原因:设置的算力和当前的gpu版本不匹配
算力匹配
gpu
百度或官网查询即可
如我的gtx2060,算力7.5
应对应设置 -gencode arch=compute_75,code=sm_75
cuda支持
到安装cuda时安装的samples查询
- cd /usr/local/cuda-8.0/samples/1_Utilities/deviceQuery
- sudo make
"/usr/local/cuda-8.0"/bin/nvcc -ccbin g++ -I../../common/inc -m64
-gencode arch=compute_20,code=sm_20
-gencode arch=compute_30,code=sm_30
-gencode arch=compute_35,code=sm_35
-gencode arch=compute_37,code=sm_37
-gencode arch=compute_50,code=sm_50
-gencode arch=compute_52,code=sm_52
-gencode arch=compute_60,code=sm_60
-gencode arch=compute_60,code=compute_60
-o deviceQuery.o -c deviceQuery.cpp
nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
-----
明显我的gpu版本算力超过6.0,这是我出错的原因,如果你的算力在其中,则可能Makefile,config,设置没设置对你相应算力。设置同查到的即可,去掉deprecated部分
我的则更换为cuda10,0后,再查:
- cd /usr/local/cuda-8.0/samples/1_Utilities/deviceQuery
- sudo make
/usr/local/cuda-10.0"/bin/nvcc -ccbin g++ -I../../common/inc -m64
-gencode arch=compute_30,code=sm_30
-gencode arch=compute_35,code=sm_35
-gencode arch=compute_37,code=sm_37
-gencode arch=compute_50,code=sm_50
-gencode arch=compute_52,code=sm_52
-gencode arch=compute_60,code=sm_60
-gencode arch=compute_61,code=sm_61
-gencode arch=compute_70,code=sm_70
-gencode arch=compute_75,code=sm_75
-gencode arch=compute_75,code=compute_75
-o deviceQuery.o -c deviceQuery.cpp
有-gencode arch=compute_75,code=sm_75,ok,cuda10何以支持了,再设置caffe的Makefile.config
Makefile.config
CUDA_ARCH := -gencode arch=compute_30,code=sm_30 \
-gencode arch=compute_35,code=sm_35 \
-gencode arch=compute_37,code=sm_37 \
-gencode arch=compute_50,code=sm_50 \
-gencode arch=compute_52,code=sm_52 \
-gencode arch=compute_60,code=sm_60 \
-gencode arch=compute_61,code=sm_61 \
-gencode arch=compute_70,code=sm_70 \
-gencode arch=compute_75,code=sm_75 \
-gencode arch=compute_75,code=compute_75
与第二步对应即可