== Progress==
TCWG-435 needless busy-wait in lowlevellock.c (3/10)
* Patches ready, testing a bit slow & fiddly
TCWG-156 cortex-strings memset (3/10)
* Worried about noisy benchmarks
* Learned to use Lava
* Found some helpful internal docs
* Managed to tweak the code slightly
== Misc ==
Meetings/mail/etc 2/10
(Public) holiday 2/10
== Plan ==
Stop worrying and follow the general curve of the noisy benchmarks
(But also try to convince Maxim's spec scripts + lava to generate some
less noisy results)
Another public holiday next Monday
== Progress ==
* GDB reverse debugging on aarch64
-- Complete decoding of aarch64 data processing immediate instructions.
[TCWG-399] [1/10]
-- Complete decoding of aarch64 data processing register instructions.
[TCWG-402] [1/10]
-- Completed decoding of aarch64 exception and system instructions.
[TCWG-400] [2/10]
-- Further progress on decoding of aarch64 load store instructions.
[TCWG-401] [2/10]
-- Started implementation of aarch64 syscall record/replay. [TCWG-409]
[2/10]
* Sick Day Off on Monday [2/10]
== Plan ==
* GDB reverse debugging on aarch64
-- Further progress on decoding of aarch64 load store instructions.
[TCWG-401]
-- Further progress on aarch64 syscall record/replay. [TCWG-409]
* Public Holiday on 1st May.
Short week, Easter Monday + child care (3/10)
== Issues ==
* Toolchain64 disk still full every 2 days
== Progress ==
* Linaro GCC 4.9 2014.04 release (3/10)
- Branch merge with FSF 4.9.1
- Release tarball available on releases.linaro.org
- Announcement will follow
* Launchpad bugs: (2/10)
o LP #1169164 : including signal.h exposes various PSR_MODE #defines
- Discussed and implemented a fix
- Validation on-going
* Misc:
o Cbuildv1 baby-sitting (1/10)
o Various meetings (1/10).
o Resolved cards :
- TCWG-343. Make LRA the default for the ARM backend
- TCWG-422. LP-bug 1296676 : ICE in assign_by_spills building linux
btrfs module
== Next ==
- One day off (Labor day)
- Submit patch for LP #1169164
- TCWG-345. Analyse performance of LRA for ARM
== Progress ==
PGO - AArch64 (TCWG-179) (4/10)
* Completed SPEC2006 runs ( -O3 + PGO) in chroot + qemu-arm64-saucy
with Linaro branch(4.9).
* Perlbench train run failed, issue with qemu.
* DealII train runs failed due to system libstdc++.so.6 not compatible
with GCC 4.9.
Changing the LD_LIBRARY_PATH to point to the libstdc++.so.6 that
gets built along with GCC 4.9 solves the problem.
* CPU2000 runs are completed my maxim so functional testing of PGO for
spec benchmarks completed.
* Wating on Arm64 hardware to benchmark SPEC2000 and SPEC2006 and
compare against the PGO runs for x86.
Misc (2/10)
* Setup arm64-trusty chroot and built GCC 4.9 compiler
* Maxim 1-1 discussions and set up opennx client
* Checked libssp patch upstream status
* Read about gcc debug counters.
* Upgrade laptop to Ubuntu 14.04 LTS
Short week (22nd and 23rd leave (4/10))
== Plan ==
* Bug fixing.
== Issues==
* Waiting on hardware to Benchmark SPEC2006 PGO runs in hardware.
== Issue ==
* None
== Progress ==
* More tests on shrink-wrap changes (TCWG-133, 6/10).
- Cortex-m3 tests exposes a data flow issue: New BB created after
dfinit does not have correct df_lr info.
- Collect Spec2k benchmarks on X86-64 and ARM.
* Linaro 4.9 binaries release (2/10).
- Update Linaro crosstool-ng config and samples to support 4.9.
- Disable parallel build for gcc manual (pdf & html).
- Prebuild binaries for aarch64:
http://cbuild.validation.linaro.org/binaries/4.9-prerelease-2014.04
* Ping aarch64 fcel patch. But still no comments.
* Fix a trunk build fail issue @r209556.
* One day off (2/10).
== Plans ==
* Send the shrink-wrap related patches for review.
* PING the pending patches.
* Investigate move-loop-invariants heuristics.
== Planed leaves ==
* May 1-3: Labour day holiday.
== Progress ==
* Got Jenkins working with matrix builds so we can utilize all LAVA
slaves. (TCWG 1387 - 3/10)
* Write script to take a list of revisions and then build and test
them all and diff the results. (TCWG 448 - 4/10)
* More experimenting with Kugan's benchmarking branch. (1/10)
* Meetings and Misc (2/10)
== Plan ==
* More work on matrix builds. (TCWG 1387)
* More work on regression test analysis and reporting. (TCWG 448)
== Progress ==
* Short week (21st and 25th are public holidays) (4/10)
* TCWG-447 (5/10)
* Implemented and tested fenv target hooks, necessary built-ins and md
patterns
* Posted RFC patches for review for both arm and aarch64
* http://gcc.gnu.org/ml/gcc-patches/2014-04/msg01743.html
* http://gcc.gnu.org/ml/gcc-patches/2014-04/msg01744.html
* TCWG-413 Spec2006 (1/10)
* Finished the set-up
* On hold for now
== Plan ==
* upstream zero/sign extension elimination activities
* start with literal pool merging
== Progress ==
* Holidays (3 days)
- Clearing emails/tasks backlog
- Some post-trip illness
* AArch64 vs. ARM64
- Comparing performance of both back-ends
* Named Register
- Re-implementing after code review
- http://reviews.llvm.org/D3261
* Time
- CARD-1246 4/10
- Others 6/10
== Plan ==
* Finish named register in LLVM, check Clang
* Continue testing and benchmarking ARM64 back-end
* Have a try at CBuildv2
== Progress ==
* Back to work part-time (50%) for the next 3 weeks
* Easter Monday (1/5)
* (No Jira card - 2/5)
Analyzed at GCC trunk validations and reported some regressions.
The current commit rate is very high since stage1, and our compute
farm isn't fast enough. Should be OK after the week-end.
Modified my scripts so that I move these validations out of Jenkins.
* (No Jira card - 2/5)
Continued converting Neon intrinsics tests for inclusion in GCC testsuite.
To make the review easier, I now believe that it would be better that
I convert most of the testsuite first, and rework the patches order.
* Meetings and conf calls(1/5)
== Next ==
* Continue to closely monitor GCC trunk validations
* Neon intrinsics tests
* Off Thursday/Friday next week
=== Previous ===
libvpx NEON intrinsics investigation [TCWG-429 14/10]
. if interested, you can read the report in card 429:
https://cards.linaro.org/browse/TCWG-429?focusedCommentId=22247
tested division optimisation patches against current trunk and pinged
on list [2/10]
good friday bank holiday [2/10]
=== Next ===
Resume looking at NEON scheduling TCWG-135, which ties into libvpx
Upgrade laptop to Ubuntu 14.04 LTS
== Progress ==
* GDB reverse debugging on aarch64
-- Scored through a64 instruction set details. [1/10]
-- Completed testing and implementation of all basic structures.
[TCWG-398] [1/10]
-- Started decoding aarch64 load store instructions for recording.
[TCWG-401] [2/10]
-- Completed decoding of aarch64 branch instructions. [TCWG-400] [2/10]
-- Scored through a32 instruction set and researched on a32 recording.
[2/10]
* Miscellaneous [2/10]
-- UK visa application process
-- Tools cauldron and TCWG sprint bookings etc.
-- Meetings
== Plan ==
* GDB reverse debugging on aarch64
-- Complete decoding of aarch64 load store instructions. [TCWG-401]
-- Completed decoding of aarch64 exception and system instructions.
[TCWG-400]
== Progress ==
PGO - AArch64 (TCWG-179) (4/10)
* Completed SPEC runs -O3 -mcpu=cortex-a57 in chroot + qemu
arm64-saucy with Linaro release source (4.8) march 2014.
* Built Linaro 4.9 GCC branch based tool chain under chroot + qemu arm64-saucy
* Created a PGO config file for aarch64 as a peak configure and use --tune=peak
SPEC gave warning that FDO cannot be used as base flags and will
ignore the flags.
Also setting basepeak = yes gave issues and ran base runs and
setting it to no does both base and peak runs
* Started SPEC2006 runs for PGO under chroot + qemu arm64-saucy
* Perlbench train run failed in qemu.
* Other benchmarks are progressing.
GLIBC Systemtap (2/10)
* Ran glibc make check under qemu. Seen some illegal instructions and
ntpl tests hung in qemu aarch64. Will has completed these tests in
hardware.
Misc (2/10)
* Tested cbuildv2 to build gcc natively under qemu-arm64-chroot-ubuntu-saucy.
git gave issues not able to clone and hangs when downloading
binutils and gdb repos.
* 17th was Public Holiday (2/10)
== Plan ==
* Go through the results for SPEC2006 PGO runs.
* Start functional PGO runs for SPEC2000
* Bug fixing.
== Issues ==
* Toolchain64 disk full
== Progress ==
* Linaro GCC 4.9 2014.04 (3/10)
- Created FSF Linaro 4.9 branch
- Testing FSF 4.9.0 RC
* Launchpad bugs: (3/10)
o LP #1169164 : including signal.h exposes various PSR_MODE #defines
- Two possible ways to fix the issue (to be discussed with maintainers)
* Misc:
o Cbuildv1 baby-sitting (1/10)
o Various meetings (1/10).
== Next ==
- Easter Monday off
- Continue on LP #1169164
- Linaro GCC 4.9 release (when FSF release will be made)
== Week of April 14th ==
- Benchmarked autoprefetcher patches. (TCWG-388, 2/10)
- Continued working on spec2xxx-utils scripts. These are now gaining support for AArch64 and SPEC2006. (TCWG-238, 4/10)
- Various discussions. (2/10)
- Short week due to Good Friday / Easter. (2/10)
== Week of April 21st ==
- Take Tyler B.'s LAVA test definition for SPEC benchmarking and adjust spec2xxx-utils scripts to match LAVA environment.
- Investigate regressions on gap, eon and wupwise from autoprefetcher patches.
-- SPEC2000's gap in particular is sensitive to 1st scheduling pass with register pressure, and, apparently has a 5% performance improvement potential (at -O2 Cortex-A15).
- Various discussions.
- Even shorter week due to Easter Monday and ANZAC Day on Friday. (4/10)
--
Maxim Kuvyrkov
www.linaro.org
== Progress ==
* TCWG-238 (4/10)
- Created scripts and spectools for spec2006 to work with cbuild2
* TCWG-413 Spec2006 (4/10)
- Worked out Spec2006 config and src.alt for v1.1 src we have to work
correctly
- Got it to work natively with Maxim's scripts
- Did a trail run in apm openembedded; Needs ccrypt, tar with xz
compression support and system binutils with -mabi=ilp64 support
* 18th is Public Holiday (2/10)
== Plan ==
Benchmarking and FENV support
== Progress ==
* Attempted to get qemu-aarch64 working, fails on saucy, works on
precise. Did native build, hung during 'make check'. Setup
qemu-aarch64 chroot on all TCWG x86_64 build farm
machines. (4/10)
* Tracked down problem that's kept TCWG Jenkins from
working. (1/10)
* Tried Kugan's benchmarking branch of cbuildv2. (1/10)
* More work on Neon intrinsics tests. (TCWG-322 - 1/10)
* Meetings and Misc (3/10)
- Worked on benchmarking doc with Ryan.
== Plan ==
* Get Jenkins working with matrix builds so we can utilize all LAVA
slaves.
* More experimenting with Kugan's benchmarking branch.
== Issue ==
* None.
== Progress ==
* multilib can not work with multiarch anymore, so change Linaro
crosstool-ng to make 2014.04 release (2/10).
- For 4.8 release, we will revert the change and keep it align with
previous releases.
* Linaro 4.9 toolchain binaries pre-releases (Cross & native) for
aarch64 (2/10):
http://cbuild.validation.linaro.org/binaries/4.9-prerelease-2014.04.
* Try to set up native spec2006 tests on Chromebook. But got "out of
memory" when building the tools (1/10).
* Clean up and test for shrink-wrap enhancement (TCWG-133: 2/10)
* Rebase the aarch64 fcsel patch and testing. (1/10)
* ARM internal work (2/10)
== Plans ==
* Send the shrink-wrap related patches for review.
* PING the pending patches.
== Planed leaves ==
* April 21: Team event.
* May 1-3: Labour day holiday.
== This week ==
- Resolved Launchpad bug 1305042 as user error and provided proper asm
coding [TCWG-290] [2/10]
- Investigated infinite loop bug [TCWG-290][6/10]
- Determined the infinite loop is occurring in function
vt_find_locations in the variable tracking pass
- Still debugging to determine cause of infinite loop
- Absent due to illness on April 14th
== Next week ==
Vacation
== Future ==
Getting the following cross compiling for RPi on Debian Wheezy x64 using
the gcc-linaro-arm-linux-gnueabihf-raspbian toolchain within Eclipse
Kepler. Compilation works fine using native Debian x64 GCC and native
GCC compiler on the RPi.
out of dynamic memory in yy_create_buffer()
collect2: error: ld returned 2 exit status
make: * [cwebsocket] Error 1
I'm stumped. Can't find much info on the issue. Any ideas?
== Progress ==
* One bank holiday, half day childcare (3/10)
* Submit two iterations of malloc microbenchmark for glibc (4/10, TCWG-160)
* Submitted a patch for aarch64 ld bug with SystemTap notes (1/10)
* Investigated strcmp implementations for ARM (1/10, TCWG-153)
* Patch review, bugs, support, etc. (1/10)
== Issues ==
* None
== Plan ==
* malloc benchmarking/implementation work
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
TCWG-156 cortex-strings memset (3/10)
* Got a full set of benchmarks (for my 2 targets)
* Cleaned up code
* Sped up small memsets (for A9, A15 results pending)
glibc performance bug in lowlevellock.c (1/10 - I'll make a card next week)
* Learned to build and test glibc
* Understood the bug and worked out how to fix
Misc
* Meetings/mail (1/10)
* Infrastructure/workflow fiddling (1/10)
* 2 days holiday
== Plan ==
More TCWG-156
Finish off glibc performance bug
Public holidays Friday & Monday
== Week of April 7th ==
- Made a how-to wiki page/script on how to use QEMU and schroot for system emulation (TCWG-179, 2/10).
-- This allows you to get armhf or armel or aarch64 Ubuntu or Debian system within minutes, e.g., for PGO or LTO bootstraps of GCC.
-- https://collaborate.linaro.org/display/TCWG/ARM+ubuntu+chroot+with+QEMU+use…
-- I can move this page into public wiki if get enough interest from non-Linaro people.
- Fixed various bugs in instruction scheduling patch set (TCWG-388, 5/10).
- Scripted running SPEC2000 for reproducible benchmarking (TCWG-238, 3/10).
-- The scripts are now at http://git.linaro.org/toolchain/spec2xxx-utils.git
-- Scripts support both local (native) and remote (cross) benchmarking.
== Week of April 14th ==
- Document spec2xxx-utils scripts and see how they fare in LAVA environment. Add support for SPEC2006. (TCWG-238)
- Post autoprefetcher patch set upstream (TCWG-388)
--
Maxim Kuvyrkov
www.linaro.org
== Week of March 31st ==
- Investigated and scripted how to reproducibly run benchmarks (TCWG-238, 7/10)
-- Stop all services on target board
-- Use local disk
-- Stop networking
-- Bind benchmark to a specific CPU. Bind all other processes to a different CPU.
-- Disable frequency scaling.
- Various discussions (3/10)
--
Maxim Kuvyrkov
www.linaro.org
PGO - AArch64 (TCWG-179) (4/10)
* Installed QEMU user static for aarch64 and use them from
chroot environment in ubuntiu 13.10
* Tried installing chroot + qemu arm64-saucy in another ubuntu 12.04
machine. Issue with binfmts. looks to be corrupted.
* Bootstrap GCC with PGO with --enable-languages=c,c++,fortran
completed under arm64-chroot.
* Built GCC and binutils in native mode under chroot + qemu arm64-saucy .
* SPEC runs -O3 -mcpu=cortex-a57 in chroot + qemu arm64-saucy with
linaro release source (4.8).
Calculix did not finish under chroot + qemu arm64-saucy
GLIBC Systemtap (4/10)
* Built systemtap under chroot + qemu arm64-saucy. Built glibc with
the patch and used --enable-systemtap.
* Able to see setjmp probe in gdb, but longjmp probes not loading.
Tried debugging in gdb. gdb debugging support not available yet under
qemu.
* Discussed with Will newton on my observation.
Bug fix (1/10)
* Not much progress continued looking at reload dumps.
Misc (1/10)
* Tested cbuildv2 to build gcc natively under qemu-arm64-chroot-ubuntu-saucy.
cbuildv2 reports configure error for gcc and g++multilibs. But they
are not needed for native builds. Sent note to Rob for fixing it.
* Attend 1-1 with Ryan
* Attend 1-1 with Maxim discuss PGO work.
== Plan ==
PGO runs for SPEC2006 on chroot + qemu arm64-saucy
Upstream /testing for systemtap probes in glibc
Reload bug fix.
Leave: 17-Apr local state election.
