------- Additional Comments From dje at gcc dot gnu dot org 2005-01-19 02:44 ------- Steven's experiments seem to demonstrate that the current DSE implementation is not very effective. GCC 4.0 includes RTL optimizations that will catch most if not all of these cases, so it is not as if this will cause GCC to miss some optimizations. If the concern is performance impact, then I think that we should consider adding adding some aggressive optimizations instead of retaining DSE. Even if DSE is very fast, what does GCC gain by fewer iterations of an optimization pass that has almost no benefit?
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18880
