R-Devel (82603) still passing "fno-optimize-sibling-calls" on Windows
On 7/18/22 19:14, Avraham Adler wrote:
Hello. According to my understanding of the changes in R-devel, the flag -fno-optimize-sibling-calls should no longer be forced when compiling R. Yet, as I compile R, revision 82603, from source on Windows (Skylake-X server) I see the flag being passed (example below). Is this different for Windows than for Unix? The changes header seems to be its own and not under the "Installation on Unix-Alike" header. Thank you, Avi
Hi Avi, yes, thanks for reminding me, I will remove that in the Windows make files (in about a week or so). There is no need to be more careful on Windows than Unix regarding this (actually a bit on the contrary, as most users use the same compiler). Feel free to remove that already in your builds. Tomas
Excerpt: making Lapack.d from Lapack.c gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -ffloat-store -c dlamch.f -o dlamch.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c dlapack.f -o dlapack.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c cmplx.f -o cmplx.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c la_constants.f90 -o la_constants.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c dlartg.f90 -o dlartg.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c la_xisnan.f90 -o la_xisnan.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c dlassq.f90 -o dlassq.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c zlartg.f90 -o zlartg.o gfortran -O3 -march=native -pipe -fno-optimize-sibling-calls -flto -ffat-lto-objects -fuse-linker-plugin -c zlassq.f90 -o zlassq.o
______________________________________________ R-devel at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel