Problem with cudnn samples v8 with cuda toolkit 12.1

Graphics / Linux Linux
1

Hi,
I have cuda toolkit v12.1 and a driver 530.30.02 for ubuntu 20 and I have GPU RTX 3060 Laptop. My main problem is that after installing cudnn for cuda toolkit, I face a problem when testing the proper installation of cudnn. I install cudnn-local-repo-ubuntu2004-8.8.1.3_1.0-1_amd64 but when I run the test for cudnn using cudnn_samples_v8/mnistCUDN, I get this error. Does that mean I installed cudnn properly and there is an incompatibility in the make file of the package? Can you please help me? Could it be that I am not installing Cudnn properly?

CUDA_VERSION is 12010
Linking agains cublasLt = true
CUDA VERSION: 12010
TARGET ARCH: x86_64
HOST_ARCH: x86_64
TARGET OS: linux
SMS: 35 50 53 60 61 62 70 72 75 80 86 87
/usr/local/cuda/bin/nvcc -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include -ccbin g++ -m64 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_62,code=sm_62 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_87,code=compute_87 -o fp16_dev.o -c fp16_dev.cu
nvcc fatal : Unsupported gpu architecture ‘compute_35’
make: *** [Makefile:241: fp16_dev.o] Error 1

2

compute_35,code=sm_35 is not supported any more and you also don’t need it (it’s for old Kepler based gpus). Just remove.

3

I would be so grateful if you can explain to me where can we remove it. I couldn’t find it in the makefile of the sample test
Thanks,
Mostafa

4

Thank you I figured it out.

5

I am facing the same issue. Can you please share how did you remove it ?

1 Like
6

find Makefile
vim Makefile
find below code ,and add filter-out 35, replace orginal code
ifeq ($(GENCODE_FLAGS),)
#Generate SASS code for each SM architecture listed in $(SMS)
#$(foreach sm,$(SMS),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
$(foreach sm,$(filter-out 35,$(SMS)),$(eval GENCODE_FLAGS += -gencode arch=compute_$(sm),code=sm_$(sm)))
:wq
after this Makefile edit
u can run you’re testing mnistCUDN