[LLVMdev] gsoc 2013
criswell at illinois.edu
Wed May 1 19:20:50 PDT 2013
On 4/30/13 9:02 PM, Yun-Wei Lee wrote:
> Hello, it may a bit too late.
> I want to join gsoc 2013 and my proposal is to do static analyze on
> gpu programming, such as CUDA.
> Recently, I have been studying materials relating to SSA, data flow
> analysis, fixed point theory and etc under the instruction of one of
> the Ph.D in UIUC.
I'm at Illinois as well (I work for Vikram Adve). May I ask under who
you are studying?
> After discuss with him, we came up this proposal.
> I think I can do this well because I am also familiar with many
> algorithms on graph, such as scc. (I'm a contestant of ACM ICPC in my
You will need to have a much more concrete proposal for GSoC. I'm
assuming that your static analysis is for bug finding; is that correct?
If so, your proposal should describe what sort of errors your static
analysis will find, how your analysis will work, and how it will be
integrated into Clang/LLVM. You should also state whether your analysis
will work at the Clang AST level, the LLVM IR level, or some combination
I heard the proposal deadline was Friday (although I'm not mentoring
this summer, so I could be wrong on the date). If you want feedback,
you should add these things to your proposal and send them to the
correct list (cfe-dev for Clang, llvmdev for LLVM) for feedback.
Your final proposal should also include a detailed schedule of how the
work will proceed and long it will take; it should also include an
explanation of why you're qualified for the position. You will
basically need to provide something like a CV describing your work,
school, and coding experience. I believe mentors will place more weight
on coding experience than academic credentials.
-- John T.
> Wish we can discuss more.
> Yun-Wei Lee
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-dev