Tesla P4 - 8GB / US$ 100 or even lower

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

jycai

New Member
Jan 30, 2021
5
1
3
"Brave" to try a driver from random source.

But thnx to do so :D;)
Love tesla as a car too:Dand don't afraid. Of couse I tried it in my test server on vlan.
I am planning to move my VMs from Esxi to Proxmox to utilize the P4, so some kind risk is worth to take at testing stage to know the enemy.
 
  • Like
Reactions: Clampazzo

Clampazzo

New Member
Oct 18, 2020
5
1
3
Thank you Clampazzo! I changed the VM from UEFI to default and follow this instruction Install drivers for NVIDIA RTX virtual workstations | Compute Engine Documentation | Google Cloud to install the grid driver in ubuntu 22.04, nvidia-smi is working and no crash after the reboot.

Since I am running the grid driver without license, it should only work for 20 minutes or so. Do you have the script to make it run for 24 hours?
Nice glad it works!

I have not dug into the script, but I suspect there is something in the vgpu discord, but I have not had the time to figure it out on my own. If you find one please let me know.
 
  • Like
Reactions: jycai

jycai

New Member
Jan 30, 2021
5
1
3
Nice glad it works!

I have not dug into the script, but I suspect there is something in the vgpu discord, but I have not had the time to figure it out on my own. If you find one please let me know.
I found the solution here - KrutavShah / vGPU_LicenseBypass - Linux bypass. However the method only works up to Nividia's vGPU version 14.1(510.73.06) ,later version is no go.

1.Use the command to make the driver run for 24 hours.
Code:
echo 'options nvidia NVreg_RegistryDwords="UnlicensedUnrestrictedStateTimeout=0x5A0;UnlicensedRestricted1StateTimeout=0x5A0"' | sudo tee /etc/modprobe.d/nvidia.conf
sudo update-initramfs -u
reboot

2. Reload the driver by adding a job in /etc/crontab with the below code or reboot Ubuntu VM in proxmox everyday.
Code:
echo "1" > /sys/bus/pci/devices/pci_id_of_vgpu/rescan
it shoud look like - * 3 * * * root echo "1" > /sys/bus/pci/devices/0000:00:10.0/rescan
 
Last edited:

Slothstronaut

Member
Apr 27, 2022
25
53
13
Anyone know of a way to spoof a PCIe ID in a bare-metal Windows? I want this card to operate in WDDM alongside a 2080Ti but it freaks out when the GRID driver is replaced by the GeForce driver.
 

blackholevault

New Member
Dec 24, 2022
1
0
1
Ok... this is the thread that I've desperately been seeking.


Question for the people infinitely more knowledgeable than myself. I have a small form factor PC (i3-10105 with integrated graphics and running Windows 10) that I'm using primarily as a Blue Iris server. However I'd like to occasionally use it for gaming and I'd also like to use a low profile GPU to hardware accelerate my Blue Iris. Would the Tesla P4 be a good sub $100 value proposition to help accomplish these goals? As an aside, I know absolutely nothing about virtualization, but I enjoy tinkering, so I'm not sure to what degree that would be necessary, but I'm willing to learn.
 

Xeroxxx

Member
Mar 12, 2020
89
55
18
I tried that as well. They want proof of your organisation, this will be forwarded to NVIDIA. This doesn't change that it stays B2B.

Howerver give it a try.
 
Last edited:

custom90gt

Active Member
Nov 17, 2016
180
68
28
38
Going to throw one of these in my backup servers for plex transcoding. Not too shabby of a card for a 2u system and $90 bucks.