John,

it's known explicitly-coded behavior. Previously, running multiple links in parallel used to kill performance rather than improve it. See:

     JAM_SEMAPHORE on $(targets) = <s>gcc-link-semaphore ;

in gcc.jam along with a comment above it. You can comment it out and see whether it improves performance. If so, it might become an option.

Thanks,
Volodya

On Thu, Jun 25, 2015 at 1:31 PM John Maddock <jz.maddock@googlemail.com> wrote:

Folks,

I've noticed something very strange, and very annoying with mingw builds
- the link steps all seem to be performed serially, never in parallel -
is this a known issue?

I'm using a win64 build of bjam, plus ming-x64. Building with msvc with
the same bjam executable works fine, as does building on Linux with GCC.

Even if I kill the build and restart, it still runs all the remaining
link steps in series, and it takes an age with mingw.

Any ideas?

Thanks, John.
_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build