weiliu89 / caffe

Caffe: a fast open framework for deep learning.
http://caffe.berkeleyvision.org/
Other
4.77k stars 1.67k forks source link

Compiler error[Unsupported gpu architecture 'compute_61'] #602

Open Jimjipeng opened 7 years ago

Jimjipeng commented 7 years ago

NVCC src/caffe/layers/cudnn_softmax_layer.cu nvcc fatal : Unsupported gpu architecture 'compute_61' make: [.build_release/cuda/src/caffe/layers/cudnn_softmax_layer.o] Error 1 make: Waiting for unfinished jobs.... NVCC src/caffe/layers/cudnn_relu_layer.cu nvcc fatal : Unsupported gpu architecture 'compute_61' make: *** [.build_release/cuda/src/caffe/layers/cudnn_relu_layer.o] Error 1

sangthian commented 7 years ago

delete 'compute_61' line in makefile.config, your gpu's computing ability is less than 6.1(the concrete value you can find in nvidia's website)

Jimjipeng commented 7 years ago

Many thanks