Write a Blog >>
PLDI 2021
Sun 20 - Sat 26 June 2021 PLDI
Thu 24 Jun 2021 09:10 - 09:15 at PLDI-B - Talks 3B: Architectures and Systems
Thu 24 Jun 2021 21:10 - 21:15 at PLDI-B - Talks 3B: Architectures and Systems

Traditional offline optimization frameworks rely on representative hardware, software, and inputs to compare different optimization decisions on. With application-specific optimization for mobile systems though, the idea of a representative test bench is unrealistic while creating offline inputs is non-trivial. Online approaches partially overcome these problems but they might expose users to suboptimal or even erroneously optimized code. As a result, our mobile code is poorly optimized and this results in wasted performance, wasted energy, and user frustration. In this paper, we introduce a novel compiler optimization approach designed for mobile applications. It requires no developer effort, it tunes applications for the user’s device and usage patterns, and has no negative impact on the user experience. It is based on a lightweight capture and replay mechanism. In its online stage, it captures the state accessed by any targeted code region. By re-purposing existing OS capabilities, it keeps the overhead low. In its offline stage, it replays the code region but under different optimization decisions to enable sound comparisons of different optimizations under realistic conditions. Coupled with a search heuristic for the compiler optimization space, it allows us to discover optimization decisions that improve performance without testing these decisions directly on the user. We implemented a prototype system in Android based on LLVM combined with a genetic search engine. We evaluated it on both benchmarks and real Android applications. Online captures are infrequent and each one introduces an overhead of less than 15ms on average. For this negligible effect on user experience, we achieve speedups of 44% on average over the Android compiler and 35% over LLVM -O3.

Conference Day
Thu 24 Jun

Displayed time zone: Eastern Time (US & Canada) change

09:00 - 09:40
Talks 3B: Architectures and SystemsPLDI at PLDI-B +12h
09:00
5m
Talk
Reticle: A Virtual Machine for Programming Modern FPGAs
PLDI
Luis VegaUniversity of Washington, Joseph McMahanUniversity of Washington, Adrian SampsonCornell University, Dan GrossmanUniversity of Washington, Luis CezeUniversity of Washington
DOI
09:05
5m
Talk
Revamping Hardware Persistency Models: View-Based and Axiomatic Persistency Models for Intel-x86 and Armv8
PLDI
Kyeongmin ChoKAIST, Sung-Hwan LeeSeoul National University, Azalea RaadImperial College London, Jeehoon KangKAIST
DOI
09:10
5m
Talk
Developer and User-Transparent Compiler Optimization for Interactive Applications
PLDI
Paschalis MpeisUniversity of Edinburgh, Pavlos PetoumenosUniversity of Manchester, Kim HazelwoodFacebook AI Research, Hugh LeatherFacebook
Link to publication DOI Media Attached
09:15
5m
Talk
Perceus: Garbage Free Reference Counting with Reuse
PLDI
Alex ReinkingMicrosoft Research, Ningning XieUniversity of Hong Kong, Leonardo de MouraMicrosoft Research, Daan LeijenMicrosoft Research
DOI
09:20
5m
Talk
Filling Typed Holes with Live GUIs
PLDI
Cyrus OmarUniversity of Michigan, David MoonUniversity of Michigan, Andrew BlinnUniversity of Michigan, Ian VoyseyCarnegie Mellon University, Nick CollinsUniversity of Chicago, Ravi ChughUniversity of Chicago
DOI Pre-print
09:25
5m
Talk
Boosting SMT Solver Performance on Mixed-Bitwise-Arithmetic Expressions
PLDI
Dongpeng XuUniversity of New Hampshire, Binbin LiuUniversity of New Hampshire; University of Science and Technology of China, Weijie FengUniversity of Science and Technology of China, Jiang MingUniversity of Texas at Arlington, Qilong ZhengUniversity of Science and Technology of China, Jing LiUniversity of Science and Technology of China, Qiaoyan YuUniversity of New Hampshire
DOI
09:30
5m
Talk
Automatically Enforcing Fresh and Consistent Inputs in Intermittent Systems
PLDI
Milijana SurbatovichCarnegie Mellon University, Limin JiaCarnegie Mellon University, Brandon LuciaCarnegie Mellon University
DOI
09:35
5m
Talk
Bliss: Auto-tuning Complex Applications using a Pool of Diverse Lightweight Learning Models
PLDI
Rohan Basu RoyNortheastern University, Tirthak PatelNortheastern University, Vijay GadepallyMIT Lincoln Laboratory, Devesh TiwariNortheastern University
DOI
21:00 - 21:40
Talks 3B: Architectures and SystemsPLDI at PLDI-B
21:00
5m
Talk
Reticle: A Virtual Machine for Programming Modern FPGAs
PLDI
Luis VegaUniversity of Washington, Joseph McMahanUniversity of Washington, Adrian SampsonCornell University, Dan GrossmanUniversity of Washington, Luis CezeUniversity of Washington
DOI
21:05
5m
Talk
Revamping Hardware Persistency Models: View-Based and Axiomatic Persistency Models for Intel-x86 and Armv8
PLDI
Kyeongmin ChoKAIST, Sung-Hwan LeeSeoul National University, Azalea RaadImperial College London, Jeehoon KangKAIST
DOI
21:10
5m
Talk
Developer and User-Transparent Compiler Optimization for Interactive Applications
PLDI
Paschalis MpeisUniversity of Edinburgh, Pavlos PetoumenosUniversity of Manchester, Kim HazelwoodFacebook AI Research, Hugh LeatherFacebook
Link to publication DOI Media Attached
21:15
5m
Talk
Perceus: Garbage Free Reference Counting with Reuse
PLDI
Alex ReinkingMicrosoft Research, Ningning XieUniversity of Hong Kong, Leonardo de MouraMicrosoft Research, Daan LeijenMicrosoft Research
DOI
21:20
5m
Talk
Filling Typed Holes with Live GUIs
PLDI
Cyrus OmarUniversity of Michigan, David MoonUniversity of Michigan, Andrew BlinnUniversity of Michigan, Ian VoyseyCarnegie Mellon University, Nick CollinsUniversity of Chicago, Ravi ChughUniversity of Chicago
DOI Pre-print
21:25
5m
Talk
Boosting SMT Solver Performance on Mixed-Bitwise-Arithmetic Expressions
PLDI
Dongpeng XuUniversity of New Hampshire, Binbin LiuUniversity of New Hampshire; University of Science and Technology of China, Weijie FengUniversity of Science and Technology of China, Jiang MingUniversity of Texas at Arlington, Qilong ZhengUniversity of Science and Technology of China, Jing LiUniversity of Science and Technology of China, Qiaoyan YuUniversity of New Hampshire
DOI
21:30
5m
Talk
Automatically Enforcing Fresh and Consistent Inputs in Intermittent Systems
PLDI
Milijana SurbatovichCarnegie Mellon University, Limin JiaCarnegie Mellon University, Brandon LuciaCarnegie Mellon University
DOI
21:35
5m
Talk
Bliss: Auto-tuning Complex Applications using a Pool of Diverse Lightweight Learning Models
PLDI
Rohan Basu RoyNortheastern University, Tirthak PatelNortheastern University, Vijay GadepallyMIT Lincoln Laboratory, Devesh TiwariNortheastern University
DOI