Cygwin 裡面已經有提供 OpenBLAS 與針對 OpenBLAS 編譯的 LAPACK 函式庫，可以用 CPU 處理多執行緒數值線性代數運算，以及針對這兩個東西編譯的 Python2/Python3 是值運算模組 Numpy。用這些就可以作數值運算了。
不過一般都會再裝 Scipy 和 Matplotlib 來用；要在 Cygwin 裡面用 Matplotlib 與 Scipy，就得用 setuptools 的 easy_install 或 pip 來裝；但是安裝過程有一些套件相依性要處理，也是有一點麻煩步驟的。
一些舊的資料提到可以使用 Tcl/Tk 作 Matplotlib 的互動顯示，不過在安裝的時候不斷遇到錯誤 ... 於是換成使用 Qt4/5，但是僅挑 PyQt5 的套件又不能用，要同時裝 PyQt4 才行
以下是目前整理出來，針對 Python3 的 Cygwin 套件。請直接用 Cygwin 安裝檔點選並進行安裝，這樣會一併連同相依套件一起挑出來；紅字部份是 Tcl/Tk 的部份，但是安裝後還是有問題，因此可以不用安裝。
編譯器、環境
|
gcc-fortran, gcc-g++, pkg-config
|
是值運算函式庫
|
必要：libopenblas, lapack-devel, libqhull-devel, libqrupdate-devel, libfftw3-devel, libcrypt-devel
選裝：libarpack-devel, libsuitesparseconfig-devel, libamd-devel, libcamd-devel, libccolamd-devel, libcholmod-devel, libcolamd-devel, libumfpack-devel
|
Python 環境
|
python3-devel, python3-pip, python3-numpy3, python3-pexpect, python3-six, python3-pyparsing, python3-tkinter, python3-pyqt4, python3-pyqt5
|
Matplotlib 相依的圖形相關函式庫
|
libfontconfig-devel, libpng-devel, libfreetype-devel, tcl-tk, tcl-itk
|
顯示圖案的 X-Window 系統
|
xorg-server, xinit, xlaunch
|
裝完之後，用下面指令使用 pip 安裝
user@localhost ~ $ pip3 -v install matplotlib user@localhost ~ $ pip3 -v install scipy
Matplotlib 編譯過程中跳出來的的相依性檢查
============================================================================
Edit setup.cfg to change the build options
BUILDING MATPLOTLIB
matplotlib: yes [2.0.2]
python: yes [3.6.1 (default, Mar 21 2017, 21:49:16) [GCC
5.4.0]]
platform: yes [cygwin]
REQUIRED DEPENDENCIES AND EXTENSIONS
numpy: yes [version 1.11.2]
six: yes [using six version 1.10.0]
dateutil: yes [using dateutil version 2.6.0]
functools32: yes [Not required]
subprocess32: yes [Not required]
pytz: yes [using pytz version 2017.2]
cycler: yes [using cycler version 0.10.0]
tornado: yes [tornado was not found. It is required for the
WebAgg backend. pip/easy_install may attempt to
install it after matplotlib.]
pyparsing: yes [using pyparsing version 2.1.10]
libagg: yes [pkg-config information for 'libagg' could not
be found. Using local copy.]
freetype: yes [version 2.6.5]
png: yes [version 1.6.28]
qhull: yes [pkg-config information for 'qhull' could not be
found. Using local copy.]
OPTIONAL SUBPACKAGES
sample_data: yes [installing]
toolkits: yes [installing]
tests: no [skipping due to configuration]
toolkits_tests: no [skipping due to configuration]
OPTIONAL BACKEND EXTENSIONS
macosx: no [Mac OS-X only]
qt5agg: yes [installing, Qt: 5.6.2, PyQt: 5.6.2]
qt4agg: yes [installing, Qt: 4.8.7, PyQt: 4.8.7; PySide not
found]
gtk3agg: no [Requires gtk3 development files to be
installed.]
gtk3cairo: no [Requires gtk3 development files to be
installed.]
gtkagg: no [Requires pygtk]
tkagg: yes [installing; run-time loading from Python Tcl /
Tk]
wxagg: no [requires wxPython]
gtk: no [Requires pygtk]
agg: yes [installing]
cairo: yes [installing, pycairo version 1.12.0]
windowing: no [Microsoft Windows only]
OPTIONAL LATEX DEPENDENCIES
dvipng: no
ghostscript: no
latex: no
pdftops: no
OPTIONAL PACKAGE DATA
dlls: no [skipping due to configuration]
============================================================================
Scipy 的部份就沒有截到有關資訊就不列了。
接著是各部份的安裝簡單檢查～
先來看 Numpy 安裝資訊。以下所顯示的資料提到，這個安裝套件「沒有」用 OpenBLAS？
user@localhost ~
$ python3
Python 3.6.1 (default, Mar 21 2017, 21:49:16)
[GCC 5.4.0] on cygwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> numpy.__config__.show()
blas_mkl_info:
NOT AVAILABLE
openblas_info:
NOT AVAILABLE
atlas_3_10_blas_threads_info:
NOT AVAILABLE
atlas_3_10_blas_info:
NOT AVAILABLE
atlas_blas_threads_info:
NOT AVAILABLE
atlas_blas_info:
NOT AVAILABLE
blas_info:
libraries = ['blas', 'blas']
library_dirs = ['/usr/lib']
language = c
define_macros = [('HAVE_CBLAS', None)]
blas_opt_info:
define_macros = [('NO_ATLAS_INFO', 1), ('HAVE_CBLAS', None)]
libraries = ['blas', 'blas']
library_dirs = ['/usr/lib']
language = c
openblas_lapack_info:
NOT AVAILABLE
lapack_mkl_info:
NOT AVAILABLE
atlas_3_10_threads_info:
NOT AVAILABLE
atlas_3_10_info:
NOT AVAILABLE
atlas_threads_info:
NOT AVAILABLE
atlas_info:
NOT AVAILABLE
lapack_info:
libraries = ['lapack', 'lapack']
library_dirs = ['/usr/lib']
language = f77
lapack_opt_info:
define_macros = [('NO_ATLAS_INFO', 1), ('HAVE_CBLAS', None)]
libraries = ['lapack', 'lapack', 'blas', 'blas']
library_dirs = ['/usr/lib']
language = c
>>> quit()
有點奇怪，跑一個矩陣內積運算看看，確認有沒有多執行緒進行
import numpy as np size=10000 a = np.random.random_sample((size,size)) b = np.random.random_sample((size,size)) np.dot(a,b)
看工作管理員，4 個 CPU Core 都有在跑，所以的確有使用 OpenBLAS（如果使用 Reference BLAS 的話，就只會看到一個 CPU Core 在忙而已～）
（本圖不純砍頭！！～）
接著檢查 Matplotlib 安裝。以下指令沒有跳出怎樣的 Error 就行了
user@localhost ~ $ python3 Python 3.6.1 (default, Mar 21 2017, 21:49:16) [GCC 5.4.0] on cygwin Type "help", "copyright", "credits" or "license" for more information. >>> import matplotlib >>> print(matplotlib.get_backend()) Qt5Agg >>> import matplotlib.pyplot as plt >>>
如果要即時繪圖，則需要「啟動」X Window 程序，用裡面的 Virtual Terminal 才行...
Cygwin 的 X Window 是一個背景程序，啟動後會在工作列有相關按鈕，從那裡啟用 XTerm 來跑下面從 Matplotlib 借來的範例，的確會跳出圖案
import matplotlib.pyplot as plt
import numpy as np
t = np.arange(0.0, 2.0, 0.01)
s = 1 + np.sin(2*np.pi*t)
plt.plot(t, s)
plt.xlabel('time (s)')
plt.ylabel('voltage (mV)')
plt.title('Sine Wave')
plt.grid(True)
plt.show()
本圖絕對不是網頁摳來的～～
至於 Scipy 的檢查就和 numpy 差不多，請各位找方法自行檢查
user@localhost ~ $ python3 Python 3.6.1 (default, Mar 21 2017, 21:49:16) [GCC 5.4.0] on cygwin Type "help", "copyright", "credits" or "license" for more information. >>> import scipy >>> scipy.__config__.show() openblas_lapack_info: NOT AVAILABLE lapack_mkl_info: NOT AVAILABLE atlas_3_10_threads_info: NOT AVAILABLE atlas_3_10_info: NOT AVAILABLE atlas_threads_info: NOT AVAILABLE atlas_info: NOT AVAILABLE lapack_info: libraries = ['lapack', 'lapack'] library_dirs = ['/usr/lib'] language = f77 blas_info: libraries = ['blas', 'blas'] library_dirs = ['/usr/lib'] language = c define_macros = [('HAVE_CBLAS', None)] lapack_opt_info: define_macros = [('NO_ATLAS_INFO', 1), ('HAVE_CBLAS', None)] libraries = ['lapack', 'lapack', 'blas', 'blas'] library_dirs = ['/usr/lib'] language = c blas_mkl_info: NOT AVAILABLE openblas_info: NOT AVAILABLE atlas_3_10_blas_threads_info: NOT AVAILABLE atlas_3_10_blas_info: NOT AVAILABLE atlas_blas_threads_info: NOT AVAILABLE atlas_blas_info: NOT AVAILABLE blas_opt_info: define_macros = [('NO_ATLAS_INFO', 1), ('HAVE_CBLAS', None)] libraries = ['blas', 'blas'] library_dirs = ['/usr/lib'] language = c >>> quit()
上面這樣就安裝完成，可以使用了。如果只是用原生的 Python 發行版，就能從下面網站抓套件 Python Extension Packages for Windows - Christoph Gohlke，裡面的 BLAS 用的是 Intel MKL。
更流行的方式，就是用 Anaconda 或是 Python(X,Y) 等 Python 發行版（這裡有列清單：Installing the SciPy Stack — SciPy.org），安裝就更簡單了。
雖然數值線性代數運算模組除了 OpenBLAS 之外還有 Intel Math Kernel Library，甚至最近好像很流行的 GPU 運算套件（OpenCL 還是 CUDA 還是？？等等），不過對於一般個人使用，OpenBLAS 的速度也是很不錯的。只要不是 Reference BLAS（Netlib BLAS）或是 ATLAS 就都很好處理～～
參考資料：
- 安裝部份
- 過程中可能會遇到的問題
gcc: error：spawn: No such file or directory ：發生原因是 g++ 沒有裝到
其他可能問題
- 確認功能
沒有留言:
張貼留言