• 280 likes • 426 Views
The Proposal Process, Best Practices & Policies. An open scientific discovery infrastructure for an integrated, persistent computation resource. Kent Milfeld TeraGrid Allocations Coordinator help@teragrid.org www.tacc.utexas.edu/~milfeld/pops/allocation_Proposal_Advice(9).ppt. Outline.
E N D
The Proposal Process,Best Practices & Policies An open scientific discovery infrastructure for an integrated, persistent computation resource. Kent Milfeld TeraGrid Allocations Coordinator help@teragrid.org www.tacc.utexas.edu/~milfeld/pops/allocation_Proposal_Advice(9).ppt
Outline • Terms • Overview of Proposal Types • Procedures • Resources • Awards • Entering Requests into POPS (Intent, Type, Forms, Document) • Request Document and guidelines • Research Objectives • Codes and methods to be used • Computational plan • Justification for SUs (TB) requested • Additional considerations • Review Criteria
The Lingo • Startup Development/testing/porting/benchmarking • Education Classroom, Training • Research Program (usually funded) • PI Principal Investigator • POPS Partnerships Online Proposal System • RAC Resource Allocations Committee • SU Service Unit = 1 CPU-hour • RoamingTeraGrid (Wide) Roaming (Access) Allocation Request Types Think: How much, how long, and how do I use resources
Eligibility • Principal investigator (PI) must be a researcher or educator at a U.S.-based institution, including federal research labs or commercial organizations, (Although additional information may be needed from researchers not affiliated with academic or non-profit research institutions.) • A postdoctoral researcher is eligible to serve as PI. • A qualified advisor may apply for an allocation for his or her class, but a high school, undergraduate or graduate student may not be a PI.
Web forms: Investigator, Grants, Resource Request,… Requires a written “proposal” (pdf upload) Reviewed by experts in same Field of Science 3 months from deadline to award availability For Research Projects Overview: Research Request https://pops-submit.teragrid.org/ ** • Proposals/Allocations • Limit: Unlimited • Reviewed : Quarterly • Deadlines: 15th of October, Jan., April, July • Awards begin: 1st of January, April, July, October
Web forms: Investigator, Resource Request,… Requires only an abstract Reviewed by a TG Staff (Startup Allocations Committee) 2 weeks from submission to award availability For code devel / performance eval / small-scaling computations / classroom & training instruction Overview: Startup/Education Requests https://pops-submit.teragrid.org/ ** • Requests/Allocations • Limit: 200,000SUs (less for systems <100TFLOPS) • Reviewed : Anytime • Awards begin: 2 weeks after submission
The Process: Steps • Review Systems: www.teragrid.org/userinfo/hardware/resources.php • Determine Type of Project • Login: https://pops-submit.teragrid.org (“Create POPS login” if first time.) • Select Intent (New, Renewal, Suppl/Justif/Progress/Ext) • Select project Type (Research > 200K SUs > Startup) • Fill in forms: PI/Co-PI Info, Proposal Info, Supporting Grants, Resource Request (alloc. request/machine). • Upload proposal document (for Research reqeust), and CV for all requests. • (Update at anytime and “Save to date”) • Press “Final Submission” when finished.
The Resources: Compute www.teragrid.org/userinfo/hardware/resources.php ** TeraGrid Resources Catalog ** Select: One or more resources or TeraGrid Roaming (Allocation across variety of systems) 579 TFLOPS Ranger System, TACC Startup/Education allocations are available on all systems, or as TeraGrid Roaming. SMP Blue/Gene Clusters Pools (condor) TFLOPS
The Resources: Storage • All compute allocations still include some default tape storage • Long-Term Allocable Storage: GPFS-WAN NCSA Proj. Wrk Space IU Data Capacitor IU HPSS (Tape) Archive NCSA Tape Storage SDSC Tape Storage • Data Collections IU Database/Collections NCSA Database (+Service) SDSC Database/Collections
The Resources: Visualization • Purdue TeraDRE • 48-node Purdue Condor Pool subcluster • NVIDIA GeForce 6600 GT GPUs (Gelato) • TACC SPUR • 8 Sun Fire x4000 series • 32 NVIDIA FX5600 GPUs, 128 cores • Access to Ranger queues and file systems • UC/ANL • 96 node, dual Intel Xeon • 1 Nvidia GeFORCE 660GT ABP card/node • Access to GPFS-WAN
The Awards • One per PI (generally) • 1-year duration, or multi-year • Unused SUs are forfeited at the end of an award period • Progress report required annually in renewal requests. • Progress report requiredyearly in multi-year awards • Add users to a grant via TeraGrid User Portal
Proposal “Intent” • New First Research or Startup/Education allocation Request • Renewal Subsequent allocation Requests Others: • Supplements Request additional resources during a 12-month allocation period Not for Startups! Reviewed by RAC (Resource Allocation Committee members). • Justifications PI may submit a rebuttal when proposal is not fully awarded. For addressing specific omissions (not to salvage horrible proposals) • Extensions Can extend award period an additional 6 months, with reasonable justification No additional resources! • Project Reports It is necessary to submit a Progress Report for approval of the next year’s allocation for Multi-year awards.
Proposal Type (level) • New • Renewal, Supplement, Justification, Report, Extension and Transfer • Select Proposal first, then type/level:
Proposal Forms https://pops-submit.teragrid.org/ • Straightforward (mostly) • Once you get to the Web-based data entry forms • Latest changes • Supporting grant information • You can now login to POPS with your TeraGrid Portal login • Coming soon • Better TeraGrid portal integration
Proposal Document(s) http://www.teragrid.org/userinfo/access/allocations.php ** Key to a successful review: • Adhere to page limits! • Justify allocation request. • There are examples. ** • CV required for all requests. • Abstract for startup/education • (in forms, or as pdf document) • Proposal Document for Research request are required. Does not include figures and references.
The Options Asking for Help help@teragrid.org Multi-year Awards Possible, but not recommended for new PIs Only Progress Reports required in subsequent years Advances UP to 20% of Research request can be provided in advance
The Resources: Advanced Support Program (ASP) http://www.teragrid.org/userinfo/asp.php • Dedicated, but limited, TeraGrid staff assistance (request FTE months) • At RAC meeting reviewers rate need for ASP (0-3) • Extra info required in proposal, see URL.
“Traditional” v. Community • MRAC proposals are accepted in four general categories of research activities • Single principal investigator • Large research collaborations (e.g., MILC consortium) • Community Consortiums (e.g., NEES) • Community Services (e.g., TeraGrid Gateways) • The general requirements for proposals of all four types remain largely the same. • Whether requesting compute, storage, visualization, or advanced support or some combination
General Proposal Outline • Research Objectives • Codes and methods to be used • Computational plan • Justification for SUs (TB) requested • Additional considerations Note: Sections III and IV are often integrated.
I. Research Objectives • Traditional proposals • Describe the research activities to be pursued • Community proposals • Describe the classes of research activities that the proposed effort will support. • Keep it short: You only need enough detail to support the methods and computational plan being proposed. • TIP—Reviewers don’t want to read the proposal you submitted to NSF/NIH/etc., but they will notice whether you have merit-reviewed (grant) funding.
II. Codes, Data, and Methods • Very similar between traditional and community proposals. • For compute requests • More significant if using ‘home-grown’ codes. • Provide performance and scaling details on problems and test cases similar to those being pursued. • Ideally, provide performance and scaling data collected by you for the specific resource(s) you are requesting • For storage requests • Provide description of data to be stored (organization, formats, collection mechanisms, permissions granted or received) • Describe the amount and expected growth of data to be stored.
III. Computational Plan • Traditional proposals • Explicitly describe the problem cases you will examine • BAD: “…a dozen or so important proteins under various conditions…” • GOOD: “…7 proteins [listed here; include scientific importance of these selections somewhere, too]. Each protein will require [X] number of runs, varying [x] parameters [listed here] [in very specific and scientifically meaningful ways]…” • Community proposals • Explicitly describe the typical use-case(s) that the gateway supports and the type of runs that you expect users to make • Describe how you will help ensure that the community will make scientifically meaningful runs (if applicable) • BAD: “…the gateway lets users run NAMD on TeraGrid resources…” • BETTER: “…users will run NAMD jobs on [biological systems like this]…” • BETTER STILL: “…the gateway allows users to run NAMD jobs on up to 128 processors on problem sizes limited [in some fashion]…”
IV. Justification of SUs, TBs • Traditional proposals • If you’ve done sections II and III well, this section should be a straightforward math problem • For each research problem, calculate the SUs required based on runs (base units) defined in III and the timings in section II, broken out appropriately by resource • Reasonable scaling estimates from test-case timing runs to full-scale production runs are acceptable. • Clear presentation here will allow reviewers to award time or storage in a rational fashion • Analogous calculations should apply for storage requests
IV. Justification of SUs, TBs • Community proposals • The first big trick: Calculating SUs when you don’t know the precise runs to be made a priori. • In Year 2 and beyond • Start with an estimate of total usage based on prior year’s usage patterns and estimate for coming year’s usage patterns (justify in Section V). • From this information, along with data from sections II and III, you can come up with a tabulation of SU estimates. • Year 1 requires bootstrapping • Pick conservative values (and justify them) for the size of the community and runs to be made, and calculate SUs. • TIP—Start modestly. If you have ~0 users, don’t expect the reviewers to believe that you will get thousands (or even hundreds) next year. • Analogous calculations for TBs of storage needed
V. Additional Review Considerations • Prior progress • From prior year allocation, Startup award, or work done locally • Ability to complete the work plan described(more significant for larger requests) • Sufficient merit-reviewed funding • Staff, both number and experience • Local computing environment • Other access to HPC resources • (e.g., Campus centers, DOE centers, etc.)
V. Additional Considerations • For community proposals, these components can provide key details: • Community Support and Management Plan • Describe the gateway interface — in terms of how it helps community burn SUs or access TBs. • Describe plans for growing the user community, “graduating” users to Research alloc. awards, regulating “gateway hogs” • Progress report • The actual user community and usage patterns • Manuscripts produced thanks to this service. • Local computing environment • Other HPC resources
Proposal Review Criteria • Methodology For compute requests, the choice of applications, methods, algorithms and techniques to be employed to accomplish the stated objectives should be reasonably justified. While the accomplishment of the stated objectives in support of the science is important, it is incumbent on proposers to consider the methods available to them and to use that which is best suited. (For storage requests, the data usage, access methods, algorithms and techniques to be employed to accomplish the stated research objectives should be reasonably justified. For shared collections, proposers must describe the public or community access methods to be provided.) • State Appropriateness of Computations for Scientific Simulations The computations must provide a precise representation of the physical phenomena to be investigated. They must also employ the correct methodologies and simulation parameters (step size, time scale, etc.) to obtain accurate and meaningful results. • Describe the Efficiency in Usage of Resources The resources selected must be used as efficiently as is reasonably possible. To meet this criterion for compute resources, performance and parallel scaling data should be provided for all applications to be used along with a discussion of optimization and/or parallelization work to be done to improve the applications.(For storage resources, information on required performance and expected access patterns should be provided for all data and collections to be stored and used along with a discussion of work done or planned to improve the efficiency of the data use.)
Questions? http://www.teragrid.org/userinfo/access/allocations.php http://teragrid.org/userinfo/access/accounts.php help@teragrid.org www.teragrid.org/userinfo/hardware/resources.php https://pops-submit.teragrid.org/ http://www.teragrid.org/userinfo/asp.php Example Proposals at: http://www.teragrid.org/userinfo/access/allocations.php http://www.tacc.utexas.edu/~milfeld/pops/Aliu.pdf http://www.tacc.utexas.edu/~milfeld/pops/MColvin.pdf