[R-pkg-devel] NOTE: Installation took CPU time 3.5 times elapsed time

[Dirk Eddelbuettel]

Hi Ivan,

On 29 July 2025 at 13:33, Ivan Krylov via R-package-devel wrote:
| On Mon, 28 Jul 2025 06:45:06 -0500
| Dirk Eddelbuettel <edd using debian.org> wrote:
| 
| > And I just validated it: removing that last line from DESCRIPTION
| > removes the compilation time NOTE.
| 
| This must be the culprit. The 00install.out file [1] shows -flto=8 being
| used, which GCC documentation [2] describes as "use 8 threads for LTO":
| 
| >> If you specify the optional n [in -flto=n], the optimization and
| >> code generation done at link time is executed in parallel using n
| >> parallel jobs by utilizing an installed make program. The
| >> environment variable MAKE may be used to override the program used. 
| 
| I think this was set in the Makeconf on the machine running the
| "gcc-san" special checks; I don't see anything in the package that
| would set 8 threads in particular. R itself, when configured with
| --enable-lto, only sets -flto, not -flto=8.

That's good sleuthing, as usual. But note that my machine has no -flto=$N for
some N but rather '-flto=auto' which also results in the NOTE.

So an extra check may be needed to not issue the nag if LTO is used?

Dirk

| -- 
| Best regards,
| Ivan
| 
| [1]
| https://win-builder.r-project.org/incoming_pretest/boostmath_1.0.1_20250727_090857/specialChecks/gcc-san/package/00install.out
| 
| [2]
| https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html#index-flto
| 
| ______________________________________________
| R-package-devel using r-project.org mailing list
| https://stat.ethz.ch/mailman/listinfo/r-package-devel

-- 
dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org