This program is tentative and subject to change.

Thu 24 Oct 2024 14:00 - 14:20 at San Gabriel - Compilers and Optimisation 2 Chair(s): Manu Sridharan

Although an optimizing compiler often produces a fast program, we cannot expect it to produce an optimal program. Moreover, the amount of headroom between the program we get and the optimally fast program is usually unknown. A superoptimizing compiler—that performs a meaningful search of the program space—can help narrow this gap. We created Minotaur, a superoptimizing LLVM pass that uses program synthesis to improve the quality of its code generation, with a specific focus on integer and floating-point SIMD code. It discovers many profitable optimizations that are missing from LLVM, some of which have been implemented in LLVM as a result of our work. On an Intel Cascade Lake processor, Minotaur achieves an average speedup of 7.3% on the GNU Multiple Precision library (GMP)’s benchmark suite, with a maximum speedup of 13%. On SPEC CPU 2017, Minotaur produces an average speedup of 1.5%, with a maximum speedup of 4.5% for 638.imagick.

This program is tentative and subject to change.

Thu 24 Oct

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

13:40 - 15:20
Compilers and Optimisation 2OOPSLA 2024 at San Gabriel
Chair(s): Manu Sridharan University of California at Riverside
13:40
20m
Talk
Hydra: Generalizing Peephole Optimizations with Program Synthesis
OOPSLA 2024
Manasij Mukherjee University of Utah, John Regehr University of Utah
DOI Pre-print
14:00
20m
Talk
Minotaur: A SIMD-Oriented Synthesizing SuperoptimizerOOPSLA 2024 Distinguished Paper Award
OOPSLA 2024
Zhengyang Liu University of Utah, Stefan Mada University of Utah, John Regehr University of Utah
DOI
14:20
20m
Talk
PolyJuice: Detecting Mis-Compilation Bugs in Tensor Compilers with Equality Saturation Based Rewriting
OOPSLA 2024
Chijin Zhou Tsinghua University, Bingzhou Qian National University of Defense Technology, Gwihwan Go Tsinghua University, Quan Zhang Tsinghua University, Shanshan Li National University of Defense Technology, Yu Jiang Tsinghua University
DOI
14:40
20m
Talk
SparseAuto: An Auto-Scheduler for Sparse Tensor Computations Using Recursive Loop Nest Restructuring
OOPSLA 2024
Adhitha Dias Purdue University, USA, Logan Anderson Purdue University, Kirshanthan Sundararajah Virginia Tech, Artem Pelenitsyn Purdue University, Milind Kulkarni Purdue University
DOI Pre-print
15:00
20m
Talk
Understanding and Finding Java Decompiler Bugs
OOPSLA 2024
Yifei Lu Nanjing University, Weidong Hou Nanjing University, Minxue Pan Nanjing University, Xuandong Li Nanjing University, Zhendong Su ETH Zurich
DOI