Re: [R-SIG-Mac] Link-Time Optimization (LTO)

2020-07-15 Thread Prof Brian Ripley

On 15/07/2020 09:50, Jeroen Ooms wrote:

On Tue, Jul 14, 2020 at 4:22 PM Prof Brian Ripley  wrote:


This is a rather technical post about how libraries of compiled code can
be further optimized.  LTO generally produces smaller[*] and faster code
(typically by a few percent) at the expense of increased installation
time and is being used for large projects such as browsers and soon for
some Linux distributions.

I have committed a series of enhancements to LTO support in R-devel and
will shortly port the more important of these to R-patched.


Would it be worthwhile looking into this for Windows? We did enable
support for LTO in the rtools40 toolchains*, but those are gcc-8.3.0
and some of the benefits require gcc-9.

* 
https://github.com/r-windows/rtools-packages/blob/master/mingw-w64-gcc/PKGBUILD#L166


Way off topic for R-sig-mac, but it is under discussion for Windows once 
all the planned LTO changes are in.


A minor point which is relevant here: the recommended gfortran 
distribution for macOS (which is from GCC 8.2) contains gcc and g++.  So 
Mac users could try that to get C/Fortran consistency checks.  However, 
only much later versions are compatible with Catalina's SDK 
(https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90835), and from trying on 
High Sierra it looks like Apple's linker does not understand GCC's LTO 
format.


--
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] Link-Time Optimization (LTO)

2020-07-15 Thread Jeroen Ooms
On Tue, Jul 14, 2020 at 4:22 PM Prof Brian Ripley  wrote:
>
> This is a rather technical post about how libraries of compiled code can
> be further optimized.  LTO generally produces smaller[*] and faster code
> (typically by a few percent) at the expense of increased installation
> time and is being used for large projects such as browsers and soon for
> some Linux distributions.
>
> I have committed a series of enhancements to LTO support in R-devel and
> will shortly port the more important of these to R-patched.

Would it be worthwhile looking into this for Windows? We did enable
support for LTO in the rtools40 toolchains*, but those are gcc-8.3.0
and some of the benefits require gcc-9.

* 
https://github.com/r-windows/rtools-packages/blob/master/mingw-w64-gcc/PKGBUILD#L166

___
R-SIG-Mac mailing list
R-SIG-Mac@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-mac