Ethan Wu and Viansa Schmulbach

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2023-268

December 12, 2023

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-268.pdf

Numerous works have tackled the problem of offloading deserialization workloads to specialized hardware accelerators, however few target interoperability with “managed” languages which employ a garbage collector and other advanced runtime features. We develop a deserialization accelerator that can directly create “native” Java objects that have no additional software overhead for Java code to interact with, as if they had been created by Java code. In the process, we explore how a hardware accelerator’s memory allocation and object creation functionality can be integrated with the HotSpot Java Virtual Machine, without sacrificing accelerator performance.

---

BibTeX citation:

@techreport{Wu:EECS-2023-268,
    Author= {Wu, Ethan and Schmulbach, Viansa},
    Title= {A Deserialization Architecture for GC Languages},
    Year= {2023},
    Month= {Dec},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-268.html},
    Number= {UCB/EECS-2023-268},
    Abstract= {Numerous works have tackled the problem of offloading deserialization workloads to specialized hardware accelerators, however few target interoperability with “managed” languages which employ a garbage collector and other advanced runtime features. We develop a deserialization accelerator that can directly create “native” Java objects that have no additional software overhead for Java code to interact with, as if they had been created by Java code. In the process, we explore how a hardware accelerator’s memory allocation and object creation functionality can be integrated with the HotSpot Java Virtual Machine, without sacrificing accelerator performance.},
}

EndNote citation:

%0 Report
%A Wu, Ethan 
%A Schmulbach, Viansa 
%T A Deserialization Architecture for GC Languages
%I EECS Department, University of California, Berkeley
%D 2023
%8 December 12
%@ UCB/EECS-2023-268
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2023/EECS-2023-268.html
%F Wu:EECS-2023-268