2011/7/13 Michael Meissner <meiss...@linux.vnet.ibm.com>: > I just ran a spec 2006 run on the powerpc (32-bit) last night setting the > reassociation to 2. I do see a win in bwaves, but unfortunately it is not > enough of a win, and it is still a regression to GCC 4.5. However, I see some > regressions in 3 other benchmarks (I tend to omit differences of less than > 2%): > > 401.bzip2 97.99% > 410.bwaves 113.88% > 436.cactusADM 93.96% > 444.namd 93.74% >
I tried it on AMD, Core i5, ARM (Cortex A9) and Atom using SPEC2000 and EEMBC suites. All platforms showed both gains and regressions. The only platform whose gains outweighted regressions enough without additional tuning was Atom. Surely we should avoid tree height reduction in some cases and may tune it but I think it will be nice to have feature checked in first since at least one platform benefits from it. Ilya
