Monday 30th May 2011
This month's meetings
<< < 2011 / 5 > >>
- Review action items from last meeting
- Walled gardens
- Are we working as part of mainline?
- Are we contributing to mainline at the right level? Patches? Test results? Builds? Bug reports? Triage? Review?
- Does mainline know we exist?
- ...that we have consolidation branches?
- ...and do they know the purpose and contents?
- Do they feel they can get involved with the consolidation branches?
- Exception table sizes (Ken)
- SPEC libc write up (Dave)
771900 - profiled bootstrap
721531 - testing the Thumb bit on function pointers
Action Items from this Meeting
Action Items from Previous Meeting
(The UK has a bank holiday) (Wednesday is a German holiday)
- Will leave until we have the full group
- How does everyone use it?
- IBM has their own version
- Ira: builds with with cross, copy it to device, run with -autogo, use default iterations
- Revital: used the Linaro number of iterations
- For Telecom, didn't look into it, use defaults or something similar
- Chung-Lin: EEMBC is run using a automated system, using standard configuration
- Don't need to standardise running, as long as the reported is standardised
- Licensing problems kick in
Ken linked off to his build results on the wiki KenWerner/Sandbox/libunwind#overhead_of_the_ARM_specific_unwind-tables
- Shows a text size and file size increase
- Numbers are always under 10 %, normally under 5 %
- Some cases the text size got smaller...
- Could be build system related, is pretty confident that Firefox was built correctly though
- Would like to add eh frame information to see the upper limit
Working around issues using GCC version numbers:
- Discussion has gone upstream
- We'll see how it plays out
Is anyone going to the GCC London Google unconference?
- Will ask again later in the week
WorkingGroups/ToolChain/Meetings/Archive/2011-05-30 (last modified 2013-08-30 11:49:16)