Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.compilers > #3557 > unrolled thread

Array accesses are boring

Started byJohn R Levine <johnl@taugh.com>
First post2024-03-06 11:42 -0400
Last post2024-03-15 21:50 +0100
Articles 2 — 2 participants

Back to article view | Back to comp.compilers


Contents

  Array accesses are boring John R Levine <johnl@taugh.com> - 2024-03-06 11:42 -0400
    Re: Array accesses are boring Florian Weimer <fw@deneb.enyo.de> - 2024-03-15 21:50 +0100

#3557 — Array accesses are boring

FromJohn R Levine <johnl@taugh.com>
Date2024-03-06 11:42 -0400
SubjectArray accesses are boring
Message-ID<24-03-001@comp.compilers>
They instrumented a large set of Java benchmarks to look at the array
accesses, over 3 billion of them in total.  The majority fell into five
simple access patterns.  See the pictures on page 8 of the PDF.

https://arxiv.org/abs/2403.02416

Regards,
John Levine, johnl@taugh.com, Taughannock Networks, Trumansburg NY
Please consider the environment before reading this e-mail. https://jl.ly

[toc] | [next] | [standalone]


#3559

FromFlorian Weimer <fw@deneb.enyo.de>
Date2024-03-15 21:50 +0100
Message-ID<24-03-003@comp.compilers>
In reply to#3557
* John R. Levine:

> They instrumented a large set of Java benchmarks to look at the array
> accesses, over 3 billion of them in total.  The majority fell into five
> simple access patterns.  See the pictures on page 8 of the PDF.
>
> https://arxiv.org/abs/2403.02416

It's strange that they observed array lengths changing at the JVM
bytecode level.  That shouldn't happen.
[Different sizes on different calls to the routine? -John]

[toc] | [prev] | [standalone]


Back to top | Article view | comp.compilers


csiph-web