Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Optimization: Opportunities identified in Octane benchmark #25

Open
11 of 18 tasks
john-yan opened this issue Dec 2, 2016 · 10 comments
Open
11 of 18 tasks

Optimization: Opportunities identified in Octane benchmark #25

john-yan opened this issue Dec 2, 2016 · 10 comments
Labels
enhancement

Comments

@john-yan
Copy link

john-yan commented Dec 2, 2016 •
edited by jbajwa
Loading

Common:

  • Create a normalized crankshaft opcode distrubution
  • Optimize Allocation sequences to use prefetch - @joransiu

PPC:

  • Use vector instruction to optimize ArithmeticD. Can use non-clobbered mode @jbajwa
  • optimize atomic load/store sequence @jbajwa
  • use OV32/CA32 for overflow checking (applies to add/sub/mul instructions) @jbajwa
  • exploit MAD (multiply and add) instruction
  • exploit mod[s|u][d|w] (modulo) instruction @jbajwa

S390:

  • Use vector instruction to optimize ArithmeticD. Can use non-clobbered mode @john-yan efd92a1
  • check-non-smi can use TM to reduce gap @sampsongao
  • MathAbs can use load positive to optimize @sampsongao
  • MacroAssembler::LoadRepresentation: remove lgbr and lghr
  • BoundsCheck length can be improved by using UseAtStart
  • MathSqrt can use sqdb memory version @john-yan
  • Better optimize add r,r,m. (1% of AddI)
  • LoadDoubleLiteral can use lzdr to optimize load 0.0 @john-yan
  • CmpSmiLiteral can use CIH for 64-bit @john-yan
  • Add/SubSmiLiternal can use AIH for 64-bit @john-yan
  • Remove redundant LGFRs in crankshaft codegen. @john-yan
@jbajwa
Copy link

jbajwa commented Dec 20, 2016 •
edited
Loading

PPC: optimize atomic load/store sequence
https://codereview.chromium.org/2590543003/

@sampsongao
Copy link

S390: check-non-smi can use TM to reduce gap
https://codereview.chromium.org/2582973002/
https://codereview.chromium.org/2588163002/

@john-yan
Copy link
Author

CmpSmiLiteral can use CIH for 64-bit
Add/SubSmiLiternal can use AIH for 64-bit
https://codereview.chromium.org/2593803003/

@sampsongao
Copy link

MathAbs can use load positive to optimize
https://codereview.chromium.org/2591643005/

@joransiu
Copy link
Member

S390: Optimize Allocation sequences to use prefetch
https://codereview.chromium.org/2601563002/

@john-yan
Copy link
Author

john-yan commented Jan 3, 2017

https://codereview.chromium.org/2611773003/

@john-yan
Copy link
Author

john-yan commented Jan 5, 2017

https://codereview.chromium.org/2615683003/

@john-yan
Copy link
Author

john-yan commented Jan 5, 2017

https://codereview.chromium.org/2617733003/

@jbajwa
Copy link

jbajwa commented Jan 13, 2017

PPC: Use vector instruction to optimize ArithmeticD. Can use non-clobbered mode
https://codereview.chromium.org/2619763003/

@jbajwa
Copy link

jbajwa commented Jan 13, 2017

PPC: exploit mod[s|u][d|w] (modulo) instruction
https://codereview.chromium.org/2625013002/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@jbajwa @john-yan @joransiu @sampsongao