Write a Blog >>
VEE 2021
Fri 16 Apr 2021
Fri 16 Apr 2021 10:00 - 10:30 - JIT Compilers and Managed Runtimes Chair(s): Khanh Nguyen

Although Graphics Processing Units (GPUs) have become pervasive for data-parallel workloads, the efficient exploitation of their tiered memory hierarchy requires explicit programming. The efficient utilization of different GPU memory tiers can yield higher performance at the expense of programmability since developers must have extended knowledge of the architectural details in order to utilize them. In this paper, we propose an alternative approach based on Just-In-Time (JIT) compilation to automatically and transparently exploit local memory allocation and data locality on GPUs. In particular, we present a set of compiler extensions that allow arbitrary Java programs to utilize local memory on GPUs without explicit programming. We prototype and evaluate our proposed solution in the context of TornadoVM against a set of benchmarks and GPU architectures, showcasing performance speedups of up to 2.5x compared to equivalent baseline implementations that do not utilize local memory or data locality. In addition, we compare our proposed solution against hand-written optimized OpenCL code to assess the upper bound of performance improvements that can be transparently achieved by JIT compilation without trading programmability. The results showcase that the proposed extensions can achieve up to 94% of the performance of the native code, highlighting the efficiency of the generated code.

Fri 16 Apr

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

10:00 - 11:30
JIT Compilers and Managed RuntimesResearch Papers
Chair(s): Khanh Nguyen Texas A&M University
10:00
30m
Talk
Automatically Exploiting the Memory Hierarchy of GPUs through Just-in-Time Compilation
Research Papers
Michail Papadimitriou University of Manchester, UK, Juan Fumero University of Manchester, UK, Athanasios Stratikopoulos The University of Manchester, Christos Kotselidis KTM Innovation / The University of Manchester
DOI
10:30
30m
Talk
Multiple-Tasks on Multiple-Devices (MTMD): Exploiting Concurrency in Heterogeneous Managed Runtimes
Research Papers
Michail Papadimitriou University of Manchester, UK, Eleni Markou , Juan Fumero University of Manchester, UK, Athanasios Stratikopoulos The University of Manchester, Florin Blanaru , Christos Kotselidis KTM Innovation / The University of Manchester
DOI
11:00
30m
Talk
Automated Bug Localization in JIT Compilers
Research Papers
Heuichan Lim , Saumya Debray University of Arizona
DOI