error: identifier "__float128" is undefined

[Marty Chang]

On Ubuntu Server 14.04 LTS running on an Amazon EC2 g2.2xlarge instance, trying to compile Caffe gets me an "error: identifier "__float128" is undefined". Can anyone help me troubleshoot or resolve this problem?

Below is the output from the make command.

$ make all
PROTOC src/caffe/proto/caffe.proto
CXX .build_release/src/caffe/proto/caffe.pb.cc
CXX src/caffe/internal_thread.cpp
CXX src/caffe/util/db_lmdb.cpp
CXX src/caffe/util/blocking_queue.cpp
CXX src/caffe/util/db_leveldb.cpp
CXX src/caffe/util/cudnn.cpp
CXX src/caffe/util/math_functions.cpp
CXX src/caffe/util/hdf5.cpp
CXX src/caffe/util/signal_handler.cpp
CXX src/caffe/util/im2col.cpp
CXX src/caffe/util/benchmark.cpp
CXX src/caffe/util/io.cpp
CXX src/caffe/util/upgrade_proto.cpp
CXX src/caffe/util/db.cpp
CXX src/caffe/util/insert_splits.cpp
CXX src/caffe/solvers/rmsprop_solver.cpp
CXX src/caffe/solvers/adadelta_solver.cpp
CXX src/caffe/solvers/sgd_solver.cpp
CXX src/caffe/solvers/adam_solver.cpp
CXX src/caffe/solvers/nesterov_solver.cpp
CXX src/caffe/solvers/adagrad_solver.cpp
CXX src/caffe/data_transformer.cpp
CXX src/caffe/layer.cpp
CXX src/caffe/blob.cpp
CXX src/caffe/net.cpp
CXX src/caffe/layer_factory.cpp
CXX src/caffe/layers/embed_layer.cpp
CXX src/caffe/layers/euclidean_loss_layer.cpp
CXX src/caffe/layers/loss_layer.cpp
CXX src/caffe/layers/contrastive_loss_layer.cpp
CXX src/caffe/layers/neuron_layer.cpp
CXX src/caffe/layers/hdf5_output_layer.cpp
CXX src/caffe/layers/image_data_layer.cpp
CXX src/caffe/layers/concat_layer.cpp
CXX src/caffe/layers/softmax_layer.cpp
CXX src/caffe/layers/batch_reindex_layer.cpp
CXX src/caffe/layers/filter_layer.cpp
CXX src/caffe/layers/spp_layer.cpp
CXX src/caffe/layers/cudnn_lrn_layer.cpp
CXX src/caffe/layers/exp_layer.cpp
CXX src/caffe/layers/infogain_loss_layer.cpp
CXX src/caffe/layers/sigmoid_layer.cpp
CXX src/caffe/layers/hdf5_data_layer.cpp
CXX src/caffe/layers/dummy_data_layer.cpp
CXX src/caffe/layers/tanh_layer.cpp
CXX src/caffe/layers/bnll_layer.cpp
CXX src/caffe/layers/power_layer.cpp
CXX src/caffe/layers/batch_norm_layer.cpp
CXX src/caffe/layers/hinge_loss_layer.cpp
CXX src/caffe/layers/base_conv_layer.cpp
CXX src/caffe/layers/pooling_layer.cpp
CXX src/caffe/layers/cudnn_relu_layer.cpp
CXX src/caffe/layers/memory_data_layer.cpp
CXX src/caffe/layers/inner_product_layer.cpp
CXX src/caffe/layers/reduction_layer.cpp
CXX src/caffe/layers/reshape_layer.cpp
CXX src/caffe/layers/multinomial_logistic_loss_layer.cpp
CXX src/caffe/layers/accuracy_layer.cpp
CXX src/caffe/layers/tile_layer.cpp
CXX src/caffe/layers/eltwise_layer.cpp
CXX src/caffe/layers/cudnn_sigmoid_layer.cpp
CXX src/caffe/layers/silence_layer.cpp
CXX src/caffe/layers/slice_layer.cpp
CXX src/caffe/layers/softmax_loss_layer.cpp
CXX src/caffe/layers/sigmoid_cross_entropy_loss_layer.cpp
CXX src/caffe/layers/conv_layer.cpp
CXX src/caffe/layers/dropout_layer.cpp
CXX src/caffe/layers/absval_layer.cpp
CXX src/caffe/layers/im2col_layer.cpp
CXX src/caffe/layers/lrn_layer.cpp
CXX src/caffe/layers/prelu_layer.cpp
CXX src/caffe/layers/cudnn_conv_layer.cpp
CXX src/caffe/layers/log_layer.cpp
CXX src/caffe/layers/relu_layer.cpp
CXX src/caffe/layers/cudnn_softmax_layer.cpp
CXX src/caffe/layers/deconv_layer.cpp
CXX src/caffe/layers/threshold_layer.cpp
CXX src/caffe/layers/argmax_layer.cpp
CXX src/caffe/layers/window_data_layer.cpp
CXX src/caffe/layers/mvn_layer.cpp
CXX src/caffe/layers/split_layer.cpp
CXX src/caffe/layers/base_data_layer.cpp
CXX src/caffe/layers/flatten_layer.cpp
CXX src/caffe/layers/cudnn_tanh_layer.cpp
CXX src/caffe/layers/data_layer.cpp
CXX src/caffe/layers/cudnn_pooling_layer.cpp
CXX src/caffe/layers/cudnn_lcn_layer.cpp
CXX src/caffe/parallel.cpp
CXX src/caffe/syncedmem.cpp
CXX src/caffe/solver.cpp
CXX src/caffe/common.cpp
CXX src/caffe/data_reader.cpp
NVCC src/caffe/util/math_functions.cu
./include/boost/config/suffix.hpp(510): error: identifier "__float128" is undefined


./include/boost/smart_ptr/detail/sp_counted_base_gcc_x86.hpp(75): warning: variable "tmp" was set but never used


1 error detected in the compilation of "/tmp/tmpxft_00002146_00000000-16_math_functions.compute_50.cpp1.ii".
make: *** [.build_release/cuda/src/caffe/util/math_functions.o] Error 1

[Fangzou Liao]

On my computer, I use the boost 1.60, this problem is caused by its incompatibility with CUDA.

https://github.com/boostorg/config/pull/83

the solution is get all files under https://github.com/boostorg/config

and copy the folder include/boost/  to the corresponding folder of your boost

or another way is install lower edition boost

...

[helxsz]

I had this problem before when I used boost 1.6.  try boost 1.5.5, it should be working.

...

[Marty Chang]

I no longer get this error after downgrading to Boost 1.55.0. Thank you!