User: diener Date: 2012/04/06 07:33 AM

Modified: /supercomputing/ supercomputing.tex

Log: A few tweaks.

File Changes:

Directory: /supercomputing/ ===========================

File [modified]: supercomputing.tex Delta lines: +6 -3 =================================================================== --- supercomputing/supercomputing.tex 2012-04-06 12:07:18 UTC (rev 87) +++ supercomputing/supercomputing.tex 2012-04-06 12:33:27 UTC (rev 88) @@ -36,21 +36,24 @@ {\em An abstract with a general description of the work may be sufficient, but additional information that will enhance your request may include explanations for why the project needs access to XSEDE resources, how the estimate of resources request was reached, and how the resources requested were selected.}

You'll need to submit CV's for PI/co-PI's, and then your request will be considered within a few weeks. On Kraken and Ranger, you can ask for up to 200kSUs, or roughly 20 CPU-years of time. + +The PI can be a post-doc, research staff or a faculty member so you might +have to twist the arm of your advisor. }

\frame{\frametitle{Your initial proposal}

The recommended format for an initial proposal goes something like this: One paragraph that reads like a scientific abstract. Lay out reasonable goals, and motivate the interesting nature of your proposed research.

-In the second paragraph, delve into your code in more detail. Describe how it currently runs, mention any scaling results you have, whether it is built on commonly used packages, how you have modified it, etc. Try to demonstrate that it should be running on a parallel architecture and will do so efficently. +In the second paragraph, delve into your code in more detail. Describe how it currently runs, the numerical methods used, mention any scaling results you have, whether it is built on commonly used packages, how you have modified it, etc. Try to demonstrate that it should be running on a parallel architecture and will do so efficently. }

\frame{\frametitle{Your research proposal}

Try to remember the following when submitting a larger-scale Research proposal: \begin{itemize} -\item There are numerical relativists on the XSede Resource Allocation Committee (XRAC). Don't spend page after page justifying why the BSSN formalism is a tremendous breakthrough, or why the Einstein Toolkit ior some other software package s well-suited to use on Ranger/Kraken... they already believe it! -\item You do want to demonstrate you understand how to show that codes scale well to the run sizes you propose. Plot scaling results for +\item There are numerical relativists on the XSede Resource Allocation Committee (XRAC). Don't spend page after page justifying why the BSSN formalism is a tremendous breakthrough, or why the Einstein Toolkit or some other software package is well-suited to use on Ranger/Kraken... they already believe it! +\item You do want to demonstrate that you understand how to show that codes scale well to the run sizes you propose. Plot scaling results for cases that are as close as you can reasonably get to the real runs you want to do. This is especially true if you've modified an existing code, even if it is in wide use in its released form. \end{itemize} }