Re: Toy computational "benchmark" in Ada (new blog post)

[Paul Rubin <no.email@nospam.invalid>]

David Trudgett <dktrudgett@gmail.com> writes:
> If you used the project file, it should be optimised, but it obviously
> isn't, so I don't know what you did. Try GNAT Community 2019 instead,
> perhaps?

With -O3 removed from the .gpr, I see this:

    .L58:
            movq    -72(%rbp), %rax
            cmpq    %rax, -40(%rbp)
            jg      .L69
            movq    -40(%rbp), %rax
            movq    432(%rbx), %rdx
            leaq    0(,%rax,8), %rcx
            subq    $8, %rcx
            addq    %rcx, %rdx
            movq    %rdx, -80(%rbp)
            movq    432(%rbx), %rdx
            movsd   -8(%rdx,%rax,8), %xmm1
            movq    432(%rbx), %rdx
            movsd   -8(%rdx,%rax,8), %xmm0
    .LEHE26:
            mulsd   %xmm1, %xmm0
            movsd   -112(%rbp), %xmm1
            addsd   %xmm1, %xmm0
            movsd   %xmm0, -112(%rbp)
            addq    $1, -40(%rbp)
            jmp     .L58

I probably can't do much more tonight, but might try a newer GNAT
version in the coming days.  I'll also try GCC 9 for the C++ version.