Loren P Meissner wrote:
> - And even so, I think it would take a lot of expertise to write code
> in
> assembler that runs as fast as what's produced by modern optimizers.
and James Giles wrote:
> Nope. The code produced by the compiler is known to
> the programmer. You can always at least break even with
> the compiler.
Maybe it depends on the hardware and the people. Even in the mid-1960s
the Univac 1107, with its multiple accumulators and multiple index
registers which partly overlapped, and its six-pass Fortran compiler,
was emitting code that was cleverer and more efficient than most
assembler programmers, at least at the institution where I was working,
could generate.
David Muxworthy
|