Re: [PATCH v3 00/20] kbuild: significantly speed up kernel builds

[Nick Desaulniers]

On Thu, Sep 17, 2026 at 12:42 PM Lorenzo Stoakes (ARM) <ljs@xxxxxxxxxx> wrote:
>
> On Thu, Sep 17, 2026 at 06:36:18PM +0100, Lorenzo Stoakes (ARM) wrote:
> > On Thu, Sep 17, 2026 at 10:15:26AM -0700, Linus Torvalds wrote:
> > > On Thu, 17 Sept 2026 at 09:06, Lorenzo Stoakes (ARM) <ljs@xxxxxxxxxx> wrote:
> > > >
> > > >                 == defconfig FULL build ==
> > > >
> > > >                         before    after     delta
> > > >                       ----------------------------------
> > > > Threadripper, gcc        32.5s    26.3s     -6.1s (-19%)
> > > > Threadripper, clang      34.5s    27.2s     -7.3s (-21%)
> > > > EPYC, gcc                28.7s    20.6s     -8.1s (-28%)
> > > > EPYC, clang              39.8s    32.2s     -7.6s (-19%)
> > > > M2, gcc                 519.3s   512.4s     -6.9s (-1%)
> > > > M2, clang               634.5s   569.4s    -65.1s (-10%)
> > >
> > > There's something odd going on with that M2 gcc build. Why is it such
> > > an odd duck?
> >
> > Ugh dear lord I think the LLM somehow corrupted the numbers... Let me
> > re-run those myself locally and come back with the actual times!
>
> OK numbers are in:
>
> M2, gcc                 564.7s   512.1s    -52.6s (-9%)
> M2, clang               615.1s   564.1s    -51.0s (-8%)
>
> From the v1:
>
> M2, gcc                 564.0s   512.4s    -51.6s (-9%)
> M2, clang               616.1s   569.4s    -46.7s (-8%)
>
> (Tested as per the cover letter pre-reqs, i.e. applying dependent series
> first etc.)
>
> Not sure on the -5s delta on the clang build :) that might have been me

Clang is faster on Thursdays.

> being more disciplined this time to not use the laptop at the same time.
>
> But broadly in line with the original numbers.
>
> >
> > >
> > >              Linus
> >
> > --
> > Cheers, Lorenzo
>
> --
> Cheers, Lorenzo



-- 
Thanks,
~Nick Desaulniers