How to solve this error cudacap flag 30 ()

I just download NVC 21.7 in ubuntu

but cant run any OpenACC code

pgcc -fast -acc -Minfo=accel acc_c1.c 
NVC++-S-0155-Invalid value after compiler -cudacap flag 30 ()
NVC++/x86-64 Linux 21.7-0: compilation completed with severe errors
oliver@oliver:~/OpenACC/samples/acc_c1$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Wed_Jun__2_19:15:15_PDT_2021
Cuda compilation tools, release 11.4, V11.4.48
Build cuda_11.4.r11.4/compiler.30033411_0

I found the version of my cuda in this code

Is it related??
how can I solve it?

is there some rure that is not running?

it run but not completly when I use

nvc -Minfo=accel test.c -fast -O3 -acc=gpu -gpu=cc60,cuda11.0 -o ttt

but some thing missin. It says:

Rebuild the application with -ta=tesla:cc30 to use NVIDIA Tesla GPU 0

but when i try:

nvc -Minfo=accel test.c -fast -O3 -acc=gpu -ta=tesla:cc30 -o ttt
nvc-Error-The -ta=tesla:cc30 option is no longer supported

another question , through the image it says my CUDA 11.2
but in the compiler it accept only 11.0

Hi Huda22,

Since NVIDIA discontinued support for Kepler architectures (CC 3.x), this support was removed from the compilers as of the 21.3 release. (See: Release Notes Version 21.7).

-Mat

1 Like

Thank you Mat for reply
I will try to download older version
because I cant upgrade my notebook and I cant bay new one now
cause I already have new microsof surface