Compiling 2.79b with CUDA support

Hello, may be someone can help with this question,

I have access to a machine with RTX 6000 GPUs, but I do not want to use Blender 2.80 yet… So I am trying to compile Blender 2.79b with CUDA 10(.1) RTX support. But when I follow the instructions here:

and compile I get the 2.80 version. I know next to nothing about compiling software, can anybody give me the steps to get and compile 2.79b with CUDA? And does it make any difference for the build when I compile on a machine that does not have the RTX cards (but does have the CUDA 10.1 toolkit installed)

I think it should just be an extra command to set the version I want with git, and somehow specifying the versions in these steps:

cd C:\blender-git

svn checkout lib/win64_vc14

cd C:\blender-git

git clone git://

cd blender

git submodule update --init --recursive

git submodule foreach git checkout master

git submodule foreach git pull --rebase origin master

cd C:\blender-git\blender
make full

A while ago 2.80 became master. 2.79 is a branch now.
Try this:

git checkout blender2.7
git pull

Then try to build again.

I’m not used to building under Windows but it should work.

Edit: Oh, and welcome to BA.:wink:

Hi, Chalybeum, thanks for the suggestion. Where in the process should I use these commands though, just before I make the build?

Yes, you simply checkout the branch you want and then build it.

See Building Blender and Tools/Git in the Blender docs.

Edit: Ah this might still be confusing. So: since you already downloaded the Blender sources and intialized your local repo you don’t need to do all of it again.
The two lines I gave you you and an additional

make full
make install

Should give you the latest 2.79

Just in case building is too much of a hassle, there are nightly builds of 2.79 build with the latest cuda at

Yes ! It worked well ! Thanks so much for the advice and the link for the 2.79 build.