Changeset 2a4a23 in git for m4


Ignore:
Timestamp:
Apr 20, 2018, 3:03:49 PM (6 years ago)
Author:
Hans Schoenemann <hannes@…>
Branches:
(u'fieker-DuVal', '117eb8c30fc9e991c4decca4832b1d19036c4c65')(u'spielwiese', '4bd32dfef92ec9f5ed8dceee82d14318ae147107')
Children:
eb6e1d58eb9834a39221581ba56d2e26f61499cc
Parents:
65c1bd5b526a9447123bf85f32538aff3498501c
Message:
fix: Z/p-arithmetic for all architectures
File:
1 edited

Legend:

Unmodified
Added
Removed
  • m4/cpu-check.m4

    r65c1bd r2a4a23  
    4949
    5050AS_CASE([$host_cpu],
     51dnl the following settings seems to be better on itanium processors
     52  [ia64*], [
     53      AC_DEFINE(HAVE_GENERIC_ADD,1,use branch for addition in Z/p otherwise it uses a generic add)],
    5154dnl the following settings seems to be better on i386 and x86_64 processors
    52   [i*86*|x86_64*], [AC_DEFINE(HAVE_MULT_MOD,1,multiplication is fast on the cpu: a*b is with mod otherwise using tables of logartihms)],
    53 dnl the following settings seems to be better on itanium processors
    54 dnl AC_DEFINE(HAVE_MULT_MOD,1,)
    55   [ia64*], [AC_DEFINE(HAVE_GENERIC_ADD,1,use branch for addition in Z/p otherwise it uses a generic add)],
     55dnl AC_DEFINE(HAVE_GENERIC_MULT,1,)
     56  [i*86*|x86_64*], [
     57            AC_DEFINE(HAVE_GENERIC_ADD,1,use branch for addition in Z/p otherwise it uses a generic add)
     58            AC_DEFINE(HAVE_GENERIC_MULT,1,multiplication is fast on the cpu: a*b is with mod otherwise using tables of logartihms)
     59            ],
    5660dnl the following settings seems to be better on sparc processors
    5761  [sparc*], [
    58             AC_DEFINE(HAVE_MULT_MOD,1,multiplication is fast on the cpu: a*b is with mod otherwise using tables of logartihms)
    59             AC_DEFINE(HAVE_DIV_MOD,1,division using extend euclidian algorithm otherwise using tables of logartihms)
     62            AC_DEFINE(HAVE_GENERIC_ADD,1,use branch for addition in Z/p otherwise it uses a generic add)
    6063            ],
    6164dnl the following settings seems to be better on ppc processors
    6265dnl testet on: ppc_Linux, 740/750 PowerMac G3, 512k L2 cache
    63   [powerpc*|ppc*], [AC_DEFINE(HAVE_MULT_MOD,1,multiplication is fast on the cpu: a*b is with mod otherwise using tables of logartihms)],
     66  [powerpc*|yyppc*], [AC_DEFINE(HAVE_GENERIC_MULT,1,multiplication is fast on the cpu: a*b is with mod otherwise using tables of logartihms)],
    6467  []
    6568)
Note: See TracChangeset for help on using the changeset viewer.