Re: [R-SIG-Mac] Apple Clang does support OpenMP (if libomp is available)
Jon, some time ago Apple's clang has silently dropped -fopenmp so we were able to at least keep it in the flags even if it wasn't actually using it. Still, it was only dropping it, it wasn't actually generating any parallel code, so there was real point in using it. That's when we decided to use non-Apple clang since it has proper support for parallel code generation. What you're pointing out is that more recent clang builds from Apple now do include the OMP generation pieces (in particular the pre-processor), they just don't ship any OpenMP-related headers or libraries. The problem is that it's hard to rely on something that is unsupported and works only on some systems. We could ship libomp (we already do - no need to get it from homebrew), but, for example, the omp.h headers from clang7 don't work with Apple's clang (you do need explicit -I.. and -lomp in addition to -Wp,-fopenmp because it's not only the pre-processor that needs to know about it). Also note that "-Wp,-fopenmp" is a hack - just because Apple has not patched the pre-processor and only the front-end/linker to refuse it, it doesn't mean that it actually works properly. There is no guarantee that they made changes to the code-generation which break something in OpenMP since it's not not even included in any of their tests. Hence I'd say it's a quick hack if you don't want to to install clang7 that may or may not work depending on your code, but not something I'd trust in a release. Cheers, Simon > On Jun 6, 2019, at 6:42 AM, Jon Clayden wrote: > > Dear all, > > Lack of OpenMP support in Apple’s build of Clang is cited as one > reason for not using it in CRAN builds, but this is only partly true: > after installing libomp from Homebrew, I have been adding > “-Wp,-fopenmp” to CXXFLAGS and CFLAGS (and “-lomp” to LIBS) to my > builds for a while, and everything seems to work OK. > > I’m not sure how far back this support goes (and I haven’t tried the > Xcode 11 tools yet), but is there any known reason not to take this > approach, and if not, would it make sense for R’s configure script to > try “-Wp,-fopenmp” when testing for OpenMP support? > > All the best, > Jon > > ___ > R-SIG-Mac mailing list > R-SIG-Mac@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-mac > ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Apple Clang does support OpenMP (if libomp is available)
Hi James, > > Lack of OpenMP support in Apple’s build of Clang is cited as one reason for > > not using it in CRAN builds > > From R 3.4.x forward, OpenMP has been enabled in CRAN builds as the toolchain > is using a custom compiler. I’m aware of that, but this makes building packages from source more complicated for users in my opinion, since they also have to install the custom compiler toolchains. But whatever one’s views on that, I’m suggesting that one reason given for not using Apple Clang may not be entirely correct. > > but this is only partly true: after installing libomp ***from Homebrew***, > > I have been adding > > Unfortunately, both Xcode and Xcode CLI do not ship clang with support for > OpenMP. By installing `libomp` from homebrew, the system is getting a > clang-compliant OpenMP compiler from the llvm formula. To check, please see: Not at all. The libomp formula is standalone, with no dependency on the llvm formula. See https://github.com/Homebrew/homebrew-core/blob/master/Formula/libomp.rb. I can assure you that I don’t have the llvm formula installed. > which clang /usr/bin/clang All the best, Jon ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Apple Clang does support OpenMP (if libomp is available)
Greetings and Salutations Jon, > Lack of OpenMP support in Apple’s build of Clang is cited as one reason for > not using it in CRAN builds From R 3.4.x forward, OpenMP has been enabled in CRAN builds as the toolchain is using a custom compiler. You can see this with regard to the work done to enable OpenMP with RcppArmadillo with a configure script to detect the compiler: https://github.com/RcppCore/RcppArmadillo/pull/170 https://github.com/RcppCore/RcppArmadillo/pull/219 > but this is only partly true: after installing libomp ***from Homebrew***, I > have been adding Unfortunately, both Xcode and Xcode CLI do not ship clang with support for OpenMP. By installing `libomp` from homebrew, the system is getting a clang-compliant OpenMP compiler from the llvm formula. To check, please see: which clang I wish it were the case though... Sincerely, JJB On 6/6/19, 5:43 AM, "R-SIG-Mac on behalf of Jon Clayden" wrote: Dear all, Lack of OpenMP support in Apple’s build of Clang is cited as one reason for not using it in CRAN builds, but this is only partly true: after installing libomp from Homebrew, I have been adding “-Wp,-fopenmp” to CXXFLAGS and CFLAGS (and “-lomp” to LIBS) to my builds for a while, and everything seems to work OK. I’m not sure how far back this support goes (and I haven’t tried the Xcode 11 tools yet), but is there any known reason not to take this approach, and if not, would it make sense for R’s configure script to try “-Wp,-fopenmp” when testing for OpenMP support? All the best, Jon ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
[R-SIG-Mac] Apple Clang does support OpenMP (if libomp is available)
Dear all, Lack of OpenMP support in Apple’s build of Clang is cited as one reason for not using it in CRAN builds, but this is only partly true: after installing libomp from Homebrew, I have been adding “-Wp,-fopenmp” to CXXFLAGS and CFLAGS (and “-lomp” to LIBS) to my builds for a while, and everything seems to work OK. I’m not sure how far back this support goes (and I haven’t tried the Xcode 11 tools yet), but is there any known reason not to take this approach, and if not, would it make sense for R’s configure script to try “-Wp,-fopenmp” when testing for OpenMP support? All the best, Jon ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Warning on XCode / Command Line Tools 11
On 06/06/2019 09:31, peter dalgaard wrote: FWIW, no updates are suggested for the source-building machine in my office, also running Mojave. (June 5 was Constitution Day and Election Day, so I have been away from the machine until now.) It stopping nagging me yesterday afternoon, so likely it was an Apple snafu. (I heard from others who had been caught by this, and of course Apple allows for 'automatic' updating so this could happen in the background.) We do need to keep an eye on the tools though. My current setup is an -um- eclectic mix (*) of old and new tools, some of which are 32 bit, and this is going to be a problem in 10.15 Catalina. I did try getting the tools better in line with Simon's setup at some point, but there were hiccups and then I ran out of time and couldn't risk collateral damage. You have a few months yet. The residual 32-bit applications I have are from Adobe, e.g. old uninstallers. -pd (*) ="dogs dinner" On 4 Jun 2019, at 17:10 , Prof Brian Ripley wrote: My Mojave machine today prompted me to update to version 11 of Command Line Tools (apparently 'beta 1', but I am not a Beta tester). In short: don't do that (you can revert to 10.2.1 from the URL above, at least if your Apple ID has (free) developer privileges). CoreFoundation.framework has been removed (and its headers are used in a couple of spots compiling R). Further its seems that /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg is no longer included and setting CPPFLAGS did not suffice. Hopefully we will have solutions in due course, but that version is really for 10.15 Catalina which is months away ('in the Fall', whenever that is). -- Brian D. Ripley, rip...@stats.ox.ac.uk Emeritus Professor of Applied Statistics, University of Oxford ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac -- Brian D. Ripley, rip...@stats.ox.ac.uk Emeritus Professor of Applied Statistics, University of Oxford ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Warning on XCode / Command Line Tools 11
FWIW, no updates are suggested for the source-building machine in my office, also running Mojave. (June 5 was Constitution Day and Election Day, so I have been away from the machine until now.) We do need to keep an eye on the tools though. My current setup is an -um- eclectic mix (*) of old and new tools, some of which are 32 bit, and this is going to be a problem in 10.15 Catalina. I did try getting the tools better in line with Simon's setup at some point, but there were hiccups and then I ran out of time and couldn't risk collateral damage. -pd (*) ="dogs dinner" > On 4 Jun 2019, at 17:10 , Prof Brian Ripley wrote: > > My Mojave machine today prompted me to update to version 11 of Command Line > Tools (apparently 'beta 1', but I am not a Beta tester). > > In short: don't do that (you can revert to 10.2.1 from the URL above, at > least if your Apple ID has (free) developer privileges). > > CoreFoundation.framework has been removed (and its headers are used in a > couple of spots compiling R). > > Further its seems that > /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg > is no longer included and setting CPPFLAGS did not suffice. > > Hopefully we will have solutions in due course, but that version is really > for 10.15 Catalina which is months away ('in the Fall', whenever that is). > > -- > Brian D. Ripley, rip...@stats.ox.ac.uk > Emeritus Professor of Applied Statistics, University of Oxford > > ___ > R-SIG-Mac mailing list > R-SIG-Mac@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-mac -- Peter Dalgaard, Professor, Center for Statistics, Copenhagen Business School Solbjerg Plads 3, 2000 Frederiksberg, Denmark Phone: (+45)38153501 Office: A 4.23 Email: pd@cbs.dk Priv: pda...@gmail.com ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac