Re: GCC 10.1 Release Candidate available from gcc.gnu.org
On Tue, May 5, 2020 at 7:57 AM William Seurer via Gcc wrote: > > I tried it on powerpc64 LE on power 8 and 9 and BE on power 7 and 8 and > all looks well. > > On 4/30/20 4:21 PM, Jakub Jelinek via Gcc wrote: > > The first release candidate for GCC 10.1 is available from > > > > https://gcc.gnu.org/pub/gcc/snapshots/10.1.0-RC-20200430/ > > ftp://gcc.gnu.org/pub/gcc/snapshots/10.1.0-RC-20200430 > > > > and shortly its mirrors. It has been generated from git revision > > r10-8080-g591d857164c37cd0bb96da2a293148e01f280e0f. > > > > I have so far bootstrapped and tested the release candidate on > > x86_64-linux and i686-linux. Please test it and report any issues to > > bugzilla. > > > > If all goes well, I'd like to release 10.1 on Thursday, May 7th. > > Thanks for this I tried on an Intel i7-4700MQ CPU and all works fine I will recompile some minimal packages to build a Linux system (kernel , systemd, glibc and others) Has someone found issues on common packages that require patches for GCC 10? Thanks Victor Rodriguez
Re: Question about indirect functions and PGO
On Mon, Jul 13, 2020 at 6:41 AM Erick Ochoa wrote: > > Hi, > > I just wanted to answer myself. > It seems that there are two thresholds that need to be met if a function > is to be specialized within a particular context: > > 1. --param=hot-bb-count-ws-permille=. This option controls the hotness > threshold of basic blocks and is needed for function specialization > during LTO. If a callsite is not in a hot basic block, it seems that the > callsite will not be specialized for a particular parameter. > 2. --param=ipa-cp-eval-threshold=. This option controls a heuristic that > lets constant propagation happen if a function is a good candidate for > cloning. This parameter is used for both: function specialization within > a particular context and for all contexts. > > On 10/07/2020 13:19, Erick Ochoa wrote: > > Forgot to mention that these functions take a function pointer as a > > parameter and as a result, the specialized functions are able to replace > > the indirect function call with a direct function call. > > > > On 10/07/2020 13:17, Erick Ochoa wrote: > >> Hello, > >> > >> I'm working on an optimization and I encountered this interesting > >> behaviour. There are a couple of functions that are specialized when > >> the program is not compiled with PGO (-fprofile-generate and > >> -fprofile-use) > >> > >> However, when the program is compiled with PGO the compiler does not > >> specialize the function calls. > >> > >> I printing the program just after materializing all clones. > >> > >> I am running this version of GCC: > >> Author: GCC Administrator > >> Date: Fri Jul 10 00:16:28 2020 + > >> > >> Daily bump. > >> > >> I can imagine that the profiling information was used to determine > >> that specializing these functions is a bad tradeoff between binary > >> size and speed. But I do not know this for sure. How can I find out > >> why these functions were not specialized? (I.e. is there a threshold > >> that wasn't met, and if so, where is it located and what's its value?) > >> Have you tried the new option-fprofile-partial-training ? -fprofile-partial-training can now be used to inform the compiler that code paths not covered by the training run should not be optimized for size. I tested by myself this week and it literally do not optimize at all the functions not touched by the training Regards Victor Rodriguez > >> Thanks!
Lack of latest binutils in gcc 11 docker image
HI team I am trying to use the gcc 11 docker hub image to build AMX code but it fails /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h: Assembler messages: /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h:42: Error: no such instruction: `ldtilecfg (%rdi)' /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h:49: Error: no such instruction: `sttilecfg -72(%rsp)' /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h:42: Error: no such instruction: `ldtilecfg -72(%rsp)' src/basic/test.c:11: Error: no such instruction: `tdpbssd %tmm3,%tmm2,%tmm1' /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h:56: Error: no such instruction: `tilerelease' src/basic/test.c:11: Error: no such instruction: `tdpbssd %tmm3,%tmm2,%tmm1' /usr/local/lib/gcc/x86_64-linux-gnu/11.1.0/include/amxtileintrin.h:56: Error: no such instruction: `tilerelease' I was wondering where can we send a patch to add binutils (latest ) to that docker hub image Thanks https://hub.docker.com/layers/gcc/library/gcc/11.1.0/images/sha256-1b844ca17da259215eb25ad8b3cdd7a547ff9e482a5b8477606ab028203717f4?context=explore
Re: Lack of latest binutils in gcc 11 docker image
On Fri, Jun 11, 2021 at 4:39 AM Jonathan Wakely wrote: > > > I was wondering where can we send a patch to add binutils (latest ) > > to that docker hub image > > Those images are maintained by Docker Inc. and nothing to do with the > GCC project: > > https://docs.docker.com/docker-hub/official_images/ > > "All Official Images contain a User Feedback section in their > documentation which covers the details for that specific repository. > In most cases, the GitHub repository which contains the Dockerfiles > for an Official Repository also has an active issue tracker. General > feedback and support questions should be directed to #docker-library > on Freenode IRC." > Thanks a lot, i will contact the community and let them know Regards
