[Activity] Week 50
Summary: * Validate RM toolchain * Prepare Linaro Toolchain Binaries 2012.12 release. * Collect performance data for different branch cost. Details: 1. Validate RM toolchain release. 2. Collect denbench and spec2000int performance data for branch cost tuning. 3. Fix README issue (lp:1068402) by dynamically create README.txt: * The hard coded README.txt is removed. * Update build scripts to generate the README.txt, which depends on - README.texi, which is the common template for all TARGETS. - config.texi, which sets the default configuration for the TARGET. Each TARGET has a config.texi. - version.texi, created during build, includes GCC_VERSION, GDB_VERSION and TOOLCHAIN_VERSION. 4. Update Linaro GCC and GDB to 2012.12 in Linaro crosstool-ng, workaround gdb mingw32 link fail issue (due to mbsrtowcs is redefined in gdb/gnulib), finish the smoke test and spawn release builds. Plans: * Investigate how branch cost impacts on code generation. * Release Linaro Toolchain Binaries 2012.12. Best regards! -Zhenqiang ___ linaro-toolchain mailing list linaro-toolchain@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-toolchain
RE: Clearing out past branches
This one didn't do as well as expected as it ended up causing more zero extends than necessary if my memory serves me right. This could be a course for a low priority future blueprint. Dropped. This plus flag_ree in the future? I'm not so sure that flag_ree will help AArch32 in the future but that is certainly a good topic for investigation. AFAICS, flag_ree helps with removing unnecessary sign/zero extensions when architectures have a 32 bit register (something like the w* registers on AArch64) written to and a read from the corresponding value zero extended to a 64 bit register ( i.e. the equivalent x register). So I'd expect flag_ree to remove the redundant zero extension between Write to w0 Zero_extend Xn from W0. Read from Xn . Regards Ramana -- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. ___ linaro-toolchain mailing list linaro-toolchain@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-toolchain
[ACTIVITY] 10 - 14 December
== Blueprints == InitialCurrentActual initial-aarch64-backport31 Oct 201231 Dec 2012 aarch64-baremetal-testing 31 Oct 201231 Dec 2012 backport-fma-intrinsic 31 Dec 2012Brice fused-multiply-add-support 31 Dec 2012Brice gcc-investigate-lra-for-arm 31 Dec 201231 Dec 2012 fix-gcc-multiarch-testing 31 Dec 201231 Jan 2013 == Progress == * Admin * Interviewing * Took over from Michael * New Starter prep * Updated card drafts * Discussions about LEG/TCWG interactions * Discussions about KVM * Discussions about Toolchain/Platform interactions == Next Week == * Upload card drafts into Jira. * initial-aarch64-backport and aarch64-baremetal-testing * Complete documentation == Future == * gcc-investigate-lra-for-arm * Analyse benchmarks * Run HOT/COLD partitioning benchmarks * Analyse ARM results * On x86_64 to see what the actual benefit we could get * fix-gcc-multiarch-testing * Come up with strawman proposal for updating testsuite to handle testing with varying command-line options. -- Matthew Gretton-Dann Linaro Toolchain Working Group matthew.gretton-d...@linaro.org ___ linaro-toolchain mailing list linaro-toolchain@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-toolchain
[ACTIVITY] 10 - 14 December 2012
== Progress == * gdb-linaro-7.5-2012.12 release - Updated wikis about gdb sources import from upstream and about gdb-linaro release - still some trouble with infrastructure to achieve the release * 64-bits ops in Neon: no news from upstream * disable-peeling: upstream advice is to enable tune the vectorizer cost model first, but there seems to be an agreement that if unaligned accesses have no penalty, peeling should not be considered. Updated the vectorizer-cost-model blueprint * builtin_bswap16 backport to linaro-4.7: the i686 regression was caused by the build process. Re-spawned the jobs after fixing the remaining arm issue * cbuild: briefly looked at some scripts to better understand the whole system, while tracking a problem in the gdb validation reports == Next == * handle 64-bits bitops in Neon and disable-peeling feedback from upstream if any * finish builtin_bswap16 backport == Future == * on holidays for 2 weeks (Dec 22th-Jan 6th) ___ linaro-toolchain mailing list linaro-toolchain@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-toolchain
