20.04 drivers not able to be installed

Graphics / Linux Linux
, ,
1

I’m running an AWS instance, with the following stats:

OS:
NAME=“Ubuntu”
VERSION=“20.04.3 LTS (Focal Fossa)”

GPU:
product: TU104GL [Tesla T4]
vendor: NVIDIA Corporation
width: 64 bits

I’m trying to install CUDA 11.2.0_460.27.04_linux, the cuda version suggested by the Nvidia website, and running it in a new tmp folder in /mnt/data/. This is so I can later use tensorflow-gpu

I’ve come to an error that gives this output on the nvidia log

...
   make[2]: *** [scripts/Makefile.build:297: /mnt/data/tmp/selfgz1733/NVIDIA-Linux-x86_64-460.27.04/kernel/nvidia-drm/nvidia-drm-gem-nvkms-memory.o] Error 1
   make[2]: Target '__build' not remade because of errors.
   make[1]: *** [Makefile:1907: /mnt/data/tmp/selfgz1733/NVIDIA-Linux-x86_64-460.27.04/kernel] Error 2
   make[1]: Target 'modules' not remade because of errors.
   make[1]: Leaving directory '/usr/src/linux-headers-5.15.0-1036-aws'
   make: *** [Makefile:80: modules] Error 2
-> Error.
ERROR: An error occurred while performing the step: "Building kernel modules". See /var/log/nvidia-installer.log for details.
-> The command `cd ./kernel; /usr/bin/make -k -j16  NV_EXCLUDE_KERNEL_MODULES="" SYSSRC="/lib/modules/5.15.0-1036-aws/build" SYSOUT="/lib/modules/5.15.0-1036-aws/build"` failed with the following output:

make[1]: Entering directory '/usr/src/linux-headers-5.15.0-1036-aws'
warning: the compiler differs from the one used to build the kernel
  The kernel was built by: gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
  You are using:           cc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
  SYMLINK /mnt/data/tmp/selfgz1733/NVIDIA-Linux-x86_64-460.27.04/kernel/nvidia/nv-kernel.o
  SYMLINK /mnt/data/tmp/selfgz1733/NVIDIA-Linux-x86_64-460.27.04/kernel/nvidia-modeset/nv-modeset-kernel.o
 CONFTEST: hash__remap_4k_pfn
...

After searching, I have found a similar problem, yet there are no drivers installed on the aws, let alone an ‘open’ one. I was looking to install the 460 driver through the CUDA installation as having it pre-installed caused warnings from the CUDA installer. I’m wondering if I’m missing something obvious (or not obvious) to try and get this working.

Any ideas? Thanks in advance!