Well, I'm a couple days late to the conversation, but a couple comments: If Nathan doesn't have a background knowledge in math, then I think prime counting/enumeration are very good projects, especially if he wants to look at parallizable algorithms. The current prime_pi in sage was my first real coding project, and at that point my math knowledge was groups plus learning introductory real analysis.
While tackling algorithms meeting or besting mathematica is a bit ambitious, there are plenty of algorithms out there that are faster than what I wrote, but still manageable. Most of these look very parallizable, although I haven't messed with that at all. Also, sage does not currently include a segmented sieve, which is also an extremely approachable project. Nathan feel free to contact me, I can point you to references, and may be able to clarify some stuff, although I can't promise much time with applying to grad schools and all. -- Andrew On Mon, Nov 8, 2010 at 03:48, Jason Grout <jason-s...@creativetrax.com>wrote: > On 11/7/10 4:03 AM, David Kirkby wrote: > >> On 5 November 2010 18:48, nekopczynski<nekopczyn...@plymouth.edu> wrote: >> >>> Hello Everyone, >>> I am currently a senior computer science student with a strong math >>> background. One of my professors, Dana Ernst, also a member here, >>> recommended that I post here. I am looking for a senior project >>> involving programing, algorithms, parallel processing, and >>> mathematics. Does any know of such a project that could be developed >>> or needs to be developed for the sage community. Limits on the >>> project would require that it be completed in about 3-4 months, no >>> longer. Thanks to anyone who can help. >>> >>> Nathan Kopczynski >>> >> >> I had another thought on a possible project. CUDA is something that I >> expect any computer science student would be aware of. To my >> knowledge, there is no support in Sage for CUDA, though there has been >> some interest in this. >> >> http://wiki.sagemath.org/CUDA >> >> The availability of CUDA hardware might be an issue. I don't know if >> anyone has a machine fitted with CUDA processors. >> >> My Sun Ultra 27 has an NVIDIA Quadro FX 3800 graphics accelerator >> card, which has a 192 CUDA cores, 1 GB of RAM and a memory bandwidth >> of over 50 GB/s. >> >> http://www.nvidia.com/object/product_quadro_fx_3800_us.html >> >> I've no idea if there's any way I can make a subset of those cores >> available, but I don't do anything that needs the level of performance >> this card can provide. >> >> However, I suspect the ideal way to do CUDA development would be to >> beg/borrow/buy/steal a CUDA development TWSLA card. >> >> It might be possible to borrow such a card from Nvida for a 3-month >> student project. >> >> Mathematica 8 will support CUDA >> >> http://www.wolfram.com/products/mathematica-cuda-free-white-paper.html >> >> I think have CUDA development experience on a CV would be very useful. >> > > > I was going to suggest two things: > > 1. I had a student look at the MAGMA and PLASMA libraries the other > semester in numerical analysis for a final project. > > http://icl.cs.utk.edu/magma/ > > http://icl.cs.utk.edu/plasma/ > > Unfortunately, the magma team has not yet released the source code (though > they plan to). Having yet another request for the source would be good. > > 2. If you are interested in CUDA development, there probably are some Tesla > boxes that you could get an account on. They're not mine, but I know at > least one student that was able to get an account on a box thanks to > another's generosity. Email me off list if you're interested in more > details. > > Thanks, > > Jason > > > -- > To post to this group, send an email to sage-devel@googlegroups.com > To unsubscribe from this group, send an email to > sage-devel+unsubscr...@googlegroups.com<sage-devel%2bunsubscr...@googlegroups.com> > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org > -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
