"Carlos E.R." writes:
On 19/11/2019 10.16, Richard Kettlewell wrote:
"Carlos E.R." writes:
Are you familiar with compilers? Like GCC. If you search for
"optimize" you'll see that you can optimize for size or for speed -
mutually exclusive.

In the case of compilers, the two possibilities are not mutually
exclusive. Smaller is often faster.

No, is not completely true or not always. It is true for the load time
of the process from disk or swap, of course.

What isn’t completely true? I think “smaller is often faster” really is
true, as a statement about compiler optimisations. It’s vague, but that
doesn’t make it false. A few things that are bigger and faster
(e.g. loop unrolling, alignment, a subset of inlining) do not contradict
the statement. They would contradict “smaller is always faster”, or
“smaller is almost always faster”, but nobody is claiming either.

“Size vs speed” being mutually exclusive is definitely wrong though.
Output from gcc -O1 will pretty reliably be both smaller and faster than
from gcc -O0, for example.

--
https://www.greenend.org.uk/rjk/