Advanced

Thread-level speculation for web applications

Martinsen, Jan Kasper and Grahn, Håkan (2009) Second Swedish Workshop on Multi-Core Computing p.80-88
Abstract
Thread Level Speculation (TLS) has been suggested as a

mean to automatically (or semi-automatically) extract parallelism

from sequential programs. While there have been

multiple attempts both in hardware and software to implement

real time TLS, to the best of our knowledge all attempts

have so far been on a byte code level or with statically typed

languages.

In this study, we examine the potential of TLS for

Web Applications, using the popular scripting language

JavaScript(JS).We have chosen to execute the programs by

traversing their parse trees, taking advantage of information

from the programming language that are normally lost

... (More)
Thread Level Speculation (TLS) has been suggested as a

mean to automatically (or semi-automatically) extract parallelism

from sequential programs. While there have been

multiple attempts both in hardware and software to implement

real time TLS, to the best of our knowledge all attempts

have so far been on a byte code level or with statically typed

languages.

In this study, we examine the potential of TLS for

Web Applications, using the popular scripting language

JavaScript(JS).We have chosen to execute the programs by

traversing their parse trees, taking advantage of information

from the programming language that are normally lost

when compiled to, e.g., byte code.

We have performed a test where we automatically have

divided the execution of the parsing tree among 1, 2, 4, and

8 cores for four benchmark programs. We have found that

this approach has a small number of rollbacks (i.e. error

correction when speculation fails) and significantly increases

the performance of our benchmarks. (Less)
Please use this url to cite or link to this publication:
author
publishing date
type
Contribution to conference
publication status
published
subject
pages
80 - 88
conference name
Second Swedish Workshop on Multi-Core Computing
project
EASE
language
English
LU publication?
no
id
4730b714-8ec5-47b0-99f7-be75de52504c (old id 1689353)
alternative location
http://www.bth.se/fou/forskinfo.nsf/all/6f17244e8b00c7fec12576e1002122ac/$file/tlswebmcc09_crc_final.pdf
date added to LUP
2010-10-01 11:40:59
date last changed
2016-06-29 09:15:26
@misc{4730b714-8ec5-47b0-99f7-be75de52504c,
  abstract     = {Thread Level Speculation (TLS) has been suggested as a<br/><br>
mean to automatically (or semi-automatically) extract parallelism<br/><br>
from sequential programs. While there have been<br/><br>
multiple attempts both in hardware and software to implement<br/><br>
real time TLS, to the best of our knowledge all attempts<br/><br>
have so far been on a byte code level or with statically typed<br/><br>
languages.<br/><br>
In this study, we examine the potential of TLS for<br/><br>
Web Applications, using the popular scripting language<br/><br>
JavaScript(JS).We have chosen to execute the programs by<br/><br>
traversing their parse trees, taking advantage of information<br/><br>
from the programming language that are normally lost<br/><br>
when compiled to, e.g., byte code.<br/><br>
We have performed a test where we automatically have<br/><br>
divided the execution of the parsing tree among 1, 2, 4, and<br/><br>
8 cores for four benchmark programs. We have found that<br/><br>
this approach has a small number of rollbacks (i.e. error<br/><br>
correction when speculation fails) and significantly increases<br/><br>
the performance of our benchmarks.},
  author       = {Martinsen, Jan Kasper and Grahn, Håkan},
  language     = {eng},
  pages        = {80--88},
  title        = {Thread-level speculation for web applications},
  year         = {2009},
}