{"id":46415,"date":"2023-04-26T16:48:34","date_gmt":"2023-10-03T07:03:22","guid":{"rendered":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/"},"modified":"2024-04-30T01:24:49","modified_gmt":"2024-04-29T17:24:49","slug":"46415-2","status":"publish","type":"post","link":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/","title":{"rendered":""},"content":{"rendered":"

\u601d\u3063\u305f\u3088\u308a\u7c21\u5358\u306bJupyter Notebook(iPython Notebook)\u3067Cython\u3092\u8a66\u305b\u308b\u3053\u3068\u304c\u5206\u304b\u3063\u305f\u306e\u3067\u30e1\u30e2\u3067\u3059\u3002Cython\u306f\u5b9f\u884c\u524d\u306b\u30b3\u30f3\u30d1\u30a4\u30eb\u3059\u308b\u3053\u3068\u3001\u9759\u7684\u306a\u578b\u4ed8\u3051\u3092\u884c\u3046\u3053\u3068\u3067\u51e6\u7406\u3092\u9ad8\u901f\u5316\u3057\u307e\u3059\u3002<\/p>\n

(\u672c\u8a18\u4e8b\u306e.ipynb\u30d5\u30a1\u30a4\u30eb\u306fGithub\u306e\u3053\u3053\u306b\u30a2\u30c3\u30d7\u30ed\u30fc\u30c9\u3057\u3066\u3042\u308a\u307e\u3059\u3002)<\/p>\n

\u74b0\u5883<\/h2>\n

\u8a66\u3057\u305f\u74b0\u5883\u306f\u4e0b\u8a18\u3067\u3059\u3002Mac\u3068Anaconda\u3067\u8a66\u3057\u3066\u3044\u307e\u3059\u3002Anaconda\u3092\u5c0e\u5165\u3057\u3066\u3044\u308c\u3070\u7279\u306b\u6e96\u5099\u306f\u4e0d\u8981\u3067\u3059\u3002<\/p>\n

Python 3.5.1 |Anaconda custom (x86_64)| (default, Jun 15 2016, 16:14:02) \r\n[GCC 4.2.1 Compatible Apple LLVM 4.2 (clang-425.0.28)] on darwin \r\n\r\nIPython 5.0.0 \r\n<\/code><\/pre>\n
\u3084\u3063\u3066\u307f\u3088\u3046<\/h2>\n
Cython\u30b3\u30f3\u30d1\u30a4\u30eb\u7528\u306e\u30de\u30b8\u30c3\u30af\u30b3\u30de\u30f3\u30c9\u5b9f\u884c<\/h3>\n
# Jupyter Notebook\u4e0a\u3067cython\u30d5\u30a1\u30a4\u30eb\u306e\u30b3\u30f3\u30d1\u30a4\u30eb\u304c\u3067\u304d\u308b\u3088\u3046\u306b\u3059\u308b\r\n<\/span>%<\/span>load_ext<\/span> Cython<\/span>\r\n<\/code><\/pre>\n
Cython\u306e\u95a2\u6570\u3092\u5ba3\u8a00<\/h3>\n
\u5148\u982d\u306b%%cython\u3092\u8a18\u8f09\u3057\u3066cython\u30b3\u30fc\u30c9\u3092\u66f8\u304d\u307e\u3059\u3002
\n\u4f8b\u306fCython \u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u57fa\u790e\u7de8\u304b\u3089\u4f7f\u308f\u305b\u3066\u3082\u3089\u3044\u307e\u3057\u305f\u3002<\/p>\n
# \u2193 -n <\u30d5\u30a1\u30a4\u30eb\u540d>\u3092\u3064\u3051\u308b\u3053\u3068\u3067\u5f8c\u3067\u30d5\u30a1\u30a4\u30eb\u304c\u78ba\u8a8d\u3057\u3084\u3059\u304f\u306a\u308a\u307e\u3059\u3002\r\n<\/span>%%<\/span>cython<\/span> -<\/span>n<\/span> test_cython_code<\/span>\r\ndef<\/span> fib<\/span>(<\/span>int<\/span> n<\/span>):<\/span>\r\n    cdef<\/span> int<\/span> i<\/span>\r\n    cdef<\/span> double<\/span> a<\/span>=<\/span>0.0<\/span>,<\/span> b<\/span>=<\/span>1.0<\/span>\r\n\r\n    for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>):<\/span>\r\n        a<\/span>,<\/span> b<\/span> =<\/span> a<\/span>+<\/span>b<\/span>,<\/span> a<\/span>\r\n    return<\/span> a<\/span>\r\n\r\ndef<\/span> primes<\/span>(<\/span>int<\/span> kmax<\/span>):<\/span>\r\n    cdef<\/span> int<\/span> n<\/span>,<\/span> k<\/span>,<\/span> i<\/span>\r\n    cdef<\/span> int<\/span> p<\/span>[<\/span>1000<\/span>]<\/span>\r\n    result<\/span> =<\/span> []<\/span>\r\n\r\n    if<\/span> kmax<\/span> ><\/span> 1000<\/span>:<\/span>\r\n        kmax<\/span> =<\/span> 1000<\/span>\r\n\r\n    k<\/span> =<\/span> 0<\/span>\r\n    n<\/span> =<\/span> 2<\/span>\r\n    while<\/span> k<\/span> <<\/span> kmax<\/span>:<\/span>\r\n        i<\/span> =<\/span> 0<\/span>\r\n        while<\/span> i<\/span> <<\/span> k<\/span> and<\/span> n<\/span> %<\/span> p<\/span>[<\/span>i<\/span>]<\/span> !=<\/span> 0<\/span>:<\/span>\r\n            i<\/span> +=<\/span> 1<\/span>\r\n\r\n        if<\/span> i<\/span> ==<\/span> k<\/span>:<\/span>\r\n            p<\/span>[<\/span>k<\/span>]<\/span> =<\/span> n<\/span>\r\n            k<\/span> +=<\/span> 1<\/span>\r\n            result<\/span>.<\/span>append<\/span>(<\/span>n<\/span>)<\/span>\r\n        n<\/span> +=<\/span> 1<\/span>\r\n    return<\/span> result<\/span>\r\n<\/code><\/pre>\n
\u8a66\u3057\u3066\u307f\u308b<\/h3>\n
print(fib(90))\r\nprint(primes(20))\r\n<\/code><\/pre>\n
2.880067194370816e+18\r\n[2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71]\r\n<\/code><\/pre>\n
\u3067\u304d\u305f\uff01<\/p>\n
\u751fPython\u3068\u306e\u901f\u5ea6\u6bd4\u8f03<\/h2>\n
\u540c\u3058\u51e6\u7406\u3092\u751f\u306ePython\u3067\u66f8\u3044\u3066\u5b9f\u884c\u6642\u9593\u3092\u6bd4\u8f03\u3057\u3066\u307f\u307e\u3059\u3002<\/p>\n
import<\/span> numpy<\/span> as<\/span> np<\/span>\r\n# \u6027\u80fd\u6bd4\u8f03\u7528\u3000Python\u95a2\u6570\r\n<\/span>def<\/span> pyfib<\/span>(<\/span>n<\/span>):<\/span>\r\n    a<\/span>,<\/span> b<\/span> =<\/span> 0.0<\/span>,<\/span> 1.0<\/span>\r\n    for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>):<\/span>\r\n        a<\/span>,<\/span> b<\/span> =<\/span> a<\/span>+<\/span>b<\/span>,<\/span> a<\/span>\r\n    return<\/span> a<\/span>\r\n\r\ndef<\/span> pyprimes<\/span>(<\/span>kmax<\/span>):<\/span>\r\n    p<\/span> =<\/span> np<\/span>.<\/span>zeros<\/span>(<\/span>1000<\/span>)<\/span>\r\n    result<\/span> =<\/span> []<\/span>\r\n\r\n    # \u6700\u5927\u500b\u6570\u306f1000\u500b\r\n<\/span>    if<\/span> kmax<\/span> ><\/span> 1000<\/span>:<\/span>\r\n        kmax<\/span> =<\/span> 1000<\/span>\r\n\r\n    k<\/span> =<\/span> 0<\/span>\r\n    n<\/span> =<\/span> 2<\/span>\r\n    while<\/span> k<\/span> <<\/span> kmax<\/span>:<\/span>\r\n        i<\/span> =<\/span> 0<\/span>\r\n        while<\/span> i<\/span> <<\/span> k<\/span> and<\/span> n<\/span> %<\/span> p<\/span>[<\/span>i<\/span>]<\/span> !=<\/span> 0<\/span>:<\/span>\r\n            i<\/span> +=<\/span> 1<\/span>\r\n\r\n        if<\/span> i<\/span> ==<\/span> k<\/span>:<\/span>\r\n            p<\/span>[<\/span>k<\/span>]<\/span> =<\/span> n<\/span>\r\n            k<\/span> +=<\/span> 1<\/span>\r\n            result<\/span>.<\/span>append<\/span>(<\/span>n<\/span>)<\/span>\r\n        n<\/span> +=<\/span> 1<\/span>\r\n    return<\/span> result<\/span>\r\n<\/code><\/pre>\n
\u30d5\u30a3\u30dc\u30ca\u30c3\u30c1\u6570<\/h3>\n
# 1000\u756a\u76ee\u306e\u30d5\u30a3\u30dc\u30ca\u30c3\u30c1\u6570\u3092\u7e70\u308a\u8fd4\u3057\u751f\u6210\u3057\u3066\u8a08\u6e2c\r\n<\/span>%<\/span>timeit<\/span> fib<\/span>(<\/span>1000<\/span>)<\/span>\r\n%<\/span>timeit<\/span> pyfib<\/span>(<\/span>1000<\/span>)<\/span>\r\n<\/code><\/pre>\n
cython\u306e\u65b9\u304c50\u500d\u304f\u3089\u3044\u65e9\u3044\u3067\u3059\uff01<\/p>\n
1000000 loops, best of 3: 786 ns per loop\r\n10000 loops, best of 3: 42.5 \u00b5s per loop\r\n<\/code><\/pre>\n
\u7d20\u6570\u3092\u53d6\u308a\u51fa\u3059<\/h3>\n
%<\/span>timeit<\/span> primes<\/span>(<\/span>1000<\/span>)<\/span>\r\n%<\/span>timeit<\/span> pyprimes<\/span>(<\/span>1000<\/span>)<\/span>\r\n<\/code><\/pre>\n
100 loops, best of 3: 2.12 ms per loop\r\n1 loop, best of 3: 218 ms per loop\r\n<\/code><\/pre>\n
\u3053\u3061\u3089\u306e\u8a08\u7b97\u3067\u306f100\u500d\u304f\u3089\u3044\u65e9\u3044\u3067\u3059\uff01<\/p>\n
Pandas\u306eapply\u306b\u4f7f\u3063\u3066\u307f\u308b<\/h2>\n
1000\u500b\u306e\u6574\u6570<\/p>\n
df<\/span> =<\/span> pd<\/span>.<\/span>DataFrame<\/span>(<\/span>np<\/span>.<\/span>arange<\/span>(<\/span>1<\/span>,<\/span> 10<\/span>**<\/span>4<\/span>),<\/span> columns<\/span>=<\/span>[<\/span>'num'<\/span>]<\/span> )<\/span>\r\n<\/code><\/pre>\n
apply\u95a2\u6570\u306e\u4e2d\u306b\u95a2\u6570\u6307\u5b9a\u3059\u308b\u3060\u3051\u3067\u4f7f\u3048\u307e\u3059<\/p>\n
%<\/span>timeit<\/span> df<\/span>[<\/span>'fib'<\/span>]<\/span> =<\/span> df<\/span>.<\/span>num<\/span>.<\/span>apply<\/span>(<\/span>fib<\/span>)<\/span>\r\n%<\/span>timeit<\/span> df<\/span>[<\/span>'pyfib'<\/span>]<\/span> =<\/span> df<\/span>.<\/span>num<\/span>.<\/span>apply<\/span>(<\/span>pyfib<\/span>)<\/span>\r\n<\/code><\/pre>\n
10 loops, best of 3: 39.2 ms per loop\r\n1 loop, best of 3: 2.02 s per loop\r\n<\/code><\/pre>\n
print<\/span>(<\/span>df<\/span>.<\/span>head<\/span>())<\/span>\r\n<\/code><\/pre>\n
   num  fib  pyfib\r\n0    1  1.0    1.0\r\n1    2  1.0    1.0\r\n2    3  2.0    2.0\r\n3    4  3.0    3.0\r\n4    5  5.0    5.0\r\n<\/code><\/pre>\n
\u30b3\u30f3\u30d1\u30a4\u30eb\u3055\u308c\u305fcython\u30d5\u30a1\u30a4\u30eb\u306f~\/.ipython\/cython\u306b\u683c\u7d0d\u3055\u308c\u3066\u3044\u307e\u3059\u3002
\n\u30b3\u30f3\u30d1\u30a4\u30eb\u6642\u306b%%cython -n <\u30d5\u30a1\u30a4\u30eb\u540d>\u3067\u30d5\u30a1\u30a4\u30eb\u540d\u3092\u6307\u5b9a\u3057\u3066\u3042\u3063\u305f\u5834\u5408\u306f\u305d\u306e\u30d5\u30a1\u30a4\u30eb\u540d\u3067\u3053\u3053\u306b\u683c\u7d0d\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n
ndarray\u3092\u6271\u3046<\/h2>\n
# \u30c7\u30fc\u30bf\u3092\u4f5c\u308b\r\n<\/span>rd<\/span>.<\/span>seed<\/span>(<\/span>71<\/span>)<\/span>\r\nn_data<\/span> =<\/span> 10<\/span>**<\/span>5<\/span>\r\nX<\/span> =<\/span> pd<\/span>.<\/span>DataFrame<\/span>(<\/span>rd<\/span>.<\/span>normal<\/span>(<\/span>size<\/span>=<\/span>3<\/span>*<\/span>n_data<\/span>).<\/span>reshape<\/span>((<\/span>n_data<\/span>,<\/span>3<\/span>)),<\/span> columns<\/span>=<\/span>[<\/span>\"a\"<\/span>,<\/span> \"b\"<\/span>,<\/span> \"c\"<\/span>])<\/span>\r\nprint<\/span>(<\/span>X<\/span>.<\/span>shape<\/span>)<\/span>\r\nprint<\/span>(<\/span>X<\/span>.<\/span>head<\/span>())<\/span>\r\n<\/code><\/pre>\n
(100000, 3)\r\n          a         b         c\r\n0 -0.430603 -1.193928 -0.444299\r\n1  0.489412 -0.451557  0.585696\r\n2  1.177320 -0.965009  0.218278\r\n3 -0.866144 -0.323006  1.412919\r\n4 -0.712651 -1.362191 -1.705966\r\n<\/code><\/pre>\n
ndarray\u3092\u5f15\u6570\u306b\u53d6\u308bCython\u30b3\u30fc\u30c9\u3092\u66f8\u304f<\/p>\n
%%<\/span>cython<\/span> -<\/span>n<\/span> sample_calc<\/span> \r\nimport<\/span> numpy<\/span> as<\/span> np<\/span>\r\ncimport<\/span> numpy<\/span> as<\/span> np<\/span>\r\n\r\ncpdef<\/span> np<\/span>.<\/span>ndarray<\/span>[<\/span>double<\/span>]<\/span> sample_calc<\/span>(<\/span>np<\/span>.<\/span>ndarray<\/span> col_a<\/span>,<\/span> np<\/span>.<\/span>ndarray<\/span> col_b<\/span>,<\/span> np<\/span>.<\/span>ndarray<\/span> col_c<\/span>):<\/span>\r\n    # \u5404\u5217\u306e\u578b\u30c1\u30a7\u30c3\u30af\r\n<\/span>    assert<\/span> (<\/span>col_a<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span> and<\/span> col_b<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span> and<\/span> col_c<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span>)<\/span>\r\n\r\n    # \u5404\u5217\u306e\u30b5\u30a4\u30ba\u304c\u540c\u3058\u3067\u3042\u308b\u3053\u3068\u3092\u30c1\u30a7\u30c3\u30af\r\n<\/span>    cdef<\/span> Py_ssize_t<\/span> n<\/span> =<\/span> len<\/span>(<\/span>col_c<\/span>)<\/span>\r\n    assert<\/span> (<\/span>len<\/span>(<\/span>col_a<\/span>)<\/span> ==<\/span> len<\/span>(<\/span>col_b<\/span>)<\/span> ==<\/span> n<\/span>)<\/span>\r\n    cdef<\/span> np<\/span>.<\/span>ndarray<\/span>[<\/span>double<\/span>]<\/span> res<\/span> =<\/span> np<\/span>.<\/span>empty<\/span>(<\/span>n<\/span>)<\/span>\r\n\r\n    # (a-b)\/c \u3068\u3044\u3046\u8a08\u7b97\u3092\u3059\u308b\r\n<\/span>    for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>):<\/span>\r\n        res<\/span>[<\/span>i<\/span>]<\/span> =<\/span> (<\/span>col_a<\/span>[<\/span>i<\/span>]<\/span> -<\/span> col_b<\/span>[<\/span>i<\/span>])<\/span>\/<\/span>col_c<\/span>[<\/span>i<\/span>]<\/span>\r\n    return<\/span> res<\/span>\r\n<\/code><\/pre>\n
Python\u5074\u304b\u3089\u547c\u3073\u51fa\u3059<\/p>\n
sample_calc<\/span>(<\/span>X<\/span>.<\/span>a<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>b<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>c<\/span>.<\/span>values<\/span>)<\/span>\r\n<\/code><\/pre>\n
array([-1.71804336,  1.60658332,  9.81468496, ..., -0.44683095,\r\n        0.46970409, -0.28352272])\r\n<\/code><\/pre>\n
# \u6bd4\u8f03\u7528\r\n<\/span>def<\/span> pysample_calc<\/span>(<\/span>col_a<\/span>,<\/span> col_b<\/span>,<\/span> col_c<\/span>):<\/span>\r\n    # \u5404\u5217\u306e\u578b\u30c1\u30a7\u30c3\u30af\r\n<\/span>    assert<\/span> (<\/span>col_a<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span> and<\/span> col_b<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span> and<\/span> col_c<\/span>.<\/span>dtype<\/span> ==<\/span> np<\/span>.<\/span>float<\/span>)<\/span>\r\n\r\n    # \u5404\u5217\u306e\u30b5\u30a4\u30ba\u304c\u540c\u3058\u3067\u3042\u308b\u3053\u3068\u3092\u30c1\u30a7\u30c3\u30af\r\n<\/span>    n<\/span> =<\/span> len<\/span>(<\/span>col_c<\/span>)<\/span>\r\n    assert<\/span> (<\/span>len<\/span>(<\/span>col_a<\/span>)<\/span> ==<\/span> len<\/span>(<\/span>col_b<\/span>)<\/span> ==<\/span> n<\/span>)<\/span>\r\n    res<\/span> =<\/span> np<\/span>.<\/span>empty<\/span>(<\/span>n<\/span>)<\/span>\r\n\r\n    # (a-b)\/c \u3068\u3044\u3046\u8a08\u7b97\u3092\u3059\u308b\r\n<\/span>    for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>):<\/span>\r\n        res<\/span>[<\/span>i<\/span>]<\/span> =<\/span> (<\/span>col_a<\/span>[<\/span>i<\/span>]<\/span> -<\/span> col_b<\/span>[<\/span>i<\/span>])<\/span>\/<\/span>col_c<\/span>[<\/span>i<\/span>]<\/span>\r\n    return<\/span> res<\/span>\r\n<\/code><\/pre>\n
%<\/span>timeit<\/span> sample_calc<\/span>(<\/span>X<\/span>.<\/span>a<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>b<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>c<\/span>.<\/span>values<\/span>)<\/span>\r\n%<\/span>timeit<\/span> pysample_calc<\/span>(<\/span>X<\/span>.<\/span>a<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>b<\/span>.<\/span>values<\/span>,<\/span> X<\/span>.<\/span>c<\/span>.<\/span>values<\/span>)<\/span>\r\n<\/code><\/pre>\n
100 loops, best of 3: 16.7 ms per loop\r\n10 loops, best of 3: 37.2 ms per loop\r\n<\/code><\/pre>\n
\u30e2\u30f3\u30c6\u30ab\u30eb\u30ed\u6cd5\u3067\u5186\u5468\u7387\u3092\u8a08\u7b97<\/h2>\n
# \u30c7\u30fc\u30bf\u751f\u6210\r\n<\/span>rd<\/span>.<\/span>seed<\/span>(<\/span>71<\/span>)<\/span>\r\nn_data<\/span> =<\/span> 10<\/span>**<\/span>7<\/span>\r\nX2<\/span> =<\/span> rd<\/span>.<\/span>random<\/span>(<\/span>size<\/span>=<\/span>(<\/span>n_data<\/span>,<\/span>2<\/span>)).<\/span>astype<\/span>(<\/span>np<\/span>.<\/span>float<\/span>)<\/span>\r\nX2<\/span>.<\/span>dtype<\/span>\r\n<\/code><\/pre>\n
Cython\u95a2\u6570\u306e\u5b9a\u7fa9<\/p>\n
%%<\/span>cython<\/span> -<\/span>n<\/span> calc_pi<\/span>\r\nimport<\/span> numpy<\/span> as<\/span> np<\/span>\r\ncimport<\/span> numpy<\/span> as<\/span> np<\/span>\r\n\r\ncpdef<\/span> np<\/span>.<\/span>ndarray<\/span>[<\/span>long<\/span>]<\/span>  calc_pi<\/span>(<\/span>np<\/span>.<\/span>ndarray<\/span>[<\/span>double<\/span>,<\/span> ndim<\/span>=<\/span>2<\/span>]<\/span> data<\/span>):<\/span>\r\n    cdef<\/span> Py_ssize_t<\/span> n<\/span> =<\/span> len<\/span>(<\/span>data<\/span>)<\/span>\r\n    cdef<\/span> np<\/span>.<\/span>ndarray<\/span>[<\/span>long<\/span>]<\/span> res<\/span> =<\/span> np<\/span>.<\/span>empty<\/span>(<\/span>n<\/span>,<\/span> dtype<\/span>=<\/span>np<\/span>.<\/span>int<\/span>)<\/span>\r\n\r\n    for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>):<\/span>\r\n        res<\/span>[<\/span>i<\/span>]<\/span> =<\/span> 1<\/span> if<\/span> (<\/span>data<\/span>[<\/span>i<\/span>,<\/span>0<\/span>]<\/span>**<\/span>2<\/span> +<\/span> data<\/span>[<\/span>i<\/span>,<\/span>1<\/span>]<\/span>**<\/span>2<\/span>)<\/span> <<\/span> 1<\/span> else<\/span> 0<\/span>\r\n    return<\/span> res<\/span>\r\n<\/code><\/pre>\n
\u6bd4\u8f03\u7528Python\u95a2\u6570<\/p>\n
# \u6bd4\u8f03\u7528Python\u95a2\u6570\r\n<\/span>def<\/span> pycalc_pi<\/span>(<\/span>data<\/span>):<\/span>\r\n    n<\/span> =<\/span> len<\/span>(<\/span>data<\/span>)<\/span>\r\n    res<\/span> =<\/span> [<\/span>1<\/span> if<\/span> (<\/span>data<\/span>[<\/span>i<\/span>,<\/span>0<\/span>]<\/span>**<\/span>2<\/span> +<\/span> data<\/span>[<\/span>i<\/span>,<\/span>1<\/span>]<\/span>**<\/span>2<\/span>)<\/span> <<\/span> 1<\/span> else<\/span> 0<\/span> for<\/span> i<\/span> in<\/span> range<\/span>(<\/span>n<\/span>)]<\/span>\r\n    return<\/span> res<\/span>\r\n<\/code><\/pre>\n
\u8a08\u6e2c\u3057\u3066\u307f\u308b\u3002<\/p>\n
%<\/span>time<\/span> calc_pi<\/span>(<\/span>X2<\/span>)<\/span>\r\n%<\/span>time<\/span> pycalc_pi<\/span>(<\/span>X2<\/span>)<\/span>\r\n<\/code><\/pre>\n
CPU times: user 25.2 ms, sys: 5.98 ms, total: 31.2 ms\r\nWall time: 31.1 ms\r\nCPU times: user 7.7 s, sys: 46.1 ms, total: 7.75 s\r\nWall time: 7.75 s\r\n<\/code><\/pre>\n
Cython\u306e\u65b9\u304c\u3060\u3044\u3076\u65e9\u3044\uff01<\/p>\n
# \u7d50\u679c\u304c\u540c\u4e00\u304b\u78ba\u8a8d\r\n<\/span>np<\/span>.<\/span>all<\/span>(<\/span>res<\/span> ==<\/span> respy<\/span>)<\/span>\r\n<\/code><\/pre>\n
\u3042\u3063\u3066\u308b\uff01<\/p>\n
True\r\n<\/code><\/pre>\n
# \u5186\u5468\u7387\u306e\u8a08\u7b97\r\n<\/span>np<\/span>.<\/span>sum<\/span>(<\/span>res<\/span>)<\/span>\/<\/span>n_data<\/span>*<\/span>4<\/span>\r\n<\/code><\/pre>\n
3.1413555999999998\r\n<\/code><\/pre>\n
\u63cf\u753b\u3057\u3066\u307f\u308b\u3002<\/p>\n
import<\/span> matplotlib.pyplot<\/span> as<\/span> plt<\/span>\r\n%<\/span>matplotlib<\/span> inline<\/span>\r\nimport<\/span> seaborn<\/span> as<\/span> sns<\/span>\r\nfrom<\/span> matplotlib.colors<\/span> import<\/span> LinearSegmentedColormap<\/span>\r\n\r\nsns<\/span>.<\/span>set<\/span>(<\/span>style<\/span>=<\/span>\"darkgrid\"<\/span>,<\/span> palette<\/span>=<\/span>\"muted\"<\/span>,<\/span> color_codes<\/span>=<\/span>True<\/span>)<\/span>\r\n# \u63cf\u753b\u3059\u308b\r\n<\/span>n_plot<\/span> =<\/span> 10<\/span>**<\/span>4<\/span>  # \u63cf\u753b\u3059\u308b\u70b9\u306e\u6570\r\n<\/span>plt<\/span>.<\/span>figure<\/span>(<\/span>figsize<\/span>=<\/span>(<\/span>8<\/span>,<\/span>8<\/span>))<\/span>\r\nplt<\/span>.<\/span>scatter<\/span>(<\/span>X2<\/span>[:<\/span>n_plot<\/span>,<\/span>0<\/span>],<\/span> X2<\/span>[:<\/span>n_plot<\/span>,<\/span>1<\/span>],<\/span> c<\/span>=<\/span>res<\/span>[:<\/span>n_plot<\/span>],<\/span> s<\/span>=<\/span>10<\/span>)<\/span>\r\n<\/code><\/pre>\n
\"plot.png\"<\/div>\n
\u3061\u3083\u3093\u3068\u5186\u306e\u4e2d\u3068\u5916\u3092\u5224\u5b9a\u3057\u3066\u3044\u307e\u3059\u306d\u3002<\/p>\n
\u53c2\u8003<\/h3>\n
Cython \u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u57fa\u790e\u7de8
\nhttp:\/\/omake.accense.com\/static\/doc-ja\/cython\/src\/userguide\/tutorial.html<\/p>\n
\u30aa\u30e9\u30a4\u30ea\u30fc “Cython”
\nhttps:\/\/www.oreilly.co.jp\/books\/9784873117270\/<\/p>\n
pandas 0.18.1 documentation Enhancing Performance
\nhttp:\/\/pandas.pydata.org\/pandas-docs\/stable\/enhancingperf.html<\/p>\n","protected":false},"excerpt":{"rendered":"
\u601d\u3063\u305f\u3088\u308a\u7c21\u5358\u306bJupyter Notebook(iPython Notebook)\u3067Cython\u3092\u8a66\u305b\u308b\u3053\u3068 […]<\/p>\n","protected":false},"author":10,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-46415","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"yoast_head":"\n- Blog - Silicon Cloud<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:description\" content=\"\u601d\u3063\u305f\u3088\u308a\u7c21\u5358\u306bJupyter Notebook(iPython Notebook)\u3067Cython\u3092\u8a66\u305b\u308b\u3053\u3068 […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/\" \/>\n<meta property=\"og:site_name\" content=\"Blog - Silicon Cloud\" \/>\n<meta property=\"article:published_time\" content=\"2023-10-03T07:03:22+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-04-29T17:24:49+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/cdn.silicloud.com\/blog-img\/blog\/img\/657d645137434c4406d04179\/63-0.png\" \/>\n<meta name=\"author\" content=\"\u5b87, \u534e\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"\u5b87, \u534e\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/\",\"url\":\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/\",\"name\":\"- Blog - Silicon Cloud\",\"isPartOf\":{\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/#website\"},\"datePublished\":\"2023-10-03T07:03:22+00:00\",\"dateModified\":\"2024-04-29T17:24:49+00:00\",\"author\":{\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/513018e4e121d3add1b7c5de8be21458\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/\"]}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/#website\",\"url\":\"https:\/\/www.silicloud.com\/zh\/blog\/\",\"name\":\"Blog - Silicon Cloud\",\"description\":\"\",\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/513018e4e121d3add1b7c5de8be21458\",\"name\":\"\u5b87, \u534e\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/63cd45cbc05a35fc4ff7637a163c83c4962ef58d27472726c3a3e0c9c5194f0f?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/63cd45cbc05a35fc4ff7637a163c83c4962ef58d27472726c3a3e0c9c5194f0f?s=96&d=mm&r=g\",\"caption\":\"\u5b87, \u534e\"},\"url\":\"https:\/\/www.silicloud.com\/zh\/blog\/author\/yuhua\/\"},{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/#local-main-organization-logo\",\"url\":\"\",\"contentUrl\":\"\",\"caption\":\"Blog - Silicon Cloud\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"- Blog - Silicon Cloud","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/","og_locale":"zh_CN","og_type":"article","og_description":"\u601d\u3063\u305f\u3088\u308a\u7c21\u5358\u306bJupyter Notebook(iPython Notebook)\u3067Cython\u3092\u8a66\u305b\u308b\u3053\u3068 […]","og_url":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/","og_site_name":"Blog - Silicon Cloud","article_published_time":"2023-10-03T07:03:22+00:00","article_modified_time":"2024-04-29T17:24:49+00:00","og_image":[{"url":"https:\/\/cdn.silicloud.com\/blog-img\/blog\/img\/657d645137434c4406d04179\/63-0.png"}],"author":"\u5b87, \u534e","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"\u5b87, \u534e","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"4 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/","url":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/","name":"- Blog - Silicon Cloud","isPartOf":{"@id":"https:\/\/www.silicloud.com\/zh\/blog\/#website"},"datePublished":"2023-10-03T07:03:22+00:00","dateModified":"2024-04-29T17:24:49+00:00","author":{"@id":"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/513018e4e121d3add1b7c5de8be21458"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/"]}]},{"@type":"WebSite","@id":"https:\/\/www.silicloud.com\/zh\/blog\/#website","url":"https:\/\/www.silicloud.com\/zh\/blog\/","name":"Blog - Silicon Cloud","description":"","inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/513018e4e121d3add1b7c5de8be21458","name":"\u5b87, \u534e","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.silicloud.com\/zh\/blog\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/63cd45cbc05a35fc4ff7637a163c83c4962ef58d27472726c3a3e0c9c5194f0f?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/63cd45cbc05a35fc4ff7637a163c83c4962ef58d27472726c3a3e0c9c5194f0f?s=96&d=mm&r=g","caption":"\u5b87, \u534e"},"url":"https:\/\/www.silicloud.com\/zh\/blog\/author\/yuhua\/"},{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.silicloud.com\/zh\/blog\/46415-2\/#local-main-organization-logo","url":"","contentUrl":"","caption":"Blog - Silicon Cloud"}]}},"_links":{"self":[{"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/posts\/46415","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/users\/10"}],"replies":[{"embeddable":true,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/comments?post=46415"}],"version-history":[{"count":2,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/posts\/46415\/revisions"}],"predecessor-version":[{"id":88836,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/posts\/46415\/revisions\/88836"}],"wp:attachment":[{"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/media?parent=46415"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/categories?post=46415"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.silicloud.com\/zh\/blog\/wp-json\/wp\/v2\/tags?post=46415"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}