Thu 24 Jun 2021 21:00 - 21:05 at PLDI-B - Talks 3B: Architectures and Systems
Modern field-programmable gate arrays (FPGAs) have recently powered high-profile efficiency gains in systems from datacenters to embedded devices by offering ensembles of heterogeneous, reconfigurable hardware units. Programming stacks for FPGAs, however, are stuck in the past—they are based on traditional hardware languages, which were appropriate when FPGAs were simple, homogeneous fabrics of basic programmable primitives. We describe Reticle, a new low-level abstraction for FPGA programming that, unlike existing languages, explicitly represents the special-purpose units available on a particular FPGA device. Reticle has two levels: a portable \emph{intermediate language} and a target-specific \emph{assembly language}. We show how to use a standard \textit{instruction selection} approach to lower intermediate programs to assembly programs, which can be both faster and more effective than the complex metaheuristics that existing FPGA toolchains use. We use Reticle to implement linear algebra operators and coroutines and find that Reticle compilation runs up to $100$ times faster than current approaches while producing comparable or better run-time and utilization.
Thu 24 JunDisplayed time zone: Eastern Time (US & Canada) change
09:00 - 09:40 | |||
09:00 5mTalk | Reticle: A Virtual Machine for Programming Modern FPGAs PLDI Luis Vega University of Washington, Joseph McMahan University of Washington, Adrian Sampson Cornell University, Dan Grossman University of Washington, Luis Ceze University of Washington DOI | ||
09:05 5mTalk | Revamping Hardware Persistency Models: View-Based and Axiomatic Persistency Models for Intel-x86 and Armv8 PLDI Kyeongmin Cho KAIST, Sung-Hwan Lee Seoul National University, Azalea Raad Imperial College London, Jeehoon Kang KAIST DOI | ||
09:10 5mTalk | Developer and User-Transparent Compiler Optimization for Interactive Applications PLDI Paschalis Mpeis University of Edinburgh, Pavlos Petoumenos University of Manchester, Kim Hazelwood Facebook AI Research, Hugh Leather Facebook Link to publication DOI Media Attached | ||
09:15 5mTalk | Perceus: Garbage Free Reference Counting with Reuse PLDI Alex Reinking Microsoft Research, Ningning Xie University of Toronto, Leonardo de Moura Microsoft Research, Daan Leijen Microsoft Research DOI | ||
09:20 5mTalk | Filling Typed Holes with Live GUIs PLDI Cyrus Omar University of Michigan, David Moon University of Michigan, Andrew Blinn University of Michigan, Ian Voysey Carnegie Mellon University, Nick Collins University of Chicago, Ravi Chugh University of Chicago DOI Pre-print | ||
09:25 5mTalk | Boosting SMT Solver Performance on Mixed-Bitwise-Arithmetic Expressions PLDI Dongpeng Xu University of New Hampshire, Binbin Liu University of New Hampshire; University of Science and Technology of China, Weijie Feng University of Science and Technology of China, Jiang Ming University of Texas at Arlington, Qilong Zheng University of Science and Technology of China, Jing Li University of Science and Technology of China, Qiaoyan Yu University of New Hampshire DOI | ||
09:30 5mTalk | Automatically Enforcing Fresh and Consistent Inputs in Intermittent Systems PLDI Milijana Surbatovich Carnegie Mellon University, Limin Jia Carnegie Mellon University, Brandon Lucia Carnegie Mellon University DOI | ||
09:35 5mTalk | Bliss: Auto-tuning Complex Applications using a Pool of Diverse Lightweight Learning Models PLDI Rohan Basu Roy Northeastern University, Tirthak Patel Northeastern University, Vijay Gadepally MIT Lincoln Laboratory, Devesh Tiwari Northeastern University DOI |
21:00 - 21:40 | |||
21:00 5mTalk | Reticle: A Virtual Machine for Programming Modern FPGAs PLDI Luis Vega University of Washington, Joseph McMahan University of Washington, Adrian Sampson Cornell University, Dan Grossman University of Washington, Luis Ceze University of Washington DOI | ||
21:05 5mTalk | Revamping Hardware Persistency Models: View-Based and Axiomatic Persistency Models for Intel-x86 and Armv8 PLDI Kyeongmin Cho KAIST, Sung-Hwan Lee Seoul National University, Azalea Raad Imperial College London, Jeehoon Kang KAIST DOI | ||
21:10 5mTalk | Developer and User-Transparent Compiler Optimization for Interactive Applications PLDI Paschalis Mpeis University of Edinburgh, Pavlos Petoumenos University of Manchester, Kim Hazelwood Facebook AI Research, Hugh Leather Facebook Link to publication DOI Media Attached | ||
21:15 5mTalk | Perceus: Garbage Free Reference Counting with Reuse PLDI Alex Reinking Microsoft Research, Ningning Xie University of Toronto, Leonardo de Moura Microsoft Research, Daan Leijen Microsoft Research DOI | ||
21:20 5mTalk | Filling Typed Holes with Live GUIs PLDI Cyrus Omar University of Michigan, David Moon University of Michigan, Andrew Blinn University of Michigan, Ian Voysey Carnegie Mellon University, Nick Collins University of Chicago, Ravi Chugh University of Chicago DOI Pre-print | ||
21:25 5mTalk | Boosting SMT Solver Performance on Mixed-Bitwise-Arithmetic Expressions PLDI Dongpeng Xu University of New Hampshire, Binbin Liu University of New Hampshire; University of Science and Technology of China, Weijie Feng University of Science and Technology of China, Jiang Ming University of Texas at Arlington, Qilong Zheng University of Science and Technology of China, Jing Li University of Science and Technology of China, Qiaoyan Yu University of New Hampshire DOI | ||
21:30 5mTalk | Automatically Enforcing Fresh and Consistent Inputs in Intermittent Systems PLDI Milijana Surbatovich Carnegie Mellon University, Limin Jia Carnegie Mellon University, Brandon Lucia Carnegie Mellon University DOI | ||
21:35 5mTalk | Bliss: Auto-tuning Complex Applications using a Pool of Diverse Lightweight Learning Models PLDI Rohan Basu Roy Northeastern University, Tirthak Patel Northeastern University, Vijay Gadepally MIT Lincoln Laboratory, Devesh Tiwari Northeastern University DOI |