Contribute
Register

Editing AGPM to prevent GTX4XX Freeze

Status
Not open for further replies.
cowfodder said:
Of course someone seemingly manages to come up with a fix for fermi cards literally 2 days after I order my system w/ ati 5770. May have to keep an eye on this one as nvidia has better linux support.

Don't feel too bad- the 5770 should be a great card. And while this is a fix, it is far from a perfect way to handle the trouble with the faulty NVIDIA drivers. I don't mean to diminish the work of those involved in finding this fix, as I along with many others have benefitted tremendously from this as this seems to be a fix (I've had this trouble for months now and intend on trying this fix in the next day or so once I have some free time). One thing I just want to mention is that my friend built nearly the same system as me minus a different card (he has a GTX-465) and has has a dual-monitor setup, yet he has never experienced any of these crashes and doesn't run a visualizer or anything. Any ideas as to why?!?
 
avfanatic said:
cowfodder said:
Of course someone seemingly manages to come up with a fix for fermi cards literally 2 days after I order my system w/ ati 5770. May have to keep an eye on this one as nvidia has better linux support.

Don't feel too bad- the 5770 should be a great card. And while this is a fix, it is far from a perfect way to handle the trouble with the faulty NVIDIA drivers. I don't mean to diminish the work of those involved in finding this fix, as I along with many others have benefitted tremendously from this as this seems to be a fix (I've had this trouble for months now and intend on trying this fix in the next day or so once I have some free time). One thing I just want to mention is that my friend built nearly the same system as me minus a different card (he has a GTX-465) and has has a dual-monitor setup, yet he has never experienced any of these crashes and doesn't run a visualizer or anything. Any ideas as to why?!?

When using Dual monitor, the card never goes to a lower power state. That means that the card runs most of the time at full speed and only lowers to a 2D speed when not needed. Without dual monitors, the card goes to the lowest clock speed which is known to cause the freezes/KPs.

Cheers!
 
el_charlie said:
avfanatic said:
cowfodder said:
Of course someone seemingly manages to come up with a fix for fermi cards literally 2 days after I order my system w/ ati 5770. May have to keep an eye on this one as nvidia has better linux support.

Don't feel too bad- the 5770 should be a great card. And while this is a fix, it is far from a perfect way to handle the trouble with the faulty NVIDIA drivers. I don't mean to diminish the work of those involved in finding this fix, as I along with many others have benefitted tremendously from this as this seems to be a fix (I've had this trouble for months now and intend on trying this fix in the next day or so once I have some free time). One thing I just want to mention is that my friend built nearly the same system as me minus a different card (he has a GTX-465) and has has a dual-monitor setup, yet he has never experienced any of these crashes and doesn't run a visualizer or anything. Any ideas as to why?!?


When using Dual monitor, the card never goes to a lower power state. That means that the card runs most of the time at full speed and only lowers to a 2D speed when not needed. Without dual monitors, the card goes to the lowest clock speed which is known to cause the freezes/KPs.

Cheers!

I run dual monitors on a gtx460 and get the freeze all of the time. Unfortunately as this as worked for some, it does not work for me!
 
darkstargtk said:
I run dual monitors on a gtx460 and get the freeze all of the time. Unfortunately as this as worked for some, it does not work for me!

What resolutions are you running? Same for both?
 
lukarak said:
darkstargtk said:
I run dual monitors on a gtx460 and get the freeze all of the time. Unfortunately as this as worked for some, it does not work for me!

What resolutions are you running? Same for both?


1920x1080 on the main one & 1280x1024 on the secondary monitor.
 
were you ever able to get the kext to load?

and it should be noted - this isn't a perfect fix. ultimately what we're dealing with is a driver issue that appears to be resolved in Lion's updated Fermi drivers. at the very least all of us Fermi owners have to suffer for another few months till Lion's released and the bootloaders optimized for it.
 
But a fix is a fix and works for me :lol:

However, the AGPM should be analysed to set the correct settings. And as we know, the kext behaves differently for a MacPro and for an iMac.

My system is still working fine with zero freezes or KPs and I'm considering to boot with the GUI instead of the Verbose mode because I find it very stable.

Cheers!
 
darkstargtk said:
lukarak said:
darkstargtk said:
I run dual monitors on a gtx460 and get the freeze all of the time. Unfortunately as this as worked for some, it does not work for me!

What resolutions are you running? Same for both?


1920x1080 on the main one & 1280x1024 on the secondary monitor.

i have the same problem. gtx 460, dual monitors (1680x1050 and 1280x1024) and some freezes (accompanied by sound loops) and KP - but only rarely - not every 10-30 min

modell-id = MacPro3,1

I will try this fix soon and report results.

Anyone tried to force KP by pushing the GPU into state 3?
 
i'm thinking of making this topic more general and not specific to the iMac12,1 description. should i do that, or provide a readymade kext that works on the latest MBP, MB and iMac models?

i've finally tweaked the values to the point where, at least on my machine, switching occurs between states 0 and 1 in a pretty smooth fashion without mouse lag, and will occasionally dip into state 2 when idle. the key seems to be using a long enough idle time to prevent the system from switching power states too quickly, and to prevent the occasional dip into state 3.

i've still yet to have a fermi kernel panic, so i think this is a pretty good hack till Lion rolls around. that said, there's only a handful of us who've actually tested this to make sure it works. i have another hackintosh in my house with a EP45 and MacPro3,1 smbios file that i could use to test that compatibility, but i'm lazy ATM :)
 
scorcho said:
i'm thinking of making this topic more general and not specific to the iMac12,1 description. should i do that, or provide a readymade kext that works on the latest MBP, MB and iMac models?

i've finally tweaked the values to the point where, at least on my machine, switching occurs between states 0 and 1 in a pretty smooth fashion without mouse lag, and will occasionally dip into state 2 when idle. the key seems to be using a long enough idle time to prevent the system from switching power states too quickly, and to prevent the occasional dip into state 3.

i've still yet to have a fermi kernel panic, so i think this is a pretty good hack till Lion rolls around. that said, there's only a handful of us who've actually tested this to make sure it works. i have another hackintosh in my house with a EP45 and MacPro3,1 smbios file that i could use to test that compatibility, but i'm lazy ATM :)

MacPro3,1 doesn't load AGPM kext. Only MacPro4,1 / 5,1 and newer iMacs does.

You should only post the code needed to be edited because not only you have to set the Machine type, but also the deviceID. If people have gone this far setting up their Hackintoshes, they could do a simple kext edit.

Cheers!
 
Status
Not open for further replies.
Back
Top