From maintainers-request at octave dot org Fri Jul 30 15:07:32 2004 Subject: Re: CPU flags? From: Paul THOMAS To: Paul Kienzle , octave maintainers mailing list Date: Fri, 30 Jul 2004 22:07:03 +0200 (CEST) ------=_Part_12017_8015213.1091218023742 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Paul, Do you mean the values that -march and -mcpu-type can take on? I have: # CPU types supported in gcc-3.2 and higher: athlon-xp, athlon-mp, athlon-4= , # athlon-tbird, athlon, k6, k6-2, k6-3, i386, i486, i586 (Pentium), i686 # (PentiumPro), pentium, pentium-mmx, pentiumpro, pentium2 (Celeron), penti= um3, # and pentium4. ( from http://www.linuxquestions.org/questions/archive/40/2= 004/06/2/192003 ) Paul T > Message du 30/07/04 03:29 > De : "Paul Kienzle"=20 > A : "octave maintainers mailing list"=20 > Copie =E0 :=20 > Objet : CPU flags? > Hi, >=20 > Does anyone know the flags for determining if we are running on > x86 architecture? To support 53-bit mantissa random number > generation, we found we needed some x86 specific code > in order to reduce the overhead from 2.4x to 1.6x the speed > of the 31-bit mantissa code. >=20 > Thanks, >=20 > - Paul >=20 > ------=_Part_12017_8015213.1091218023742 Content-Type: text/html Content-Transfer-Encoding: quoted-printable

Paul,

Do you mean the values that -march and -mcpu-type can take on?  I h= ave:

# CPU types supported in gcc-3.2 and higher: athlon-xp, athlon-mp, athlo= n-4,
# athlon-tbird, athlon, k6, k6-2, k6-3, i386, i486, i586 (Pentium),= i686
# (PentiumPro), pentium, pentium-mmx, pentiumpro, pentium2 (Celero= n), pentium3,
# and pentium4. ( from http://www.linuxquestions.org/= questions/archive/40/2004/06/2/192003 )

Paul T




> Message du 30/07/04 03:29
> De : "Paul Kienzle" <= PKIENZLE at USERS dot SOURCEFORGE dot NET>
> A : "octave maintainers mailing lis= t"
> Copie =E0 :
> Objet : CPU flags?=
> Hi,
>
> Does anyone know the flags for determining if= we are running on
> x86 architecture? To support 53-bit mantissa ran= dom number
> generation, we found we needed some x86 specific code> in order to reduce the overhead from 2.4x to 1.6x the speed
> o= f the 31-bit mantissa code.
>
> Thanks,
>
> - Pau= l
>
>
------=_Part_12017_8015213.1091218023742--