News for hackers. Stuff that matters.
Re:PyPy2.7 and PyPy3.5 v5.7 released by leblancfg on Tuesday March 21, 2017 @07:42AM (#13921325)
For my (and I imagine many among us) use case, this is the takeaway:

PyPy2 can now import and run many C-extension packages, among the most notable are Numpy, Cython, and Pandas. Performance may be slower than CPython, especially for frequently-called short C functions. Please let us know if your use case is slow, we have ideas how to make things faster but need real-world examples (not micro-benchmarks) of problematic code.

Re:PyPy2.7 and PyPy3.5 v5.7 released by poizan42 on Tuesday March 21, 2017 @08:21AM (#13921502)
It would probably be better to port packages like Numpy to pure PyPy[0], but of course the CPython extensions compatibility layer is by itself important to increase the adoption of PyPy.

The problem besides the jit not being able to inline external functions is that the CPython structures, and especially the refcounting used in CPython is problematic [1]

[0]: Actually specifically numpy is being rewritten in PyPy as a RPython Mixed Module: http://doc.pypy.org/en/latest/extending.html#rpython-mixed-m... [1]: http://doc.pypy.org/en/latest/discussion/rawrefcount.html

Re:PyPy2.7 and PyPy3.5 v5.7 released by yedpodtrzitko on Tuesday March 21, 2017 @12:03PM (#13923446)
There's an attempt to create such NumPy fork already: https://bitbucket.org/pypy/numpy
Re:PyPy2.7 and PyPy3.5 v5.7 released by wfunction on Tuesday March 21, 2017 @07:52AM (#13921369)
If it's useful at all, PyPy 5.7.0 was noticeably slower than CPython 2.7.11 x64 when I was using SymPy. Haven't tried out the new release though.
Re:PyPy2.7 and PyPy3.5 v5.7 released by fijal on Tuesday March 21, 2017 @08:25AM (#13921522)
Hi

I've tried optimizing sympy and it's hard - there is a chance that it uses cython (which will slow things down on pypy) but also the gains were small (up to 2x) and only noticable if you run it for a while (at least few seconds).

SymPy is a bit of an example that while python as a language is not necesarilly slow, since you can make a plan for most parts, it breeds culture of unnecessary meta-programming, tons of copies and dict lookups that makes it REALLY HARD to optimize. ORMs go in the same category

Re:PyPy2.7 and PyPy3.5 v5.7 released by sanxiyn on Tuesday March 21, 2017 @08:22AM (#13921504)
Did you try x64 version of PyPy? x86 version is necessarily much slower (both CPython and PyPy) if your program is long-heavy. SymPy is actually part of benchmark used by PyPy and PyPy should be faster if you compare x64 to x64.
Re:PyPy2.7 and PyPy3.5 v5.7 released by wfunction on Tuesday March 21, 2017 @02:47PM (#13924994)
Nope, I'm on Windows.
Re:PyPy2.7 and PyPy3.5 v5.7 released by denfromufa on Tuesday March 21, 2017 @08:38AM (#13921603)
subset of sympy can run with symengine for speedup