Saturday, May 14, 2011

CS: Could the Google Exacycle program be used to Map the Donoho-Tanner phase transition ?

I cannot find the exact reference but I seem to recall that about 3 CPU-years or about 26,000 CPU hours, were needed to produce the Donoho-Tanner phase transition. But what if you were proposed to have access to 100,000,000 CPU hours or 3,800 times these 3 CPU years ? This is what the Google proposes in the Google Exacycle program. Could we look at several extensions of this transition (noisy,different measurement matrices, solvers,....) as mentioned earlier ? The application deadline is 11:59 p.m. PST May 31, 2011. From the site:

What is the Google Exacycle for Visiting Faculty Grant Program?
Google Exacycle for Visiting Faculty is a new grant program for high-performance, CPU-intensive computing. In its first year, the program invites proposals for large-scale, computationally intensive research projects. The program awards sizable allocations on Google’s computing infrastructure to address grand challenges in science and engineering. We will award a total of approximately one billion core-hours to drive transformational research in diverse fields such as astronomy, biology and medicine, earth sciences, mathematics and physics.
Google Exacycle for Visiting Faculty is not a conventional grant program. We aim to stimulate advances in science and engineering research by supporting the computational needs of projects that push boundaries and reach for remarkable breakthroughs.
Award Information
Google will award a total of approximately one billion core-hours to up to 10 distinguished researchers and postdoctoral scholars worldwide. We are looking for projects that can consume at least 100 million core-hours. All grantees, including those outside of the U.S., will be invited to work on-site at specific Google offices in the U.S. or abroad. The exact office location will be determined at the time of project selection. Award announcements will be ongoing.
Google Exacycle for Visiting Faculty is open to all academic researchers who require unusually large computer time allocations in their pursuit of transformational advances in science and engineering. The intent of the program is to support computationally intensive projects that are enabled through the availability of massive computation capability. Both U.S. and international researchers are eligible to apply. Please note that travel requirements, such as passport and visa, as well as expenses related to relocation, travel and cost of living will be the grant recipient’s responsibility.
Usage Agreement
Awardees will participate through Google’s Visiting Faculty Program; faculty members need to have full-time status at an academic institution; postdoctoral scholars are required to have an academic appointment confirmed by a university. Awardees sign a (limited) employee agreement with Google. Program participants are urged to review the application details and guidelines and consult with their appropriate institutional representatives should they receive an award under this solicitation.
Awardees will work on a flexible schedule (part-time, full-time or semester-based) for up to one year.
Technical Specifications and Requirements
Proposals that are ideal for Google Exacycle include, but are not limited to, research projects like Folding@HomeRosetta@Home, variousBOINC projects, and grid parameter sweeps. Other examples include large-scale genomic search and alignment, protein family modeling and sky survey image analysis.
The best projects will have a very high number of independent work units, a high CPU to I/O ratio, and no inter-process communication (commonly described as Embarrassingly or Pleasantly Parallel). The higher the CPU to I/O rate, the better the match with the system. Programs must be developed in C/C++ and compiled via Native Client. Awardees will be able to consult an on-site engineering team.
Preference will be given to projects that are fairly high-risk/high-reward with the potential to drastically transform the scientific landscape. Even projects that yield negative results can still provide public data that the community can continue to analyze. At completion of the project, we recommend, but do not require, that all the researcher's data be made freely available to the academic community.
Applying to the Google Exacycle for Visiting Faculty Grant Program
Researchers and full-time faculty members from universities worldwide may apply by sending the following items to

  1. Two-page proposal in PDF format, including these items/sections:
  • Visitor's full name and contact information (postal address, email, phone number)
  • Visitor's affiliation (University, School, College and/or Department)
  • Research abstract and goals
  • Project description
  1. Description should include estimates of total resource as well as individual instance resource (CPU, RAM, I/O, data set) consumption
  1. Description of software required and evidence that it could be recompiled to Google Native Client
  1. Description of the data processing pipeline to prepare data for and process results from the CPU intensive part of Exacycle
  • Expected outcomes and results
  1. A current CV
Due Dates
Proposals will be accepted electronically starting today (no paper applications). If you require special accommodations to apply, please let us know by writing to The application deadline is 11:59 p.m. PST May 31, 2011. Applicants are encouraged to send in their proposals early as awards will be granted starting in June. Access to Google resources will be granted following the execution of appropriate agreements.

Figure excerpted from this presentation in 2009 at Los Alamos.

No comments: