DAMP 2012

Declarative Aspects and Applications of Multicore Programming

Philadelphia, PA, USA

Saturday, January 28, 2012

To be held in conjunction with POPL 2012

Welcome!

Thanks to the generous support of Intel Corporation, our registration fees this year are very affordable. We are also able to provide you with lunch at the conference hotel, allowing for more interaction during lunch time.

Looking forward to seeing you in Philadelphia.

Program

Full (long) papers will appear at the ACM site. Link to available short papers are provided below.

Session I: Invited Talk

• 9:00-10:00 Simon Peyton Jones, Microsoft Research. The Replicate Problem in Nested Data Parallelism.

  Abstract: Guy Blelloch's showed us how to transform a nested data-parallel program (the one we want to write) into a flat data-parallel program (the one we want to run). But there's a problem: if done naively, you risk duplicating so much data that the asymptotic complexity of the algorithm gets worse, which pretty much defeats the object of the exercise. This is the "replicate problem."

  We bumped into the replicate problem when implementing Data Parallel Haskell. There are known solutions for first-order programs, but that doesn't work for us, because Haskell is higher order. (Indeed, moving to higher order already requires a pervasive change to the original flattening transformation.) In this talk I'll describe the flattening transformation, explain what the replicate problem is, and how we solve it.

Session II

• 10:30-11:00 Programmable Data Dependencies and Placements. Eva Burrows and Magne Haveraaen
• 11:00-11:20 Efficient Primitives for Creating and Scheduling Parallel Computations. Umut Acar, Arthur Chargueraud and Mike Rainey
• 11:20-11:40. Declarative Multicore Programming of Software-based Stateful Packet Processing. Harjot Gill, Dong Lin, Trisha Kothari and Boon Thau Loo.
• 11:40-12:00 Bottom-Up Logic Programming for Multicores. Flavio Cruz, Michael P. Ashley-Rollman, Seth Copen Goldstein, Ricardo Rocha and Frank Pfenning

Session III: Invited Talk

• 14:00-15:00 Rob Knauerhase, Intel, Extreme-scale programming systems
• 15:00-15:30 Expressive array constructs in an embedded GPU Kernel programming language. Koen Claessen, Mary Sheeran and Bo Joel Svensson

Session IV

• 16:00-16:30 Controlling loops in parallel Mercury code. Paul Bone, Zoltan Somogyi and Peter Schachte
• 16:30-17:00 Or-Parallel Prolog Execution on Multicores Based on Stack Splitting. Rui Vieira, Ricardo Rocha and Fernando Silva
• 17:00-17:20 Compilers must speak properties, not just code. Alex Shafarenko, Raimund Kirner and Frank Penczek

What is new?

Short papers:

Not exceeding 4 pages in ACM SIGPLAN format, short papers can present work-in-progress, describe applications of existing systems, can articulate a position, or present proposals for discussion at the workshop. We hope that short papers will enables researcher to obtain feedback on their ongoing work.

Long papers:

Not exceeding 10 pages in ACM SIGPLAN format, these papers should present previously unpublished, original work.

Applications and Failures:

We encourage papers on applications. Somewhat unconventionally, we invite all researchers to consider submitting papers not just on their successful results, but also on their failed attempts at solving a problem. Such papers should discuss reasons for failure and the lessons deduced and they can be submitted in the short or long paper formats.

Overview

The advent of multicore architectures has profoundly increased the importance of research in parallel computing. Multicore architectures, now commonplace, introduce several new dimensions of variability in both performance guarantees and architectural contracts, such as the memory model, while making it highly attractive and even necessary to develop novel programming languages, models, and paradigms for taking advantage of the benefits of parallelism.

Programs written in declarative languages, which control the use of side effects, can greatly simplify development of parallel programs by eliminating or limiting data races. Such languages include purely functional languages, (constraint-) logic programming languages, many data-driven or reactive languages, and other domain specific languages (e.g., MapReduce).

DAMP 2012 is the seventh in a series of one-day workshops seeking to explore ideas in declarative programming language design that will greatly simplify programming for multicore architectures, and more generally for tightly coupled parallel architectures. Starting this year, we welcome papers on a diverse set of topics ranging from language design to applications and practical experience. To foster discussion and enable exchange of ideas between different communities, we will accept both short and long papers. Short papers aim to provide an opportunity to receive feedback on incomplete, ongoing, or even failed work. We welcome reports of successes as well as failures.

Specific topics include, but are not limited to:

• language and compiler design and implementation
• run-time systems for supporting parallelism (e.g., garbage collection, scheduling)
• parallel applications and practical experience
• architectural support for parallel languages
• type systems and analysis for accurately detecting dependencies, aliasing, side effects, and impure features
• languages for the description of data placement and distribution
• technology for debugging parallel programs
• design and implementation of domain-specific declarative languages for multicore programming

Submissions

Submission Guidelines

We welcome both short and long communications. Long papers should not exceed 10 pages in ACM SIGPLAN conference format. Short papers should not exceed 4 pages, and may present work-in-progress, position statements on the state of the art, describe applications of existing systems, or just present proposals for discussion at the workshop. Somewhat unconventionally, we would like to welcome all researchers to consider submitting papers not just on their successful results, but also on their failed attempts with an emphasis on the reasons for failure and what lessons can be learned from them.

Both long and short communications will be refereed. Long communications will be published in the ACM Digital Library. Short communications will be made available informally at the DAMP web site but will not be published.

Papers to be published in the ACM Digital Library must adhere to the SIGPLAN Republication Policy. Concurrent submissions to other conferences, workshops, journals, or similar forums of publication are not allowed.

Important Dates

Steering Committee

Current Members

• Manuel Carro

• Maria Garcia de la Banda

• Gabriele Keller

• Leaf Petersen

• Simon Peyton Jones

• Enrico Pontelli

• John Reppy

Previous Workshops

Steering Committee Rules and Responsibilities

Steering Committee Responsibilities

• To provide continuity of the DAMP workshop and maintain the vision and vigor of the workshop.

• To officially represent the DAMP workshop with ACM SIGPLAN and to consult with the ACM SIGPLAN Chair on policy and planning.

• To select the date and location and appoint the General Chair and Programme Chair of each DAMP workshop instance.

• To provide support to the General Chair and Programme Chair for their organisation and running of the DAMP workshop.

Membership

Rules

• The Steering Committee will observe all ACM SIGPLAN rules pertaining to steering committees.

• At the conclusion of each instance of the workshop:

  • The General Chair and the Programme Chair of the oldest instance of the workshop and the oldest serving Member-At-Large shall step down from the Steering Committee.

  • The General Chair and the Programme Chair of the just finished instance of the workshop shall join the Steering Committee.

  • The old Steering Committee shall select a new Member-At-Large to join the Steering Committee.

  • The new Steering Committee shall select one of its members to be Steering Committee Chair.

• Around the time of each instance of the workshop, the Steering Committee shall:

  • Select the timeframe, location, and colocation, if any, of the next instance of the workshop.

  • Appoint the General Chair and Programme Chair for the next instance of the workshop.

• The Programme Chair shall select the Programme Committee members and obtain the Steering Committee's approval.

• Any vacancies in the Steering Committee shall be filled by the Steering Committee and the appointed person shall serve the remainder the leaving member's term.

• These rules can be changed by a majority vote of the Steering Committee.