On Tue, 26 Aug 2025 at 12:15, Avi Kivity <avi_at_[hidden]> wrote:

> On Tue, 2025-08-26 at 11:01 +0100, Jonathan Wakely via Std-Proposals wrote:
>
>
>
> On Mon, 25 Aug 2025 at 02:12, Adrian Johnston via Std-Proposals <
> std-proposals_at_[hidden]> wrote:
>
> Hello,
>
> (If you spend a lot of time looking at generated assembly you might
> want to skip this one.)
>
> As we all know the compiler's budget for optimizing C++ is an
> implementation-defined metric where it knows best what is needed and
> we are not supposed to be doing the compiler's job for it.
> Unfortunately, this view has never really worked well in practice.
> Recently I spent some time looking at the generated assembly from gcc
> and clang and it doesn't look like the situation has improved in 30
> years.
>
>
> This is nonsense.
>
> Optimizations like autovectorization and conditional devirtualization do
> lots that wasn't possible 30 years ago. A problem is that many programs are
> much bigger than they were 30 years ago.
>
>
> IIUC, the big leap came with Scalar Reduction of Aggregates (SRA). Small
> structs can now be broken up intro their constituent scalars, each of which
> can be allocated registers separately, spilled to their stack separately,
> and have their lifetimes tracked separately. IIRC it was directly intended
> to reduce the C++ abstraction cost.
>

Yes, the abstraction penalty is not zero, but it's much lower than it was
15-20 years ago, let alone 30 years ago.

Early inlining also made a difference for GCC.