[SciPy-user] Real-time plotting and data storage format questions
Thu Apr 3 12:39:51 CDT 2008
Gary Pajer wrote:
> On Tue, Apr 1, 2008 at 12:36 PM, firstname.lastname@example.org <email@example.com> wrote:
>> A small team of us are developing a new simulation package from the ground
>> up. Our legacy approach relied on MATLAB and other proprietary software. A
>> hope pf ours is to be able to shed the use of MATLAB for the analysis of our
>> simulation results and instead use python with scipy/numpy/matplotlib etc.
>> I've successfully installed and compiled optimized numpy/scipy and all the
>> supporting packages (ATLAS, FFTW, etc).
>> So far so good.
>> To the point - I have two questions:
>> 1) We would like to have a "scope" to monitor simulation outputs in real
>> time. We're using one tool that can take data over a tcp/ip port, but is
>> clunky and only works on a single platform. Does such a thing exists within
>> the python realm for plotting data in real time?
> Not sure of your detailed needs, but I can point out some options.
> One typical problem in python 'scope applications is speed. Most
> current graphics solutions weren't designed with realtime use in mind.
> Nonetheless, the ones that are being actively developed (e.g.
> matplotlib) are better than they used to be.
Where can I find information about Pwm-Blt ?
> This is pretty fast, but development has all but stopped.
> But not entirely stopped. Too bad. It's the fastest solution that I
> found in a not-exhaustive search. Look out: the Linux version of the
> current release has a library bug, and it's necessary to set an
> environment variable to get it to work. This is not documented, AFAIK
> I'm currently writing a realtime data acquisition/display application,
> but speed is not critical. I'm using Enthought Tool Suite (ETS) and
> Traits, and Chaco for data display. IMHO, I think that in five years
> ETS/Traits will be the most commonly used framework for scientific
I'm too writing one/two scope-like units (rewrite of PyPlot), because
MatPlotLib was too slow ;-)
I'm aiming at 16 channels 1 kHz per channel.
>> 2) Our simulation creates large (1-4 GB) data sets. Since we're writing this
>> simulation ourselves (C++) we can save the data in any format. Does anyone
>> have a suggestion for a specific format or API that's been found to be
>> optimal in terms of memory usage and ability to import into python for
>> analysis and plotting?
>> Thank you for any suggestions. We're still new with Python, so I apologize
>> if these questions seem mundane.
>> SciPy-user mailing list
> SciPy-user mailing list
More information about the SciPy-user