Generic tuning for all Cortex-A devices

Registered by Michael Hope

The TSC have asked for a GCC optimisation option that gives code that performs well across all of the Cortex-A series. Discuss what that means, how to balance across the targets, how to present it to the user, approach for implementing, and how to qualify the results. Is it worth discussing adding -mcpu=native at the same time?

Blueprint information

Status:
Complete
Approver:
Michael Hope
Priority:
Medium
Drafter:
Ramana Radhakrishnan
Direction:
Needs approval
Assignee:
None
Definition:
Approved
Series goal:
Accepted for 4.6
Implementation:
Informational Informational
Milestone target:
milestone icon backlog
Started by
Matthew Gretton-Dann
Completed by
Matthew Gretton-Dann

Related branches

Sprints

Whiteboard

Work Items (Generic tuning):
Identify best 'generic' cost model: TODO
Define interface (-march, -mcpu, -mtune): TODO
Implement prototype: TODO
Benchmark comparisons: TODO
Discuss upstream: TODO
Final implementation: TODO
Commit upstream: TODO
Commit to Linaro 4.6: TODO
Backport to Linaro 4.5: TODO

Note that /generic/ may mean do best on the A9, do well on others, avoid anything that is terrible on other cores.

michaelh split out native tuning into a new blueprint and put this back into the backlog. Some work has been done but not enough to warrant being 'Started'.

[2013-05-22 matthew-gretton-dann] This blueprint is too broad, and should be a roadmap card. With the move to Jira this will be carded up.

(?)

Work Items

Dependency tree

* Blueprints in grey have been implemented.

This blueprint contains Public information 
Everyone can see this information.