EOL'd Quadro won't work with driver 388.71 despite support for Geforce equivalents

I'm really confused. I've been trying for the past couple days to get my Fermi-based Quadro 1000M to work with the latest Geforce driver version 388.71 (which happens to support the 500M series, the Geforce equivalents to my Quadro 1000M), after migrating through a complete uninstall w/ DDU from an outdated version 362.77 Quadro driver installation, which has been the last official update released by Nvidia for my graphics card for a while now, with the update being released in mid-2016.

But let's get back on to topic. The 388.71 driver doesn't seem to be working properly, introducing graphical glitches and wrong colors in applications like RetroArch (which by default uses OpenGL as its display driver) and crashes it when I turn on Vulkan support. It also crashed some other applications. I ended up having to roll back to the old driver for stability reasons.

Is there any way to fix this or am I going to have permanently settle for the outdated version 362.77?

 

Which OS?

 

Windows 7 Professional Edition x86-64.

I gave it another go today, making sure to restart the OS once the version 382.71 driver got installed, but still no luck. I couldn't even get to the Nvidia control panel, getting an error saying that "Display settings are not available".

Grrr. There has to be a way. An Nvidia Quadro should still be essentially the same to a Geforce in the hardware department... right?

 

I don't personally have a solution to this, but there are 1 or 2 people here on Guru3D that I've seen offer custom inf modded drivers. Their forum handles escape me at the moment, but perhaps if they see this they can help or supply some answers.

 

Ive added the quadro 1000m to the inf for 388.71 win7

here is the plain text

https://paste.ofcode.org/m9e7FyZWWn8Cpfm25x3PeN

paste into a plain text file, then rename it to nv_dispi.inf, and replace the inf in the C:\NVIDIA\DisplayDriver\388.71\display.driver folder

then install via the setup.exe under disable driver signature enforcement mode.

should install normally,

there is an alternative edit that can be done, but you should check to see if this works first

also fermi based gpus didn't get vulkan support, even thought the hardware is capable there is no vulkan driver.

 

Really? So that explains the reason why I kept getting those error messages in the virtual console for the vulkaninfo executable, located in the VulkanRT directory in Program Files (x86). Which is a shame, honestly, because besides from the security fixes included with the newer versions of the Nvidia drivers, I was actually really excited for the included Vulkan API support. Thermal throttling of my quad-core Intel i7 CPU is an issue that I experience on my computer on the regular, slowing down my CPU to a crawl, even with thermal paste applied. So having driver support for Vulkan, which purports to reduce power consumption (which generates heat) *AND* CPU usage is almost a godsend, or at least for my computer.

But enough rambling (again). I edited my nv_dispi.inf to include user1's changes and installed, and... still no dice. I'm experiencing the same issues as before, with the same wrong colors and graphical artifacts in RetroArch as before, and I still keep getting the same error messages when I try to enter the Nvidia control panel, as shown below:



Arggh.

 

if thats the case, then well have to see what a more complicated edit will do, so i have to ask, what is the full device id

in my case the full device id for a 560m is
PCI\VEN_10DE&DEV_1251&SUBSYS_204A1043&REV_A1

this can be found under hardware ids in the details section of properties of the gpu in question in device manager

basically the edit i provided earlier is a generic edit, so the next step is to find if your specific variant of the quadro 1000m has its own configuration. (which it seems that it does) and find the closest geforce variant and mimic that instead.

i should also mention, you should probably disable secure boot when testing, it can cause issues if your using a modified inf.

 

I had a look around the device manager a bit, and here's what I found for my Quadro 1000M's hardware ID: PCI\VEN_10DE&DEV_0DFA&SUBSYS_21D117AA&REV_A1
It should be what you're looking for.

And my UEFI BIOS doesn't even have an option for Secure Boot, and I boot Windows through the UEFI BIOS' legacy mode, which enables the UEFI BIOS to emulate a classical BIOS, so there should be no problems regarding it.

 

ok so the good news, is that it would appear that there is a quadro 2000m uses a mostly identical config, and shares the same subsys id (from the same company i assume)

https://paste.ofcode.org/cxwDybyLvmsq3ZEj7ya6r3

here is another inf however this one is nvlti.inf which is the original location for your gpu in the r362 driver,

install the same way i said earlier just replacing nvlti.inf instead.

also i forgot , fastboot may cause problems aswell, recommend turning that off aswell, even if it doesn't help getting the driver to install, it will help recover windows if you encounter the dreadful nvidia black screen issue that can be triggered with an incorrect inf.

if this works going forward you'll essentially you want to mimic the NVIDIA_DEV.0DDA.21D1.17AA = "NVIDIA Quadro 2000M" entry since it has the closest configuration.

all you will have to do is find the NVLT inf and change all entrys of this string NVIDIA_DEV.0DDA.21D1.17AA to NVIDIA_DEV.0DFA.21D1.17AA

and of course change the string NVIDIA_DEV.0DDA.21D1.17AA = "NVIDIA Quadro 2000M" to NVIDIA_DEV.0DFA.21D1.17AA = "NVIDIA Quadro 1000M" if you want it to be named correctly.



and of course if this doesn't work, then clearly there is a problem with the driver that nvidia didn't bother to fix or there is a hw bug with this specific revision of Gf108 that nvidia decided to ignore rather than work around.

the most odd part is that the gt 550m is essentially exactly the same , just clocked alittle faster and uses a marginally newer chip revision.