gpu programming with python

theano seems to have an advanced api, but it might be too advanced since it uses its own optimization and other magic goodies, and it seems so focused on its own objects that it's like a metaprogramming language. still, dangerously interesting. uses cuda, so nvidia only. here's some free advice from their tutorial:

Only computations with float32 data-type can be accelerated. Better support for float64 is expected in upcoming hardware but float64 computations are still relatively slow (Jan 2010).

Matrix multiplication, convolution, and large element-wise operations can be accelerated a lot (5-50x) when arguments are large enough to keep 30 processors busy.

Indexing, dimension-shuffling and constant-time reshaping will be equally fast on GPU as on CPU.

Summation over rows/columns of tensors can be a little slower on the GPU than on the CPU

Copying of large quantities of data to and from a device is relatively slow, and often cancels most of the advantage of one or two accelerated functions on that data. Getting GPU performance largely hinges on making data transfer to the device pay off.

the same website has a link to cudamat, which might be a more cooperative if lower-level way to go. target seems to be basic matrix and element-wise ops. actively developed.

pystream was developed as another cuda wrapper until about mid 2008, then abandoned when the company went off to develop gpulib, a cuda api for idl and (?) matlab.

pycuda handles the background stuff, but you still have to feed it c code (though there are tools for run-time code generation). looks like it's actively developed, though, with an impressive list of users, and like the others it does play nicely with numpy arrays. i think this is the place to start.

pygpu uses pycg and pyglew to generate cg code directly from python. so you write python and it will run on nvidia or ati hardware, under both linux and windows. unfortunately, neither the homepage nor the google code page show any signs of activity in the last few years. pycg (developed by the same guy) seems to have trickled off in late 2007, though ubuntu packages were uploaded to launchpad just a year ago. too bad, this looked like it might have been a good one.

the gpu stuff on scikits oddly seems intended for actual graphics stuff.

pyopencl ? here's a faq page contrasting cuda and opencl.

pycublas maybe just does matrix mult.

bond spread data

fred (research.stlouisfed.org) has a lot of economic indicator historical data, including moody's bond yields, but only for aaa and baa. i'd like to find something for more junky bonds....

moodys.com certainly would have these data, along with others i'd like to see. but registration is required and i don't know if that stuff is available for free.

Best Practices in Estimating the Cost of Capital: Survey and Synthesis

Robert F. Bruner, Kenneth M. Eades, Robert S. Harris, and Robert C. Higgins

nice peek at the popularity of various financial analysis techniques. just a bit dated now, as it came out in 1998, but still worth a look.

tea party != conservative populism

jeffrey friedman makes some good points about why the tea party leaders should not assume (or maybe even aim for) a conservative populism. here's the most sobering stat for me:

An April Rasmussen survey found that only 60 percent of Americans now believe that capitalism is better than socialism. Among those under 30, socialism and capitalism are nearly tied at 33 percent and 37 percent.

oil spill -> green pork

charles krauthammer lets fly a critique of obama's face-the-nation. seems a bit harsh for ck, but he does an effective job of deconstructing the basic argument.

obama vs. science

jonah goldberg makes an interesting connection between the anti-science claims against bush and the drilling moratorium. a minor point, perhaps, but one to put on the record.

portfolioscience

here's a company that sells software for portfolio optimization and risk analysis. maybe i should fill out their form and check out the demo some time, just to see how they do it and which metrics seem to be featured. in particular i'm curious about the riskapi efficient frontier optimizer.

aorda is another one, started circa 2006 by a ufl prof who was one of the early proponents of cvar. free download for crippleware, but have to register first. pay versions are very highly priced: commercial license is $10k/year.