shinmorino / sqaod Goto Github PK

Solvers/annealers for simulated quantum annealing on CPU and CUDA(NVIDIA GPU).

Home Page: https://github.com/shinmorino/sqaod/wiki

License: Other

Python 26.72% Shell 0.14% C++ 53.48% Makefile 1.06% M4 0.46% Cuda 18.03% C 0.02% Batchfile 0.09%

cuda accelearated monte-carlo-simulation linux windows cplusplus-11 quantum-computing python nvidia-gpu quantum-annealing gpu

sqaod's Issues

Test python interface, including input type/size checks.

Adding aligned Eigen matrix.

Adding aligned Eigen matrix for SSE and AVX2 for matrices that are critical in performance.
Relating to #11.

update docs.

Build/intsall instruction.

configure args,
'--enable-cuda=<path>|<ver>'
'--with-blas'
CFLAGS to enable SIMD(AVX, SSE)

CUDA/GPU requirement.

Kepler(Cc3.5) or later.
Recommendation is Maxwell or later.
Developed and tested with CUDA 9.0/9.1. Recommendation is CUDA 9.0 or later.

Simple usage.

Update examples

Recheck temp storage usage for segmented sum module.

Refine initialization of solvers.

Temporal number of trotters should be given in constructor.
Calling order of randomize_q() and set_x() should be defined in relation with init_anneal().

Compling Python c-extensions on pip installation.

Parallelize CPU-based brute-force searchers.

Release alpha binary.

Current testsuite executions are required.

Unittest in c++.
Unittest in python.
run every python script.
run perf app to check solutions.
add instructions for instsallation/build on doc.

Default SIMD setting does not work.

Import Error in python when built from source for cuda

Update(9/26/2018)
Please visit Installation at sqaod wiki and follow installation instructions.

This issue was caused by version mismatch between 'self-built native libraries' and 'python package installed by pip'.
Discussion in this issue and resolutions have been already reflected to Installation.

(shinmorino)

Do I have to build also for CPU?

ubuntu@ip-172-31-6-165:~$ g++ --version
g++ (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

ubuntu@ip-172-31-6-165:~$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Wed_Apr_11_23:16:29_CDT_2018
Cuda compilation tools, release 9.2, V9.2.88

ubuntu@ip-172-31-6-165:~$ nvidia-smi
Wed Jun  6 09:24:09 2018       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 396.26                 Driver Version: 396.26                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla K80           On   | 00000000:00:1E.0 Off |                    0 |
| N/A   39C    P8    30W / 149W |      0MiB / 11441MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

Python 3.6.4 |Anaconda, Inc.| (default, Jan 16 2018, 18:10:19) 
[GCC 7.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import sqaod
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/ubuntu/anaconda3/lib/python3.6/site-packages/sqaod/__init__.py", line 10, in <module>
    from . import cpu
  File "/home/ubuntu/anaconda3/lib/python3.6/site-packages/sqaod/cpu/__init__.py", line 1, in <module>
    from . import formulas
  File "/home/ubuntu/anaconda3/lib/python3.6/site-packages/sqaod/cpu/formulas.py", line 4, in <module>
    from . import cpu_formulas as cext
ImportError: /home/ubuntu/anaconda3/lib/python3.6/site-packages/sqaod/cpu/cpu_formulas.so: undefined symbol: _ZN5sqaod12aligned_freeEPv

Give a local memory store to host memory allocator.

Tests for python solvers and python c-extensions

Calling all methods in python solvers to confirm if all solvers work as expected.

Use python built-in exceptions for c-extensions.

update setPreference method signature.

Solver:::setPreference() needs sq::Preference as the parameter.
Changing it to Solver::setPreference(enum name, V parameter)..

Setting tileSize to a multiple of 256.

set_x() with m x N matrix.

Renaming set_x() to set_q().
Implement 2 versions of the method for N-length vector and m x N-length vector.

Fix repository to make signing enabled.

Apply vectorization for CUDADenseGraphAnnealer

Update python interface.

Add MT19937 for CUDA solvers.

Installation on AWS DeepLearning AMI 9.0 Ubuntu 16.04

Hi!

I am stacking on

 libsqaodc-cuda-9-0 : 依存: cuda-cublas-9-0:amd64 しかし、インストールすることができません
                      依存: cuda-cudart-9-0:amd64 しかし、インストールすることができません
                      依存: cuda-curand-9-0:amd64 しかし、インストールすることができません
                      依存: cuda-cublas-9-0 しかし、インストールすることができません
                      依存: cuda-cudart-9-0 しかし、インストールすることができません
                      依存: cuda-curand-9-0 しかし、インストールすることができません

Also, the default python there is 3.6.4 anaconda, is it not supported?

Thanks

Validation on search range coverage for BF searchers.

Test cases for 2d kernels with (65535 < rows).

Relating to #11, tests for cases where (65535 < rows) are needed since the y size of max 2D grid is 65535.

Dense graph solvers are unable to accept upper/lower trianglar matrices.

All dense graph solvers are able to deal only with symmetric matrices.
symmetrize() has been added to enable upper/lower triangle matrices, there's a misunderstanding on the equation.

Fix:

Check if a given matrix is symmetric.
1-1. If symmetric, the matrix will be used as is.
If asymmetric, check if the matrix is upper/lower triangular.
2-1 If triangular, symmetrize the mat.
Error: Raise exception.

This fix will be applied to all dense graph solvers and functions.

Documentation to enable SIMD.

CPUBipartiteGraphBFSearcher raises exception on Windows when AVX(2) enabled.

This exception is raised in Eigen.
Windows is not the primary target platform, thus, the priority is low.
Since it may be an issue on memory alignment, this issue will be re-investigated after implementing #11.
Workaround is to disable AVX(2) in VC setting.

shinmorino / sqaod Goto Github PK

sqaod's Issues

Recommend Projects

Recommend Topics

Recommend Org