Joshua Gatcomb <[EMAIL PROTECTED]> wrote:
> incidently - things are running much slower now using
> --optimize and -j -O2
> so how much is "much slower"
> for primes2.pasm it was taking on average 2.15 real
> seconds and is now taking on average 2.9
Indeed. Strange. It ought to be faster. Jumping
Joshua Gatcomb <[EMAIL PROTECTED]> wrote:
> This was not a requirement prior to the changes I
> outlined on the 16th.
Its in config/auto/gcc.pl, which needs obviously fixing for cygwin.
leo
incidently - things are running much slower now using
--optimize and -j -O2
so how much is "much slower"
for primes2.pasm it was taking on average 2.15 real
seconds and is now taking on average 2.9
Joshua Gatcomb
a.k.a. Limbic~Region
__
Do you
--- Leopold Toetsch <[EMAIL PROTECTED]> wrote:
> If you are using gcc 3.0 or above,
> -mno-accumulate-outgoing-args must be set.
This was not a requirement prior to the changes I
outlined on the 16th.
I was at gcc version 3.3.1 but I rebuilt 3.4.1 from
source. Adding the compiler flag appears
Joshua Gatcomb <[EMAIL PROTECTED]> wrote:
> 2. Almost everything but the most trivial JITed code
> was coredumping.
That's then caused by turning USE_CGP on. Strange. Cygwin gcc must
produce different code then somehow, which breaks things. If you are
using gcc 3.0 or above, -mno-accumulate-outg
As is part of my daily (to be read work day) routine,
I do a make realclean, a fresh CVS checkout, rebuild
parrot.
This morning I was noticing 2 problems
1. object tests 40-45 were blowing up - thanks to
Nicholas for pointing it out and Leo for fixing it
2. Almost everything but the most trivia
