Cuda 11.7 - error: attribute "__malloc__" does not take arguments

[eric20]

Just upgraded Ubuntu from 20.04 LTS to 22.04 LTS. Facing the same problem as here…

CUDA 11.5 samples throw multiple ... error: attribute "__malloc__" does not take arguments

Hi, After updating the Ubuntu distribution (from 20.04 to 21.10), compiling the examples throws errors … … cd /usr/local/cuda/samples make make[1]: Entering directory ‘/usr/local/cuda-11.5/samples/0_Simple/vectorAdd_nvrtc’ GCC Version is greater or equal to 5.1.0 <<< make[1]: Nothing to be done for ‘all’. make[1]: Leaving directory ‘/usr/local/cuda-11.5/samples/0_Simple/vectorAdd_nvrtc’ make[1]: Entering directory ‘/usr/local/cuda-11.5/samples/0_Simple/template’ /usr/local/cuda-11.…

I have downloaded and installed hpc_sdk version 11.7.

(base) eric@cray:~$ gcc --version
gcc (Ubuntu 11.2.0-19ubuntu1) 11.2.0
Copyright (C) 2021 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

(base) eric@cray:~$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Tue_May__3_18:49:52_PDT_2022
Cuda compilation tools, release 11.7, V11.7.64
Build cuda_11.7.r11.7/compiler.31294372_0

[2358200405]

Also ,I met this problem on Ubuntu22.04,and later I try to compile it on Ubuntu20.04 ,It works well,resignedly

[brian.maddox]

It’s also happening with Cuda 11.2 on 22.04. Have to stay in the 11.2 range for Tensorflow. Is there a public patch we can apply that anyone knows about?

[marcel.meinders]

Dear all, I met the same problem with CUDA 11.4 on UBUNTU 22.04 LTS when making/compiling mnistCUDNN. Also want to use it for TensorFlow. Has anyone already an idea how to solve this?