Improve the auto increment/decrement pass

Registered by Michael Hope on 2011-05-26

Improve GCC's auto increment/decrement pass, with particular emphasis on NEON loads and stores.

Status:

- Requires a change to the Cortex A9 scheduling description that can only really be provided by ARM employees.

- Current patches posted here:

    http://lists.linaro.org/pipermail/linaro-toolchain/2011-December/001944.html
    http://lists.linaro.org/pipermail/linaro-toolchain/2011-December/001945.html

Blueprint information

Status:
Complete
Approver:
Michael Hope
Priority:
Medium
Drafter:
Michael Hope
Direction:
Needs approval
Assignee:
Ramana Radhakrishnan
Definition:
Approved
Series goal:
Accepted for 4.7
Implementation:
Informational Informational
Milestone target:
None
Started by
Michael Hope on 2011-07-15
Completed by
Matthew Gretton-Dann on 2013-05-22

Related branches

Sprints

Whiteboard

Blocked on post inc address writeback pipeline description.

[2011-12-21] michaelh1 took this from Richard and put it in the backlog

[2013-05-22 matthew-gretton-dann] This made no difference to performance and there is no point in modelling this in the pipeline. Not being moved into a Jira task

(?)

Work Items

Work items:
Discuss with upstream: DONE
Investigate and design improvement: DONE
[ramana] Model pre/post increment/decrement in the pipeline: TODO
Verify on core only code: INPROGRESS
Verify on vectorised/NEON code: INPROGRESS
Upstream and backport: INPROGRESS

Dependency tree

* Blueprints in grey have been implemented.

This blueprint contains Public information 
Everyone can see this information.