Your SlideShare is downloading. ×
  • Like
LCA13: GCC Performance
Upcoming SlideShare
Loading in...5

Thanks for flagging this SlideShare!

Oops! An error has occurred.


Now you can save presentations on your phone or tablet

Available for both IPhone and Android

Text the download link to your phone

Standard text messaging rates apply

LCA13: GCC Performance


Resource: LCA13 …

Resource: LCA13
Name: GCC Performance
Date: 04-03-2013
Speaker: Matthew Gretton-Dann

Published in Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads


Total Views
On SlideShare
From Embeds
Number of Embeds



Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

    No notes for slide


  • 1. ASIA 2013 (LCA13) Toolchain Working Group GCC Update Matthew Gretton-Dann Technical Lead - Toolchain Working Group
  • 2. ASIA 2013 (LCA13) Agenda Who are we? Purpose of session What we've done since LCE'12 AArch64 Enablement ARMv7-A Performance What we're going to do next AArch64 Enablement ARMv7-A Performance Discussions on Performance issues
  • 3. ASIA 2013 (LCA13) Who are We?
  • 4. ASIA 2013 (LCA13) Purpose of Session Report on what we've done Start discussions on what we're going to do next iteration These are ARMv7-A Performance focused These will carry on into hacking sessions Session on Friday will wrap-up these discussions and deliver tasks for the next quarter. What about AArch64? We'll do enablement work as necessary We won't focus on AArch64 for performance as we have no hardware. However, we expect most ARMv7-A performance work to also have some benefits for AArch64.
  • 5. ASIA 2013 (LCA13) What We've Done AArch64 Enablement Boehm GC & Libatomic GC Sections ARMv7-A Performance Improvements Vectorizer Cost Model Branch Cost Investigations Improved support for REV intrinsics Improved core to Neon register transfers Better 64-bit shifts in Neon Improved costs model for 64-bit ops in Neon vs Core registers
  • 6. ASIA 2013 (LCA13) What We're Going To Do – A Plea What matters to you? Give us your poor code!
  • 7. ASIA 2013 (LCA13) What We're Going To Do ARMv7-A Performance Disable Peeling for alignment in the Vectorizer Improve Neon Intrinsic code generation Jump Threading Remove redundant zero/sign extensions Shrinkwrapping Improve model for Branches vs Conditional Execution
  • 8. More about Linaro Connect: More about Linaro: More about Linaro engineering: ASIA 2013 (LCA13)