Hanjun-Dai / graphnn

Training computational graph on top of structured data (string, graph, etc)
MIT License
288 stars 91 forks source link

Do not find the 'modify' command #15

Closed FFrankyy closed 6 years ago

FFrankyy commented 6 years ago

I have problems in Build Static Library section, modify configurations in make_common file make -j8 these two commands do not work: do not find the modify command, what's the problem?

Hanjun-Dai commented 6 years ago

Hi there, could you provide more details of your problem?

You can paste your modified make_common file, and the relevant error log, your os detail (mac, ubuntu), etc.

FFrankyy commented 6 years ago

make_common file: dir_guard = @mkdir -p $(@D)

CUDA_HOME := /usr/local/cuda INTEL_ROOT := /opt/intel MKL_ROOT = $(INTEL_ROOT)/mkl TBB_ROOT = $(INTEL_ROOT)/tbb USE_GPU = 1

FIND := find CXX := g++ NVCC := $(CUDA_HOME)/bin/nvcc NVCCFLAGS += --default-stream per-thread CXXFLAGS += -Wall -O3 -std=c++11 LDFLAGS += -lm -L$(CUDA_HOME)/lib64 -lcudart -lcublas -lcurand -lcusparse -lmkl_rt -ltbb -lfmt CUDA_ARCH := -gencode arch=compute_30,code=sm_30 \ -gencode arch=compute_35,code=sm_35 \ -gencode arch=compute_50,code=sm_50 \ -gencode arch=compute_50,code=compute_50

I do not know how to modify configurations in make_common file, and when I enter "make -j8" it failed: In file include/nn/variable.h:7:24:fatal error:fmt/printf.h: no such file or directory

Could you please tell me how can I address this problem?

Hanjun-Dai commented 6 years ago

Could you please sync the latest version of my code and try again?

I should have removed that dependency already.

FFrankyy commented 6 years ago

I am sorry maybe it's the version problem, now the above problem does not occur, but a new problem comes, when I enter "make -j8", it still failed Makefile:53:recipe for target 'build objs/cxx/nn/row_selection.o' failed make: *** [build/objs/cxx/nn/row_selection.o] Error 1

Do you know why? Thanks!

Hanjun-Dai commented 6 years ago

have you tried to do make clean first?

FFrankyy commented 6 years ago

I have tried, but still failed, returned the same problem

FFrankyy commented 6 years ago

And I still do not know how to: modify configurations in make_common file

Hanjun-Dai commented 6 years ago

could you try the single thread build: make and then give me the snapshot of the error log? Instead of the last line of log, I need something like which line triggers the error.

FFrankyy commented 6 years ago

when I run make it returned:

g++ -Wall -O3 -std=c++11 -I/usr/local/cuda/include -I/opt/intel/mkl/include -I/opt/intel/tbb/include -Iinclude -fPIC -DUSE_GPU -MMD -c -o build/objs/cxx/nn/row_selection.o src/nn/row_selection.cpp src/nn/row_selection.cpp:3:21: fatal error: tbb/tbb.h: no such file or directory

include "tbb/tbb.h"

                 ^

compilation terminated. Makefile:53: recipe for target 'build/objs/cxx/nn/row_selection.o' failed make: *** [build/objs/cxx/nn/row_selection.o] Error 1

Hanjun-Dai commented 6 years ago

Thanks. Then it makes sense.

Have you installed intel parallel studio? If so, the default path should be /opt/intel. Check and see whether you have it. If not, you need to first install the intel parallel studio. Students will have free license.

FFrankyy commented 6 years ago

Thanks, I have just installed intel parallel studio, and the previous problems seemed resolved, but when I run the mnist example and enter command: make it returned problems: g++ -Wall -O3 -std=c++11 -DUSE_GPU -I/usr/local/cuda/include -I/opt/intel/mkl/include -I../../include -Iinclude -o build/mnist mnist.cpp ../../build/lib/libgnn.a -L../../build/lib -lgnn -lm -lmkl_rt -ltbb -L/usr/local/cuda/lib64 -lcudart -lcublas -lcurand -lcusparse /usr/bin/ld: 找不到 -lmkl_rt /usr/bin/ld: 找不到 -ltbb collect2: error: ld returned 1 exit status Makefile:21: recipe for target 'build/mnist' failed make: *** [build/mnist] Error 1

How can it be solved? Thanks!

Hanjun-Dai commented 6 years ago

As I mentioned in the README, please run the intel script to setup environment:

in ~/.bashrc, add the following sentence

source {path_to_your_intel_root/name_of_parallel_tool_box}/bin/psxevars.sh

For example: source /opt/intel/parallel_studio_xe_2017.4.056/bin/psxevars.sh

Please adapt the path to your own system

FFrankyy commented 6 years ago

Good! Finally Got it!! Thanks!

Hanjun-Dai commented 6 years ago

close the issue since it is resolved.