Jump to content
Rampage Dev

New Nvidia Retail Driver Install Solution

87 posts in this topic

Recommended Posts

New Nvidia Retail Driver Install Solution

Posted on May 16, 2013by rampagedev

about:

New Nvidia Retail drivers have been released today. It introduces a new Preference Pane under System Preferences that allows you to enable the Nvidia Web Driver or OS X Default Driver The new drivers require information to be written to the NVram. Without this you will not be able to Meklort, who developed the NVram Module that we currently use, assisted me in understanding why the NVram information showed up on some systems and not others. If you are using my Recovery Partition And Enable All Features Of iCloud and iMessages fix you must remove nvram.log from /var/log/ before you can begin.

If you plan to enable ECC, for a Nvidia card that support this feature, you must add the slot identification edit in your DSDT or the option to enable ECC on that card will not be available.

Accreditation:

Meklort for NVram Module and boot.plist Flag

Originally posted here

Guide:

Part 1: Install

1. Download Mac OS X Driver Release 313.01.01 here

2. Install the drivers

3. Install the latest Bootloader found here

4. In your Extra folder if not present already create a folder called modules:



screen-shot-2013-05-15-at-10-36-58-pm.pn


5. From i386/modules folder copy FileNVRAM.dylib into Extra/modules:


screen-shot-2013-05-15-at-10-37-31-pm.pn


6. Continue to Part 2


Part 2: Fusion Drive Installs or Systems Without FileNVRAM


1. In your Extra folder open org.chameleon.Boot.plist:


screen-shot-2013-05-15-at-10-39-15-pm.pn


2. Under Kernel Flags add nvda_drv=1:


screen-shot-2013-05-15-at-10-41-25-pm.pn


3. Continue to Part 3


Part 3: Enabling


1. Go to System Preferences:


screen-shot-2012-11-23-at-7-45-43-pm.png


2. Go to Nvidia Driver Manager:


screen-shot-2013-05-15-at-10-43-46-pm.pn


3. In the Preference Pane change the settings to match the following:


screen-shot-2013-05-15-at-10-45-11-pm.pn


4. Restart


5. The new kexts should now be loaded:


screen-shot-2013-05-15-at-10-47-27-pm.pn


6. If you are using my Recovery Partition And Enable All Features Of iCloud and iMessages fix please redo Parts 4-7 here

Share this post


Link to post
Share on other sites
Advertisement

it seems that following your guide these new drivers works fine on the first pc in signature...

if i'll report something strange i'll tell you. :)

cheers

 

p.s. thank you!

Share this post


Link to post
Share on other sites

After installing the driver for the 560 TI (which isn't supported according to http://www.nvidia.co...f03-driver.html), turning it on from system prefs doesn't stick after a reboot. Though turning it on and then running LuxMark resulted in a lock up (confirmed kexts were loaded via 3rd party system info app).

 

Cosmetic bug:

 

post-810930-0-83759400-1368747116_thumb.png

 

There is a way to force it but its not a good idea since it is cosmetic.

Share this post


Link to post
Share on other sites

After installing the driver for the 560 TI (which isn't supported according to http://www.nvidia.co...f03-driver.html), turning it on from system prefs doesn't stick after a reboot. Though turning it on and then running LuxMark resulted in a lock up (confirmed kexts were loaded via 3rd party system info app).

 

can you check your log looking for the lock up was caused by the well know NVDA Channel Exception??

 

Thanks, and good luck

Share this post


Link to post
Share on other sites

Cosmetic bug:

 

post-810930-0-83759400-1368747116_thumb.png

 

There is a way to force it but its not a good idea since it is cosmetic.

 

I am facing same problem

Share this post


Link to post
Share on other sites

I am facing same problem

 

Do you not see that the drivers are loaded in my pic? There is no issue just cosmetic. If you add the boot flag you can guarantee the drivers are loaded.

Share this post


Link to post
Share on other sites

Hi,. excuse me for the silly questions....

I am trying unsuccessfully -for several days :wallbash: -to get a GTX titan to work on my desktop...this is a real MacPro5,1 2xQuad-Core Intel Xeon, 96 GB RAM...with extra power supply.

I read somewhere that with 10.8.4 the titan was supported....but got a blank screen with the latest version. I was able to boot and login after I moved NVDA*.kext.

Now I see my two cards in System information, one is a Qadro 4000, the other ( Vendor: NVIDIA (0x10de) Device ID: 0x1005 Revision ID: 0x00a1 Kernel Extension Info: No Kext Loaded ). If I rebuild the cache and reboot it hangs, I tried installing NVidia drivers but they not compatible with 10.8.4.... I have no real clue if I should revert to 10.8.3 and use new nvidia drivers, change these kext things, any suggestions would be greatly apreciated.

 

S>

Share this post


Link to post
Share on other sites

Hi,. excuse me for the silly questions....

I am trying unsuccessfully -for several days :wallbash: -to get a GTX titan to work on my desktop...this is a real MacPro5,1 2xQuad-Core Intel Xeon, 96 GB RAM...with extra power supply.

I read somewhere that with 10.8.4 the titan was supported....but got a blank screen with the latest version. I was able to boot and login after I moved NVDA*.kext.

Now I see my two cards in System information, one is a Qadro 4000, the other ( Vendor: NVIDIA (0x10de) Device ID: 0x1005 Revision ID: 0x00a1 Kernel Extension Info: No Kext Loaded ). If I rebuild the cache and reboot it hangs, I tried installing NVidia drivers but they not compatible with 10.8.4.... I have no real clue if I should revert to 10.8.3 and use new nvidia drivers, change these kext things, any suggestions would be greatly apreciated.

 

S>

 

I have 4 users using the Titan in 10.8.4 with no issues with the latest beta. Start with Beta 12E36 and up.

Share this post


Link to post
Share on other sites

Mine Works simply from change within system prefs, without modifying org.chameleon or something.

I'm using Screen%20Shot%202013-05-18%20at%2010.27.46%20AM.png

 

Just to be sure the driver is loaded properly check CUDA prefpane from system pref

 

Screen%20Shot%202013-05-18%20at%2010.30.28%20AM.png

Share this post


Link to post
Share on other sites

Mine Works simply from change within system prefs, without modifying org.chameleon or something.

I'm using Screen%20Shot%202013-05-18%20at%2010.27.46%20AM.png

 

Please read the guide. It says only for fusion installs and or not having Nvram needs the flag. Thank you.

Share this post


Link to post
Share on other sites

How about a clover based installation ?

Tried to use the prefpane to switch to web drivers with clover 1654 and it didn't worked (not a cosmetic issue !) :wallbash:

It works with nvda_drv=1 in boot args.

Is there a way to make it work with the nvidia nvram prefpane with clover ?

Share this post


Link to post
Share on other sites

How about a clover based installation ?

Tried to use the prefpane to switch to web drivers with clover 1654 and it didn't worked (not a cosmetic issue !) :wallbash:

It works with nvda_drv=1 in boot args.

Is there a way to make it work with the nvidia nvram prefpane with clover ?

 

Did you check system info? Otherwise go ask Slice.

Share this post


Link to post
Share on other sites

Yes I checked the system info. Apple driver is loaded.

Nvidia driver can't be loaded unless I use the nvda_drv=1 boot arg.

Will ask Slice.

Thank you.

Share this post


Link to post
Share on other sites

Yes I checked the system info. Apple driver is loaded.

Nvidia driver can't be loaded unless I use the nvda_drv=1 boot arg.

Will ask Slice.

Thank you.

 

Thanks for the feed back. Let us know what slice says.

Share this post


Link to post
Share on other sites

 

5. From i386/modules folder copy FileNVRAM.dylib into Extra/modules:

screen-shot-2013-05-15-at-10-37-31-pm.png?w=640

 

I can't find this file.. (might be a noob thing sorry)

Share this post


Link to post
Share on other sites

[/center]

 

I can't find this file.. (might be a noob thing sorry)

 

Its in the folder of the complied binaries of the boot loader.

Share this post


Link to post
Share on other sites

Hey there,

 

I couldn't find it so I went ahead and tried the reboot without moving that file. It looks like my resolution problem has been fixed. :D

 

Here is a screenshot of the inside of my i386 folder... No modules folder, and no .dylib file. What does it do when you move the file? Should I bother pursuing it?

 

image.jpg

Share this post


Link to post
Share on other sites

Hey there,

 

I couldn't find it so I went ahead and tried the reboot without moving that file. It looks like my resolution problem has been fixed. :D

 

Here is a screenshot of the inside of my i386 folder... No modules folder, and no .dylib file. What does it do when you move the file? Should I bother pursuing it?

 

image.jpg

 

This is because you failed to follow the instructions on the guide on where to get the bootloader.

Share this post


Link to post
Share on other sites

I have 4 users using the Titan in 10.8.4 with no issues with the latest beta. Start with Beta 12E36 and up.

 

Thanks for the suggestion, but I can't seem to be able to locate anything older than 12E52..., maybe I 'll try again with 10.8.3, new drivers enabled, nvda_drv=1, and enabling the latest nvidia version, etc... I think I tried that already,...it seems to be something to do with the "Kernel Extension Info: No Kext Loaded"

is there another way to load this kext, I tried downloading some kext files, but osx86tool does not like to open those files, somehow hardware detection in osX86tools doesn't work, and I get errors when I try to setup a custom geforce card with over 5GB using GFX string creator.... any suggestion?

Thanks, S>

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

  • Similar Content

    • By arsradu
      Hi guys,
       
      So, this little tutorial is meant to help people fix a new issue with Mojave: opening jpg files (looks like it happens with larger files, not so much with smaller ones) using Quick Look (on spacebar key pressed) or Preview app using the dedicated GPUs.
      The issue is that, when trying to open jpg files (for some reason, it doesn't occur with pngs), the pop-up window would remain in an endless loop. And from that point forward, any other app trying to open the same file, will either become unresponsive (in the case of the Preview app) or just stay in endless loop (in case of Quick Look)
       
      As a workaround, in case the tutorial below doesn't work for you, you can use Grab (built-in app in MacOS) to open said files.
       
      This issues DOES NOT occur using the iGPU. So...if you're comfortable using the iGPU, just switch to that and you should be good to go. You don't need this tutorial.
       
      This tutorial is based on @Cyberdevs' research and testing on this topic (especially useful if you're using an RX series AMD video card), and my own very limited time testing this on my rig. I just thought it would be useful to create a new thread for it so that people can discuss this particular issue here.

       
      With that being said, I'll just move on to the actual tutorial.
       
      New (probably better) method.
       
      Switch your SMBIOS in Clover config in regards to the dedicated GPU you're using, instead of the CPU. Take a look here and choose the corresponding SMBIOS closest to your hardware.
      Yeah, that's it.
       
      In case you lose CPU PM with this method, there are multiple ways to fix that, and they do not make the object of this thread.  It worked perfectly fine for me though and I didn't need to change anything else.
       
      Alternatively, you can also try one of the MacPro SMBIOSes, such as MacPro5,1 or MacPro6,1. It was tested multiple times and it showed good results. 
      Keep in mind that choosing these will most likely not display you an iGPU in System Report, since those machines generally (probably all of them) use Xeon CPUs with no iGPUs either.
       
      A simple Lilu plugin
       
      Thanks to vulgo for the plugin and @Matgen84 for sharing and testing it, we now have yet another possible solution to this issue. It's a Lilu plugin (which means that it requires Lilu.kext in order to work) it's called NoVPAJpeg, and it can be downloaded from here.
      You use it as a regular kext, alongside Lilu. I prefer (and recommend) using it within the EFI folder, in Clover/kexts/Other. But if you want to use it in S/L/E, that should work just fine.
       
      More info can be found in this post here.
       
      In case you wanna try alternative medicine  ...
       
      The main thing we need to do is actually very simple: we need to enable the iGPU (integrated GPU) by default, in BIOS, while actually being connected to the dedicated GPU.
       
      For ATI GPUs (curtesy of @Cyberdevs' post here).
       
      The discreet GPU should be set as primary in BIOS and your displays should be connected to your discreet GPU
      The iGD/iGPU/iGFX must be enabled in BIOS/UEFI as the secondary GPU.
       
      I removed the DSDT Reconfigured the config.plist to and removed some flags which weren't needed because of the DSDT added the InjectATI + Ornioco Framebuffer (I removed the Orinoco framebuffer because of some weird glitches but it might be useful for other AMD GPUs if you inject the correct framebuffer according to your GPU. No need to inject anything for nVidia GPUs unless your GPU needs it) Added the InjectIntel + ig-platform-id for my intel HD530 (for intel HD 530 I use the 0x19120001 ig-platform-id)  
      Steps 5 to 7 might not be necessary I just did them anyway, if anyone succeeded to get quicklook working without those steps let me know so I can exclude those steps.

           Also confirmed to work with AMD R9 270X on a SandyBridge rig (without an iGPU) using WhateverGreen + Lilu kexts, which also helped to solve the transparency issue along with the Quick Look problem, it also resolved the buggy Screen Capture (Command+Shift+5) issue as well. 
       
          Also, do:
      nvram cleanup kextcache & prelinked kernel rebuild reboot  More info on that here.
       
       
      For NVIDIA GPUs
       
      This is what I needed for my NVIDIA rig and the MSI motherboard in the signature. Your milage may vary, depending on your hardware. But, the idea should be the same regardless.
       
         For MSi motherboads (check out my motherboard details in the signature)
       
      Set your BIOS to use PEG (dedicated GPU) so you can boot.
      Since you're reading this, chances are that you're already connected to the dedicated GPU and the BIOS is set to PEG. So just go into BIOS and see if you're using the iGPU or the dedicated (PEG) graphics device by default.
      If you're already using this, skip to Step 3 (the part AFTER Reboot). Reboot (I think you don't necessarily need to boot into the OS...but you do need to reboot at least until Clover UI) Reboot back into BIOS and set it from PEG to IGD (integrated graphics) WITHOUT removing the cable. Save Reboot again.  
         The result should be that the OS is tricked into thinking you're booting with the IGPU when in fact you're using the dedicated GPU.
       
         For Gigabyte motherboards
       
         Set:
      Peripherals -> Initial Display Output -> iGFX Chipset -> Internal Graphics -> Enabled
           DVMT Pre-Allocation -> default 64M.
           DVMT Total Gfx Mem -> default 256M Save Reboot  
         You should do this with the cable connected to the dedicated GPU.
       
      You will also need to patch the iGPU from Clover config.
      1. Graphics -> Inject Intel (UNCHECK Inject Nvidia or ATI if it's checked).
      2. ig-platform-id (in my case for Intel HD 4600 Desktop) it was: 0x0d220003
       
      And use IntelGraphicsFixup.kext (now included into WhateverGreen.kext) in Clover/kexts/Other. In my particular case, without the kext, I would get freezes in Safari.
       
      Another way to activate the iGPU while having the dGPU set as primary is by going into BIOS/SAG/Graphics Configuration/iGPU and setting Multi-Monitor to [Enabled], all the rest remains on [Auto], and set Inject/Intel to [True] in Clover config.*.
       
      Thanks to @gjw for the tip.
       
      *Motherboard dependent. But the multi-monitor option might be the same regardless of manufacturer.
       
       
      This might be a bug with Mojave and it could be fixed by Apple in the future. but in the meantime, feel free to use this if it helps you fix the issue.
       
      For HP motherboards (including laptops), so far no solution, unless the BIOS somehow allows to switch from eGPU to iGPU and vice-versa.
       
      Cheers!
    • By shmn
      Hi, I have been trying to follow @RehabMan 's guide to disable my dGPU ( https://www.tonymacx86.com/threads/guide-disabling-discrete-graphics-in-dual-gpu-laptops.163772/ ) for two days now, but haven't found a way to disable my 1080 Ti on my desktop Clover hack.

      I have a dual-boot (Win/Mac) machine on a Z370 Aorus 7. Since there's no nvidia drivers for Mojave atm, I want to stick with the iGPU under Mac, but still use the 1080 Ti under Win.

      The hack boots just fine when I remove the 1080 Ti, if it's plugged in via PCIe I get a kernel panic.
      Please find attached my ACPI folder, my clover config.plist and the kernel panic details.

      Might there even be a less hacky way than patching these ACPI files? I appreciate any kind of support! Thanks!
      ACPI.zip
      config.plist

      systemReport.txt
    • By rioredstone
      if a web driver is NOT installed, what happens??
       
      I used AMD RX560 in Sierra, but it is very unstable. usually gpu crash or window server crash occurred.
      so now I attached nvidea Quadro FX1700 to my Sierra. I know FX1700 works well in High Sierra, but I haven't tried it in Mojave yet.
      FX1700 works well without any additional kext, even nvidia web driver.
       
      I plan to buy a nividia card. I would like to know if a nvidia card, especially kepler, maxwell, pascal, works without web driver.
      just curious.
       
       
×