A Low Level Virtual Machine Backend for F2F Computing

Main contributer: Oleg Knut (oknut) Other contributers: Ulrich Norbisrath (ulno), Artjom Lind (artjom85) Here is small overview of what is done: 1) Backend for running LLVM code on F2F client side; ... and what has to be done: 2) Parser for code on client side to prevent run of potentially unsafe code.


2011.04.01    LLVM Computing Engine updated to support LLVM 2.8

Client backend for F2F:

Current status: Implemented in stable branch General idea: Right now F2F supports only running Python code, which is delivered from server application to client and sending back execution results. Idea of the project is to extend client application accept LLVM intermediate representation (IR) and of course make it possible to run it on client side. Additional functionality will allow us support any platforms and operation systems, that LLVM currently support. Another huge advantage is that F2F will be not stuck with single language to create computation instructions, but it's possible to create backend for any language (currently already done: C++, C, ObjC, Python) to create LLVM IR.