------- Comment #15 from hubicka at gcc dot gnu dot org 2008-09-18 16:49 ------- We seem to make nice progress on the testcase ;)
Conversion to FUD would probably help here, but looking at the DF dumps, the testcase don't look like your average testcase showing DU/UD chain explosion. The chains seem relatively short at least in manual inspection through first part of dump. Yet we produe about 40 DU/UD records live at once for every instruction ever emit, so it seems we are doing some wrong still. If someone more familiar with dataflow branch could look into why we have so many DU/UDs, it would be great. Honza -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37448
