Blender Q: amdgpu or admgpu-Pro What driver "should you use" ❓

I’ll start with system specs using screenfetch :

In my current workflow, I have the amdgpu driver that is already in the Linux kernel. I have heard but can’t confirm that the amdgpu-Pro driver is needed for better performance with Blender.
I went through the settings and the gpu does not show up in when choosing a render.

I did a cpu render with no problem on a small project, but that can be an issue down the line with scenes, animations etc.

Does anyone have any comment or personal experience with using the open source driver vs. the proprietary driver for rendering in Blender?

Currently I use open source driver and it’s pretty good. Using either one gives much better results than cpu.
Proprietary driver is supposed to give better performance, but I have no way of comparing them.
I did have some trouble with it last time I used it, but I can’t say for sure that the driver was to blame for that. I don’t have the same problem on open source driver, but it’s entirely new installation, so the problem might be with something else.

You can also check the benchmarks: AMDGPU-PRO - Phoronix
It looks like amdgpu-pro often gets worse results in games.

Is it an improvement over the Cycles render? I can’t even choose the gpu in OpenCL to even begin to compare. I just did a nuke and pave last week with this system and outside of wanting better performance for Rendering everything just works. What render do you use? EEVEE or Cycles?

I usually work on eevee, just because I can see the results in real time, but once I’m in the finishing stage of the project I plan to move into cycles and then focus more on textures and everything that might look different on eevee.
Working on GPU is noticeably better but cycles still take time, especially on higher settings. If gpu option isn’t available in blender, try getting it from instead of repository.

1 Like

Hey there @Definitive_Linux ! I know this is an older post, but I wanted to share with you that in order to get HIP working in Blender, it takes a little massaging to get the OpenCL stuff you need out of the AMD proprietary driver (if you didn’t know already).

This way you can keep the standard AMD driver and still make use of Blender’s HIP rendering engine for Cycles!

Here’s a link to a post I wrote about how I did it on the Fedora discourse:

And here’s a copy paste in case you don’t want to head off site. Hope this helps!

Hey all,

I recently joined this community last week and I experienced quite a faff in order to get my use case for this desktop to work. I did some searching on this Discourse and couldn’t find the info I needed, so I thought I would contribute it here in case anyone needed it in the future.

Here’s my use case:
I’m a Blender user who switched to Fedora. I recently acquired an AMD card which falls under the newer proprietary AMD driver capabilities (Radeon RX 6700 XT). I want to make use of Blender’s HIP GPU rendering under AMD which doesn’t work without their proprietary OpenCL magic.

Here’s what I had to do to make it work! I’m going to start out by saying I did not figure this process out on my own. I followed a guide here from reddit:

Step 1: Add /etc/yum.repos.d/rocm.repo with these contents:


Step 2: Download the official AMDGPU driver installer for RHEL 9:

Then install it using sudo dnf install ./amdgpu-install-

Step 3: Navigate to: Index of /amdgpu/22.10.3/rhel/8.6/main/x86_64/

And download these files:

Step 4: Install the above packages from their location (Downloads folder) using this command (all at once):

sudo dnf install amdgpu-core- libdrm-amdgpu-common- libdrm-amdgpu-

Now, DNF told me that amdgpu-core failed to install, and that the Transaction failed, but the other two packages installed successfully which is what’s important!

Step 5: Install the rest of the components with the installer installed previously from the (also already added) ROCm repo:

amdgpu-install --usecase=opencl,hip --no-dkms

Now open up Blender 3.2 and check the system settings; you should see the HIP tab and your GPU there!