RE: [Bug-gnubg] Re: gnubg bug?

[Ingo Macherius]

Problem can not be reprosuced on Linux x86_64 with gcc 4.3 either, despite heavy Optimization flags used. All gnubg settings are compile time default.

 

gcc version 4.3.2 (Debian 4.3.2-1.1)

Target: x86_64-linux-gnu

 

gcc -pipe -combine -march=core2 -O3 -funsafe-loop-optimizations -funsafe-math-optimizations -ffast-math -freciprocal-math -ftree-vectorize -mfpmath=sse -mssse3 -msse3 -msse -msse2 -fomit-frame-pointer -msahf

 

(inim)set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA

[...]

(inim) hint 1
    1. Cubeful 0-ply    21/17 19/18                  MWC:  22.81%
       0.000 0.000 0.000 - 1.000 1.000 0.085
        0-ply cubeful [expert]

Ingo

-----Original Message-----
From: address@hidden [mailto:address@hidden On Behalf Of Michael Petch
Sent: Wednesday, August 05, 2009 4:44 AM
To: Zulli, Louis P
Cc: address@hidden; Øystein Johansen; address@hidden
Subject: Re: [Bug-gnubg] Re: gnubg bug?

I can confirm Zulli’s findings as well. On the MAC the issue doesn’t exist (Intel  with OS/X 10.5.7). Pruning doesn’t alter the behavior (It just works as expected) I used both version of GCC on apple (4.0.1 and 4.2.1):
 

 On 04/08/09 7:42 PM, "Zulli, Louis P" <address@hidden> wrote:

Anyway, with Pruning On or Off and at all ply levels I seem to get 21/17 19/18 being best.

Maybe there's a gcc-4.3 related problem? Maybe I'll try Intel's icc later.

Louis