Mailman 3 March 2012 - NumPy-Discussion

cannot reference NpyAccessLib assembly for DLR app
by William Johnston March 28, 2012

March 28, 2012

Hello, I cannot add a reference to NpyAccessLib in an IronPython C# DLR app. The error message is: Failed while initializing NpyCoreApi: DllNotFoundException:Unable to load DLL 'NpyAccessLib': The specified module could not be found. (Exception from HRESULT: 0x8007007E) NumpyDotNet stack trace: at NumpyDotNet.NpyCoreApi.GetNativeTypeInfo(Int32& intSize, Int32& longsize, Int32& longLongSize, Int32& longDoubleSize) at NumpyDotNet.NpyCoreApi..cctor() The type initializer for 'NumpyDotNet.NpyCoreApi' threw an exception. Again, I cannot add the reference. Attached is a screenshot. Any suggestions? Regards, William Johnston

1 0

Looking for people interested in helping with Python compiler to LLVM
by Travis Oliphant March 28, 2012

March 28, 2012

Hey all, I gave a lightning talk this morning on numba which is the start of a Python compiler to machine code through the LLVM tool-chain. It is proof of concept stage only at this point (use it only if you are interested in helping develop the code at this point). The only thing that works is a fast-vectorize capability on a few functions (without for-loops). But, it shows how creating functions in Python that can be used by the NumPy runtime in various ways. Several NEPS that will be discussed in the coming months will use this concept. Right now there is very little design documentation, but I will be adding some in the days ahead, especially if I get people who are interested in collaborating on the project. I did talk to Fijal and Alex of the PyPy project at PyCon and they both graciously suggested that I look at some of the PyPy code which walks the byte-code and does translation to their intermediate representation for inspiration. Again, the code is not ready for use, it is only proof of concept, but I would like to get feedback and help especially from people who might have written compilers before. The code lives at: https://github.com/ContinuumIO/numba -Travis

15 26

different percentile implementations ?
by Pierre Haessig March 28, 2012

March 28, 2012

Hi, A quick question I've had in mind for some time but didn't find a solution : Is there a significant difference between "numpy.percentile" and "scipy.stats.scoreatpercentile" ? Of course the signatures are somewhat different, but I have the feeling that the overall purpose is the same. Am I missing something ? Best, Pierre

2 3

Numpy Memory Error with corrcoef
by Nicole Stoffels March 27, 2012

March 27, 2012

Dear all, I get the following memory error while running my program: *Traceback (most recent call last): File "/home/nistl/Software/Netzbetreiber/FLOW/src/MemoryError_Debug.py", line 9, in <module> correlation = corrcoef(data_records) File "/usr/lib/python2.7/dist-packages/numpy/lib/function_base.py", line 1992, in corrcoef c = cov(x, y, rowvar, bias) File "/usr/lib/python2.7/dist-packages/numpy/lib/function_base.py", line 1973, in cov return (dot(X, X.T.conj()) / fact).squeeze() MemoryError* Here an easy example how to reproduce the error: *#!/usr/bin/env python2.7 from pybinsel import open from numpy import * if __name__ == '__main__': data_records = random.random((459375, 24)) correlation = corrcoef(data_records) *My real data has the same dimension. Is this a size problem of the array or did I simply make a mistake in the application of corrcoef? I hope that you can help me! Thanks! Best regards, Nicole Stoffels -- Dipl.-Met. Nicole Stoffels Wind Power Forecasting and Simulation ForWind - Center for Wind Energy Research Institute of Physics Carl von Ossietzky University Oldenburg Ammerländer Heerstr. 136 D-26129 Oldenburg Tel: +49(0)441 798 - 5079 Fax: +49(0)441 798 - 5099 Web : www.ForWind.de Email: nicole.stoffels(a)forwind.de

5 5

"arg" searchsorted
by Bryan Van de Ven March 27, 2012

March 27, 2012

I recently got asked about an "arg" version of searchsorted, basically a version that could take a sorter as an additional argument. For instance: In [13]: a = np.array([5,6,8,1,6,9,0]) In [14]: s = np.argsort(a) In [17]: s Out[17]: array([6, 3, 0, 1, 4, 2, 5]) In [18]: np.searchsorted(a, [1.5, 7.2], sorter=sorter) Out[18]: array([2, 5]) In [19]: np.searchsorted(a, -1.5, sorter=s) Out[19]: 0 In [20]: np.searchsorted(a, 11.5, sorter=s) Out[20]: 7 In [21]: np.searchsorted(a, 8.5, sorter=s) Out[21]: 6 In [32]: np.searchsorted(a, 6, side='left', sorter=sorter) Out[32]: 3 In [33]: np.searchsorted(a, 6, side='right', sorter=sorter) Out[33]: 5 I've already implemented this, I wanted to ping the list and see if there were any thoughts about this small feature. Or should I just submit a PR for consideration? Thanks, Bryan

2 4

Linking against MKL but still slow?
by Christoph Dann March 27, 2012

March 27, 2012

Dear list, so far I used Enthoughts Python Distribution which contains a compiled version of numpy linked against MKL. Now, I want to implement my own extensions to numpy, so I need my build numpy on my own. So, I installed Intel Parallel studio including MKL and the C / Fortran compilers. I linked against the same libraries as Enthought: In [2]: np.show_config() lapack_opt_info: libraries = ['mkl_lapack95_lp64', 'iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] blas_opt_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] lapack_mkl_info: libraries = ['mkl_lapack95_lp64', 'iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] blas_mkl_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] mkl_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] and used the intel compilers to build numpy. Since I have an i7 processor I compiled including AVX instructions. Yet, EPD is double as fast as my own build executing the simple benchmark from: http://dpinte.wordpress.com/2010/01/15/numpy-performance-improvement-with-t… I expected at least comparable performance. How is such a decrease possible? Did I miss a significant part to make numpy really fast? Thanks Christoph

2 2

AttributeError with shape command
by Stephanie Cooke March 27, 2012

March 27, 2012

Hello, I am new to numpy. When I try to use the command array.shape, I get the following error: AttributeError: 'list' object has no attribute 'shape' Is anyone familiar with this type of error? Thanks

3 2

How to Extract the Number of Rows and Columns in a Matrix
by Stephanie Cooke March 26, 2012

March 26, 2012

Hello, I would like to extract the number of rows and columns of a matrix individually. The shape command outputs the rows and columns together, but are there commands that will separately give the rows and separately give the columns? Thanks

3 2

label NA and datetime as experimental
by Ralf Gommers March 26, 2012

March 26, 2012

Hi, We decided to label both NA and datetime APIs as experimental for the 1.7.0 release. I made a PR that does this, please review: https://github.com/numpy/numpy/pull/240 Ralf

3 4

Linking against MKL but still slow?
by Christoph Dann March 26, 2012

March 26, 2012

Dear list, so far I used Enthoughts Python Distribution which contains a compiled version of numpy linked against MKL. Now, I want to implement my own extensions to numpy, so I need my build numpy on my own. So, I installed Intel Parallel studio including MKL and the C / Fortran compilers. I linked against the same libraries as Enthought: In [2]: np.show_config() lapack_opt_info: libraries = ['mkl_lapack95_lp64', 'iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] blas_opt_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] lapack_mkl_info: libraries = ['mkl_lapack95_lp64', 'iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] blas_mkl_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] mkl_info: libraries = ['iomp5', 'mkl_def', 'mkl_intel_lp64', 'mkl_intel_thread', 'mkl_core', 'pthread'] library_dirs = ['/opt/intel/lib/intel64', '/opt/intel/mkl/lib/intel64'] define_macros = [('SCIPY_MKL_H', None)] include_dirs = ['/opt/intel/include/'] and used the intel compilers to build numpy. I have an Intel i7 processor and compile including AVX instructions. Yet, EPD is double as fast as my own build executing the simple benchmark from: http://dpinte.wordpress.com/2010/01/15/numpy-performance-improvement-with-t… I expected at least comparable performance. How is such a decrease possible? Did I miss a significant part to make numpy really fast? Thanks Christoph

1 0