Typed Architecture: Architectural Support for Lightweight Scripting

Abstract: Dynamic scripting languages are becoming more and more
widely adopted not only for fast prototyping but also for developing
production-grade applications. They provide high-productivity
programming environments featuring high levels of abstraction with
powerful built-in functions, automatic memory management,
object-oriented programming paradigm and dynamic typing. However,
their flexible, dynamic type systems easily become the source of
inefficiency in terms of instruction count, memory footprint, and
energy consumption. Besides, the bytecode dispatch loop incurs
significant performance cost due to redundant computation and
hard-to-predict indirect jumps. This overhead makes it challenging to
deploy these high-productivity programming technologies for
production, especially on emerging single-board computers for IoT
applications.

In this talk I will present Typed Architecture, a high-efficiency,
low-cost execution substrate for dynamic scripting languages, where
each data variable retains high-level type information at an ISA
level. Typed Architecture calculates and checks the dynamic type of
each variable implicitly in hardware, rather than explicitly in
software, hence significantly reducing instruction count for dynamic
type checking. To reduce the recurring cost of bytecode dispatch, I
also introduce Short-Circuit Dispatch (SCD), an architectural
extension that enables fast, hardware-based bytecode dispatch with
fewer instructions. Our evaluation using a fully synthesizable RISC-V
RTL design on FPGA demonstrates that Typed Architecture augmented with
SCD achieves geomean speedups of 24.3% and 19.8% with maximum speedups
of 65.0% and 63.5% for two production-grade scripting engines for Lua
and JavaScript, respectively.


Bio:
Jae W. Lee is an associate professor in the Department of Computer
Science and Engineering at Seoul National University (SNU), Korea. His
research areas include computer architecture, compilers, and parallel
computing, and he has co-authored more than 40 research papers in
these areas with 3,600+ citations. His work has been recognized with
two test-of-time awards (IEEE VLSI Symposium 2017, ACM ASPLOS 2014),
and several top/highlights papers (ACM ASPLOS 2017, IEEE PACT 2010,
IEEE/ACM MICRO 2010). Before joining SNU, he was an assistant
professor of Semiconductor Systems Engineering at Sungkyunkwan
University (SKKU), Korea. Prior to that, he was a research associate
at Princeton University, and a researcher and engineer at
Parakinetics, Inc. (Princeton, NJ), where he conducted research on
multicore software optimization. He received his M.S. degree in
Electrical Engineering from Stanford University and Ph.D. degree in
Computer Science from MIT.