1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" 2 "http://www.w3.org/TR/html4/strict.dtd"> 3<!-- Material used from: HTML 4.01 specs: http://www.w3.org/TR/html401/ --> 4<html> 5<head> 6 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1"> 7 <title>Polly - Polyhedral optimizations for LLVM</title> 8 <link type="text/css" rel="stylesheet" href="menu.css"> 9 <link type="text/css" rel="stylesheet" href="content.css"> 10</head> 11<body> 12<div id="box"> 13<!--#include virtual="menu.html.incl"--> 14<div id="content"> 15 <!--*********************************************************************--> 16 <h1>Open Projects</h1> 17 <!--*********************************************************************--> 18 19 LLVM Polly keeps here a list of open projects which each of themselves would 20 be a great contribution to Polly. All of these projects are meant to be self 21 contained and should take a newcomer around 3-4 months of work. The projects 22 we propose are all suitable as <a 23 href="https://developers.google.com/open-source/gsoc/">Google Summer of 24 Code</a> projects. In case you are interested in a Google Summer of code 25 project make sure to reach out via the Polly <a 26 href="https://groups.google.com/group/polly-dev">mailing list</a> early to 27 discuss your project proposal. 28 29 <h3>Integrate Polly with the LLVM vectorizers</h3> 30 Polly is not only a self-contained optimizer, but also provides a powerful 31 dependence and other program analyses. Currently, these analyses are only used 32 for our own optimizations. However, LLVM passes such as the loop vectorizer 33 would clearly benefit from having direct access to the available Polly 34 analyses. In this project, you would define in collaboration with the LLVM 35 community and considering existing dependence analysis interface a new 36 dependence analysis interface for Polly that allows passes to directly query 37 Polly analysis. Even though this project sounds straightforward at a first 38 glance, sorting out how to actually make this happen with the current and 39 the new pass managers, understanding how and when to invalidate the Polly 40 analysis and if dependence information can be computed on-demand make this 41 still a challenging project. If successful, this project may be a great way 42 to bring features of Polly to standard -O3 optimizations. 43 44 <h3>Register tiling to obtain fast BLAS kernels with Polly</h3> 45 Even though Polly is already able to speed up compute kernels significantly, 46 when comparing to the best BLAS routines we still are at least one order of 47 magnitude off. In this project you will investigate what is needed to close 48 this performance gap. Earlier investigations have shown that register tiling 49 is one important piece towards this goal. In combination with good tile size 50 models and some back-end work, this project is shooting to make common blas 51 operations, but also many non-blas kernels competitive with vendor math 52 libraries and outperforming the code icc/gcc currently generate. 53 54 <h3>Polly support for Julia - First steps</h3> 55 <a href="https://julialang.org/">Julia</a> is a new matlab style programming 56 language that provides C like performance for scientific computing. Even 57 though Julia also translates to LLVM-IR, parsing and optimizing Julia code 58 poses new challenges that currently prevent Polly from optimizing Julia 59 code despite the clear need for optimizations such as loop-tiling for Julia. 60 In this project you will -- starting from first proof-of-concept patches -- 61 integrate Polly into Julia and ensure that Julia code can benefit from the 62 same high-level loop optimizations as todays C code already does. If time 63 permits, making Polly's recent bound-check elimination logic work in Julia 64 code would allow the optimization of Julia code, even if save out-of-bound 65 checking is used. 66 <h3>Interactive Polyhedral Web Calculator</h3> 67 At the core of Polly we use the isl math library. isl allows us to describe 68 loop transformations with relatively simple higher level operations while 69 still providing the full expressiveness of integer polyhedra. To understand 70 and describe the transformations we are performing it is often very convenient 71 to quickly script example transformations in a scripting language like python. 72 isl already comes with a python binding generator, with 73 pypyjs there is a python interpreter for the web and with emscriptem isl 74 itself can also be compiled to javascript. In this project you combine all 75 these components to obtain an interactive polyhedral web calculator, that uses 76 latest web technology to nicely illustrate the integer polyhedra you obtain. 77</div> 78</div> 79</body> 80</html> 81