kyuepublic / ptychopy Goto Github PK
View Code? Open in Web Editor NEWHigh performance ptychography reconstruction python package running on GPU
High performance ptychography reconstruction python package running on GPU
I run on the code on CUDA 11.2 with GCC 9.4 on Ubuntu OS.
The program can make the right reconstruction result. In src/examples/testEPIE.cpp, after write the result to disk "phaser->writeResultsToDisk; ", the program delete the pointer of class IPhaser "delete phaser; ", then the program got the "Segmentation fault".
I try to figure its location, but have no idea. Maybe the CUDA version make the problem. Would you please help me to figure it out?
I cannot compile ptychopy because the extra compiler options are not being passed correctly?
(ptychopy) dching@RADIO:~/Documents/ptychopy$ python setup.py build -v
running build
running build_ext
building 'ptychopy' extension
Warning: Can't read registry to find the necessary compiler setting
Make sure that Python modules winreg, win32api or win32con are installed.
C compiler: g++ -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -pipe -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -pipe -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /home/dching/miniconda3/envs/ptychopy/include -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /home/dching/miniconda3/envs/ptychopy/include -fPIC
compile options: '-I/home/dching/miniconda3/envs/ptychopy/lib/python3.7/site-packages/numpy/core/include -I['/home/dching/miniconda3/envs/ptychopy/lib/python3.7/site-packages/numpy/core/include'] -I/usr/local/cuda/include -I/usr/local/cuda/samples/common/inc -I/home/dching/miniconda3/envs/ptychopy/include -I/home/dching/miniconda3/envs/ptychopy/include/python3.7m -c'
extra options: 'nvcc gcc g++'
nvcc: ./src/ProbeKernels.cu
nvcc: ./src/DiffractionsKernels.cu
nvcc: ./src/ScanMeshKernels.cu
nvcc: ./src/CudaSmartPtr.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: ./src/IPtychoScanMesh.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: ./src/ThreadPool.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: ./src/Sample.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: ./src/Diffractions.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: ./src/utilities.cpp
nvcc fatal : Unknown option '-fpic'
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Mon_Oct_12_20:09:46_PDT_2020
Cuda compilation tools, release 11.1, V11.1.105
Build cuda_11.1.TC455_06.29190527_0
setuptools is version 49.6.0
Hello!
I'm having this problem with the installation process following your README file.
My device is python 3.6, CUDA 10.2, gcc and g++ are both version 4.8.5, and after running . /install.sh, I got an error halfway through the
installation:
error: Command "g++ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -I/home/pxy/miniconda3/envs/ptychopy/lib/ python3.6/site-packages/numpy/core/include -I['/home/pxy/miniconda3/envs/ptychopy/lib/python3.6/site-packages/numpy/core/include'] - I/usr/local/cuda-10.2/include -I/usr/local/cuda-10.2/samples/common/inc -I/home/pxy/miniconda3/envs/ptychopy/include -I./src -I/home/ pxy/miniconda3/envs/ptychopy/include/python3.6m -c . /src/FileManager.cpp -o build/temp.linux-x86_64-3.6/src/FileManager.o -DEPIE_HDF5 -fpic -DHAVE_HDF5" failed with exit status 1
I don't know what the problem is, so I'm asking for your help!Thanks!
After running this command to build your program: python setup.py build -v
I get the following error message. Seems that your program doesn't support cuda 11.3 ?
running build
running build_ext
building 'ptychopy' extension
Warning: Can't read registry to find the necessary compiler setting
Make sure that Python modules winreg, win32api or win32con are installed.
C compiler: g++ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC
compile options: '-I/home/xf9/miniconda3/envs/epie/lib/python3.6/site-packages/numpy/core/include -I['/home/xf9/miniconda3/envs/epie/lib/python3.6/site-packages/numpy/core/include'] -I/usr/local/cuda-11.3/include -I/usr/local/cuda-11.3/samples/common/inc -I/home/xf9/miniconda3/envs/epie/include -I./src -I/home/xf9/miniconda3/envs/epie/include/python3.6m -c'
extra options: 'nvcc gcc g++'
g++: ./src/CartesianScanMesh.cpp
g++: ./src/CudaSmartPtr.cpp
nvcc: ./src/utilitiesKernels.cu
nvcc: ./src/ScanMeshKernels.cu
nvcc: ./src/ProbeKernels.cu
g++: ./src/Cuda2DArray.cpp
nvcc: ./src/SampleKernels.cu
nvcc: ./src/DiffractionsKernels.cu
nvcc fatal : Unsupported gpu architecture 'compute_113'
nvcc fatal : Unsupported gpu architecture 'compute_113'
nvcc fatal : Unsupported gpu architecture 'compute_113'
nvcc fatal : Unsupported gpu architecture 'compute_113'
g++: ./src/IPtychoScanMesh.cpp
nvcc fatal : Unsupported gpu architecture 'compute_113'
g++: ./src/DiffractionLoader.cpp
g++: ./src/ThreadPool.cpp
g++: ./src/ListScanMesh.cpp
g++: ./src/Sample.cpp
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.