This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
This adds back libgcc-math populated with x86_64 vectorized intrinsics for log, log2, log10, exp, sin, cos, cosf, expf, log10f, log2f, logf, powf and sinf. (The patches enabling the use of those are patch (2/3) enabling generic support for vectorizing intrinsics which is still the same as in the first post of the series, and patch (3/3) which will follow this submission) Changes from the previous version include adding a libgcc-math testsuite (with the test orginally in gcc.dg/vect/vect-fun-1.c), enabling libgcc-math also for i?86 if --enable-targets=all is given. Bootstrapped and tested on x86_64-unknown-linux-gnu (all 3 patches). Ok for mainline? Toplevel changes inline below, drop-in tarball for the libgcc-math directory attached. Thanks, Richard. 2006-11-16 Richard Guenther <rguenther@suse.de> * Makefile.def (target_modules): Add libgcc-math target module. * configure.in (target_libraries): Add libgcc-math target library. (--enable-libgcc-math): New configure switch. Enabled by default on x86_64-*-linux*. * Makefile.in: Regenerate. * configure: Regenerate. * doc/install.texi: Document --disable-libgcc-math configure switch. Index: Makefile.def =================================================================== *** Makefile.def (revision 119056) --- Makefile.def (working copy) *************** host_modules= { module= gnattools; }; *** 118,123 **** --- 118,124 ---- target_modules = { module= libstdc++-v3; lib_path=.libs; raw_cxx=true; }; target_modules = { module= libmudflap; lib_path=.libs; }; target_modules = { module= libssp; lib_path=.libs; }; + target_modules = { module= libgcc-math; lib_path=.libs; }; target_modules = { module= newlib; }; target_modules = { module= libgfortran; }; target_modules = { module= libobjc; }; Index: configure.in =================================================================== *** configure.in (revision 119056) --- configure.in (working copy) *************** target_libraries="target-libiberty \ *** 147,152 **** --- 147,153 ---- target-libstdc++-v3 \ target-libmudflap \ target-libssp \ + target-libgcc-math \ target-libgfortran \ target-boehm-gc \ ${libgcj} \ *************** AC_ARG_ENABLE(libssp, *** 312,317 **** --- 313,340 ---- ENABLE_LIBSSP=$enableval, ENABLE_LIBSSP=yes) + # Set the default so we build libgcc-math for x86_64-*-linux + AC_ARG_ENABLE(libgcc-math, + [ --enable-libgcc-math Builds libgcc-math directory],, + [ + case "${target}" in + x86_64-*-linux*) + enable_libgcc_math=yes ;; + i?86-*-linux*) + if test x$enable_targets = xall; then + enable_libgcc_math=yes + else + enable_libgcc_math=no + fi + ;; + *) + enable_libgcc_math=no ;; + esac + ]) + if test "${enable_libgcc_math}" != "yes"; then + noconfigdirs="$noconfigdirs target-libgcc-math" + fi + # Save it here so that, even in case of --enable-libgcj, if the Java # front-end isn't enabled, we still get libgcj disabled. libgcj_saved=$libgcj Index: gcc/doc/install.texi =================================================================== *** gcc/doc/install.texi (revision 119056) --- gcc/doc/install.texi (working copy) *************** should not be built. *** 1141,1146 **** --- 1141,1150 ---- @item --disable-libgomp Specify that the run-time libraries used by GOMP should not be built. + @item --disable-libgcc-math + Specify that the run-time libraries for arch and gcc specific math + functions should not be built. + @item --with-dwarf2 Specify that the compiler should use DWARF 2 debugging information as the default.
Attachment:
libgcc-math2-2.tar.bz2
Description: application/bzip
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |