Compilation of Pytorch binaries stops in the middle of the process

Hello, I would like to run my NVIDIA TESLA K40m hardware with Pytorch however Pytorch no longer officially supports below CUDA 3.7. However, my NVIDIA TESLA K40m runs at maximum with CUDA 3.5. Fortunately there is a method (https://medium.com/@jeremistderechte/how-to-get-pytorch-working-on-kepler-3-5-gpus-e-g-tesla-k40-f7275a23b186) which consists of downloading the Pytorch binaries and specify the variable export TORCH_CUDA_ARCH_LIST="3.5" to run Pytorch on my NVIDIA card. The compilation goes well up to 71% with the error:

[71%] Built target microkernels-all
[71%] Built target fbgemm_avx2
gmake: *** [Makefile:146:all] Error 2

I’m running Ubuntu Server 22.04 LTS.
My version of gcc (g++) is: 10.5.0
My version of cmake is: 3.22.1
For the NVIDIA card I have the 470.239.06 driver and CUDA 11.4.
My variables in the .bashrc file are:

export PATH=/usr/local/cuda-11.4/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-11.4/lib64:$LD_LIBRARY_PATH
export CPATH=$CPATH:/usr/local/cuda/include
export LIBRARY_PATH=$LIBRARY_PATH:/usr/local/cuda/lib64
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/nvidia/lib64
export PATH=/usr/local/cuda-11.4/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-11.4/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export TORCH_CUDA_ARCH_LIST="3.5"

Thank you in advance for your help,