Try to use pre compiled header to accelerate the build

[jorisv]

I setup pch on pinocchio_pywrap_default and obtain the following result:

Build with PCH:    ninja pinocchio_pywrap_default  702.06s user 20.62s system 1337% cpu 54.044 total
Build without PCH: ninja pinocchio_pywrap_default  979.61s user 32.80s system 1565% cpu 1:04.67 total

Also, memory consumption seem to be lower. But it's hard de measure.

[github-actions]

👋 Hi,
This is a reminder message to assign an extra build label to this Pull Request if needed.
By default, this PR will be build with minimal build options (URDF support and Python bindings)
The possible extra labels are:

• build_collision (build Pinocchio with coal support)
• build_casadi (build Pinocchio with CasADi support)
• build_autodiff (build Pinocchio with CppAD support)
• build_codegen (build Pinocchio with CppADCodeGen support)
• build_extra (build Pinocchio with extra algorithms)
• build_mpfr (build Pinocchio with Boost.Multiprecision support)
• build_sdf (build Pinocchio with SDF parser)
• build_accelerate (build Pinocchio with APPLE Accelerate framework support)
• build_all (build Pinocchio with ALL the options stated above)

Thanks.
The Pinocchio development team.

[ManifoldFR]

That's a 28.3% decrease in terms of compilation time. That's kinda nuts.

Have you also tried this on an Apple Silicon Mac ? Like the M1 Ultra Desktop we have, or perhaps @lmontaut's laptop ?

[ManifoldFR]

Also, it changes nothing about symbols etc?

[jorisv]

That's a 28.3% decrease in terms of compilation time. That's kinda nuts.

Have you also tried this on an Apple Silicon Mac ? Like the M1 Ultra Desktop we have, or perhaps @lmontaut's laptop ?

I only did the benchmark on my computer. We can compare CI time, but since we use ccache, I don't think it will be fair.

Also, it changes nothing about symbols etc?

I don't think it change anything on the symbol. But you're right, I will try to compare binary size.