Thursday 2nd August 2012

This month's meetings

WorkingGroups/ToolChain/Meetings
<< <  2012 / 8 >  >>
Mon Tue Wed Thu Fri Sat Sun
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    

Attendees

  • Michael Hope
  • Peter Maydell
  • Ramana Radhakrishnan
  • Richard Earnshaw
  • Ulrich Weigand
  • Zhenqiang Chen

Agenda

  • Stand up call

Minutes

Peter:

  • Upstream review
  • Fallout from configure related patches
  • KVM interrupt related patches has been acked
  • Next is RFC set, needs tidy up first

Zhenqiang:

  • Investigating register allocation for Thumb-1
  • IP register is available and allocated, reload picks another register to replace it
  • Collected more detail on movw/movt
  • Split after eload as the best result
  • Looked at vpr. Splitter 1 gives more spilling, after reload gives less spilling

Ramana:

  • Getting the NEON intrinsic patches upstream
  • Looking at errors in the NEON testsuite
  • Issues in how regexes are made with scan assembler
  • Reviewing Carrot's patches with immediate for some operations

Ulrich:

  • Checked in patch to change vector alignment to 8 bytes
  • Needed to check vector alignment reachable
  • Seems to still be correct, so patch is good
  • Has results for the vector align branch, eon improves, wupwise regresses
  • Reload/aarch64 bug: has a patch and cleanup, haven't heard back yet
  • Working on the GDB 7.5 release
  • Committed three patches covering some of the watchpoint, remote mode failures
  • Release probably spins on Friday, time for a couple more patches
  • Most of the new failures are tests that cover pre-existing bugs

Michael:

  • Release branch merges
  • Created a 'Merge process' page for GCC
  • Fix prettyprinter upstream to enable i686 parallel builds
  • Worked on the tasks and hiring plan
  • Signed up for Plumbers
  • Updated the Android executable note in AOSP
  • Ran benchmarks for NEON shifts, vfp core transfer, and vector alignment changes
  • Changed the benchmark systems to reboot to clear outaccumulated (and repeatable!) slowdowns

WorkingGroups/ToolChain/Meetings/Archive/2012-08-02 (last modified 2013-08-30 11:48:00)