Comments on: Extending Python with C: A case study http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/ Random scribbling about programming, translation, and Japan Fri, 03 Feb 2012 09:05:20 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Ryan Ginstrom http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-115 Ryan Ginstrom Mon, 17 Dec 2007 11:03:09 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-115 @Chris: Thanks for the comment. Yes, as I noted in the article, there are a few more things I can do with the code to get a 10% to 30% speedup. Actually though, I don't know if you really have a function-call overhead for something like min. I don't have a lot of experience with gcc, but the MS compiler would almost certainly inline that for me. I guess you'd have to look at the generated assembly language to know for sure. @Chris:

Thanks for the comment. Yes, as I noted in the article, there are a few more things I can do with the code to get a 10% to 30% speedup.

Actually though, I don’t know if you really have a function-call overhead for something like min. I don’t have a lot of experience with gcc, but the MS compiler would almost certainly inline that for me. I guess you’d have to look at the generated assembly language to know for sure.

]]> By: Chris http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-114 Chris Mon, 17 Dec 2007 10:58:59 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-114 Your blog software might have eaten my comment, you can find MIN2, MIN3 at http://untergrund.bewaff.net/~chris/min.h Your blog software might have eaten my comment,
you can find MIN2, MIN3 at http://untergrund.bewaff.net/~chris/min.h

]]> By: Chris http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-113 Chris Mon, 17 Dec 2007 10:57:20 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-113 I just wanted to mention that implementing min as a function is not c-esque. Remember that every time you call the function, a stack frame needs to be build up. Therefore, people usually define it as a preprocessor macro: #define MIN2(a, b) ((a) I just wanted to mention that implementing min as a function is not c-esque. Remember that every time you call the function, a stack frame needs to be build up. Therefore, people usually define it as a preprocessor macro:

#define MIN2(a, b) ((a)

]]> By: ben http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-112 ben Fri, 14 Dec 2007 13:46:17 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-112 The benchmarks at the shootout site above don't mean anything. For example Perl is an interpreted language written in C, and yet Perl is supposed to be faster than C on some of those benchmarks. Since the Perl is a C program and the benchmark is a C program, all it means is that the C code used in the benchmark isn't as good as the C code in Perl. It's testing the programmer's skill, not the language itself. The only thing that we can conclude is that Larry Wall and co are better C programmers than the person who wrote the benchmark. Computer "science" is riddled with this kind of pseudo science. The benchmarks at the shootout site above don’t mean anything. For example Perl is an interpreted language written in C, and yet Perl is supposed to be faster than C on some of those benchmarks. Since the Perl is a C program and the benchmark is a C program, all it means is that the C code used in the benchmark isn’t as good as the C code in Perl. It’s testing the programmer’s skill, not the language itself. The only thing that we can conclude is that Larry Wall and co are better C programmers than the person who wrote the benchmark.

Computer “science” is riddled with this kind of pseudo science.

]]> By: Ryan Ginstrom http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-110 Ryan Ginstrom Thu, 06 Dec 2007 02:40:03 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-110 @alex Yes, I don't think anyone will argue that languages like C are much faster than VHLL like Python at these micro-benchmarks. I want to use Python wherever possible, and dip down to a lower-level language to break performance bottlenecks. My goal is to write something that's 99% Python, yet still 99% as fast as a pure C program :) Thanks for the idea about browsing the Python source to look for optimization ideas. The Python developers really are wizards about writing tight code. I like their motto: "We read Knuth so you don't have to." @alex

Yes, I don’t think anyone will argue that languages like C are much faster than VHLL like Python at these micro-benchmarks. I want to use Python wherever possible, and dip down to a lower-level language to break performance bottlenecks. My goal is to write something that’s 99% Python, yet still 99% as fast as a pure C program :)

Thanks for the idea about browsing the Python source to look for optimization ideas. The Python developers really are wizards about writing tight code. I like their motto: “We read Knuth so you don’t have to.”

]]> By: alex http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-109 alex Wed, 05 Dec 2007 02:08:24 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-109 You've probably already seen it, but the Programming Lang Shootout has some interesting any lang vs. any lang benchmarks... http://shootout.alioth.debian.org/ just browsing the code can be interesting too, to see how they optimize stuff in python etc. You’ve probably already seen it,
but the Programming Lang Shootout has some interesting
any lang vs. any lang benchmarks…
http://shootout.alioth.debian.org/

just browsing the code can be interesting too, to see
how they optimize stuff in python etc.

]]> By: fsdaily.com http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-111 fsdaily.com Mon, 03 Dec 2007 18:22:45 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-111 <strong>Story added...</strong> Your story has been submitted to fsdaily.com! Come and promote your article by voting for it here: http://www.fsdaily.com/HighEnd/Extending_Python_with_C_A_case_study... Story added…

Your story has been submitted to fsdaily.com! Come and promote your article by voting for it here: http://www.fsdaily.com/HighEnd/Extending_Python_with_C_A_case_study...

]]> By: Ryan Ginstrom http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-108 Ryan Ginstrom Mon, 03 Dec 2007 01:07:10 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-108 Thanks a lot for the link! I especially appreciate the very liberal license. I haven't seen any implementations of the substring-match specialization, which is why I wrote this one. Actually I thought I had invented it myself, so I was a little disappointed to see it already described on the wikipedia page for Levenshtein distance :) Thanks a lot for the link! I especially appreciate the very liberal license.

I haven’t seen any implementations of the substring-match specialization, which is why I wrote this one. Actually I thought I had invented it myself, so I was a little disappointed to see it already described on the wikipedia page for Levenshtein distance :)

]]> By: jose http://ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/comment-page-1/#comment-107 jose Mon, 03 Dec 2007 00:53:54 +0000 http://www.ginstrom.com/scribbles/2007/12/02/extending-python-with-c-a-case-study/#comment-107 nice work. libdistance implements these fuzzy string matching algorithms in C and has python bindings: http://monkey.org/~jose/software/libdistance/ levenshtein, hamming, jaccard, and a few others. i hope you find it useful! nice work. libdistance implements these fuzzy string matching algorithms in C and has python bindings:

http://monkey.org/~jose/software/libdistance/

levenshtein, hamming, jaccard, and a few others.

i hope you find it useful!

]]>