gpu-policy=01% depending on the GPU case in operation.
sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-active=%01%00%00%00
sudo nvram gpu-policy=%01
https://github.com/ah-/switcher/issues/2
https://github.com/ah-/gmux-scripts/issues/1#issuecomment-68113930
Last edited:
No. White screen is white screen. Completely white without anything on it. Even in verbose mode. The verbose mode disappears and the screen goes white. Then after a couple of minutes it reboots. Same goes to booting without verbose mode. The apple do have a loading bar and it loads up until about 70-80%. Then screen goes black for a second and then again completely white where it hangs and reboots after a while.
Now this is interesting again.
First, I did not know that the AMD-chip would be able to really heat itself back into 'repaired' status. I thought overheating and forced shutdown would only trick the OS into booting iGPU, once. Would be a nice trick to avoid booting single user for the current procedure. But that never worked for me.
Unfortunately with High Sierra my MBP is overheating really badly if our TheFix is not applied. Indeed so quickly that it also shuts down the machine really quickly, less than a minute. Around 5 installs of the beta with self-enforced overheating shutdowns/reboots now and not once came the Radeon back to life.
Certainly, overheating by itself has never done anything good here.
As great as the efi/nvram-fix is, it is by no means a perfect solution. Even as we still strive to improve on it here.
Speaking of which:
If that indeed helps it would be great. But I doubt that currently.
Does this setting stay? While booted, across reboots?
The temp sensor on the broken GPU seems to be broken as well here.
Before reading your new nvram trick I didn't boot with gpu-policy=%01, yet GPUdiode hangs around between 6–8°C now without the fix and without any agc= variable. (Disclaimer: Currently I also use a much steeper response curve for temps that sets in earlier than the system's default. Fan-speed rarely dips below 3000rpm.)
But I will try to see what happens.
My suggestion would be to look for how much you are discharging your battery.
One easy tool for that: https://www.coconut-flavour.com/coconutbattery/
I am discharging now without your latest suggestion for gpu-policy with 10–15 Watts (spikes into 30s, light usage).
Last edited:
Hello.
My config is MacBookPro 8.3 (17", i7 2,3, AMD 6750, Sierra 10.12.6). My video chip not completely dead yet. Sometimes it even works (but not too long).
For dGPU tests I`ve used http://www.geeks3d.com/20140304/gpu...w-fp64-opengl-4-test-and-online-gpu-database/ I`m sure that it is OK for "Real Test" . 100% times switches to dGPU when it`s possible.
Yesterday I`ve spent a lot of time trying various solutions:
My current working solution is
What I`ve also tried:
My config is MacBookPro 8.3 (17", i7 2,3, AMD 6750, Sierra 10.12.6). My video chip not completely dead yet. Sometimes it even works (but not too long).
For dGPU tests I`ve used http://www.geeks3d.com/20140304/gpu...w-fp64-opengl-4-test-and-online-gpu-database/ I`m sure that it is OK for "Real Test" . 100% times switches to dGPU when it`s possible.
Yesterday I`ve spent a lot of time trying various solutions:
- gfxCardStatus - various versions. Very unstable solution for me. Even if set to "integrated only" mac sometimes switches to dGPU.
- Solution with ArchLinux - works fine until shutdown or reboot. Looks like nvram variables restores after each reboot
- https://github.com/0xbb/gpu-switch - as solution 2 it can`t handle reboots in my case.
My current working solution is
- to make following shell-script (gpufix.sh)
Code:sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00 sudo nvram SystemAudioVolume=%00
Second command is for silent boot as you know.
- CMD + S during boot
- running my script
- reboot
What I`ve also tried:
- tried to make login and logout hooks [NO LUCK]
Code:sudo defaults write com.apple.loginwindow LoginHook /<full_path>/gpufix.sh sudo defaults write com.apple.loginwindow LogoutHook /<full_path>/gpufix.sh defaults write com.apple.loginwindow LoginHook /<full_path>/gpufix.sh defaults write com.apple.loginwindow LogoutHook /<full_path>/gpufix.sh
Last edited:
If it works from linux, that means it survives a reboot.
If it works after you applied TheFix in single user, that means it survives a reboot.
You describe a situation where it doesn't survive a reboot from fully booted OS X?
Have you tried starting from a clean slate?
That means disable gfxCardStatus, gpu-switch, and the like, full reboot, then shutdown.
Reset SMC and reset NVRAM/PRAM.
Then apply sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00 ?
OK, it somehow survives only one reboot.
For example:
For example:
- single user
- nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00
- reboot
- boots and NOT switches to dGPU
- reboot from macOS
- boots and DOES switches
That means a couple of things:
Intel chips can withstand the heat much better than AMD ones. All of them. Athlons and Durons died like flies if they had their heatsinks not adequately applied. Saw dozens of them. Intel chips from that era onwards usually throttled down enough as to not damage themselves. This low Tjunction tolerance in AMD chips I would call a flaw. They are suicidal by design.
Choosing such a chip without providing the cooling capacity needed is a flaw on Apple's side. The tolerance was obviously chosen too low. Guide to Store employees: "'The machine is not a laptop but a notebook!' Never call it a laptop, because it runs so hot so regularly that your male customers might sue us for diminshed fertility."
And then making it so difficult to clean the innards, even illegal (no user serviceable parts… ), really tops it off.
I regularly cleaned everything in reach in my MBP and certainly the fans and fins around the heatsinks were clean. Compressed air, swabs, alcohols, every two months or so. You could see a bit dust come out but not much. Just this week I serviced a little LG notebook and the exhausts were covered from the inside with a really thick felt carpet of oils, dust, and cat hair, as was the single fan. Customer complained it was by now running a little louder… But it still workd and you could touch it without getting overly suspicious in that direction.
If you removed all traces of your previous attempts, including the hooks:
Seems like a novel condition in this thread?
Do you see something suspicious under
/Library/Preferences (especially /Library/Preferences/SystemConfiguration)
under /private/var/ or /private/etc?
Like
$ /private/var/db/.com.apple.iokit.graphics
$ /private/var/db/displaypolicyd
Or grepping though the logs?
Last edited:
/private/var/db/.com.apple.iokit.graphics found
/private/var/db/displaypolicyd empty folder
What does it mean?
Several pages ago somebody wrote about nvram auto-reset during reboot too.
I did. But it does not reset during reboot at the boot phase. It resets on the shutdown phase of the reboot. You can test it yourself. After you successfully log into the OS, just force shutdown the MBP by holding the power button. After you boot again, you wont need any kind of single user mode commands. It will boot straight into the OS without problems.
So, at least my logic says that the nvram is resetting on the shutdown phase, not the boot.
Exactly!
Apart from the issues that made this thraed and countless others necessary:
One thing I miss is USB3 while on the road. Who says no to a faster CPU or better colors.
Otherwise Apple does not even offer today a compelling reason to upgrade at reasonable prices.
Latest lineup: 16 GB? Superfast 1TB SSD? Effective default working resolution of 1680x1050?
Built into mine.
Plus: a matte screen, F-keys, replaceable RAM, replaceable drives, replaceable battery, several different connectors I use all the time, SD-slot, adequate size touchpad, MagSafe, 2 audio ports with optical, seperate power button, kensington lock, keys that travel a bit…
And in the laughable section a glowing Apple logo.
Too bad they really had to go all 6 Design flaws on this one. But then, it's classic.
Two thoughts on that:
Does this setting survive multiple linux boots? My guess is a strong yes.
Does this setting survive multiple recovery or single user reboots? My guess is a weak yes.
Something is written to nvram causing now unwanted behavior.
Whether it is a remnant of a program or a of a hack remains to be seen. Maybe the OS itself acts up.
Before I applied TheFix /private/var/db/.com.apple.iokit.graphics had contents different from what is written into that file now. In that same directory was another file, name related (and promptly forgotten), that I first removed and then, alas, just deleted.
If you cleaned LaunchDaemons, -Agents and Log*Hooks I would deduce: Somewhere in those locations I suggested above lurks a critter.
Of course I prefer to find the cause of this and then just fix that.
In this case I assume a clean install will not overwrite the TheFix on shutdown.
Can you try to boot from another partition/drive and see what happens there after a full GUI reboot?
If you do not have a drive like that ready, what are the results with sudo reboot ?
(That skips a few Apple procedures)
Last edited:
That's make sense. My laptop has the same behavior.
What content has your file?
Didn't work. nvram was reset.
Will try tomorrow.
Yes, gpu-policy=%01 remains after 3 restarts and 2 shutdowns.
CoconutBattery discharging with: 8.09 Watts - after 1 minute power off MagSafe connector.
ScreenShot:
https://prnt.sc/g4u0g1
nvram -p
backlight-level ~%00
boot-args -v
gpu-policy %01
SystemAudioVolumeDB %80
SystemAudioVolume %80
Update:
By the way, my sudo nvram fa4ce...gpu-power-prefs... remains after 1 month of restarts, shutdowns and 6 system installs El Capitan, Yosemite and Sierra.
sudo nvram fa4ce28d-b62f-4c99-9cc3-6815686e30f9:gpu-power-prefs=%01%00%00%00
Last edited: