Re: -fno-strength-reduce flag patch

Ralf Baechle (ralf@cobaltmicro.com)
Thu, 18 Sep 1997 11:42:23 -0700 (PDT)

>
> Ralf Baechle <ralf@cobaltmicro.com> writes:
>
> |> Only one of the optimizatons performed by the strength reduction pass
> |> was disabled. However some people say that few CPUs with just a few
> |> registers shouldn't use that optimization anyway as it tends to produce
> |> slower code, so -fno-strength-reduce is a good thing anyway. And
> |> to my knowledge the particular optimizer bug doesn't show up for
> |> other architectures.
>
> FWIW, this bug did show up on m68k as well. It seems to be rather generic.

After reading GCC's code I was assuming this as well. I however was not
able to make GCC generate bad code on != Intel for the piece of code that
was circulating as example for for this bug.

Ralf