https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65076
--- Comment #61 from Jan Hubicka <hubicka at gcc dot gnu.org> --- It seems we regressed slightly further. GCC 4.9: 27.87s GCC 5: 32.2 trunk: 37.78 furhter GGC memory use went up from 1053147 kB to 1221321 kB. GCC 4.9: phase parsing : 1.55 ( 5%) usr 1.24 (22%) sys 3.77 (10%) wall 167909 kB (16%) ggc phase lang. deferred : 2.19 ( 7%) usr 0.54 (10%) sys 2.77 ( 7%) wall 194757 kB (18%) ggc phase opt and generate : 28.52 (88%) usr 3.75 (68%) sys 32.43 (83%) wall 699106 kB (66%) ggc |name lookup : 0.36 ( 1%) usr 0.29 ( 5%) sys 0.70 ( 2%) wall 26015 kB ( 2%) ggc |overload resolution : 1.14 ( 4%) usr 0.24 ( 4%) sys 1.49 ( 4%) wall 118130 kB (11%) ggc GCC 5: phase parsing : 1.51 ( 5%) usr 1.09 (21%) sys 3.06 ( 9%) wall 172015 kB (16%) ggc phase lang. deferred : 2.17 ( 8%) usr 0.58 (11%) sys 2.76 ( 8%) wall 205684 kB (20%) ggc phase opt and generate : 24.17 (87%) usr 3.48 (67%) sys 27.69 (82%) wall 673958 kB (64%) ggc |name lookup : 0.48 ( 2%) usr 0.19 ( 4%) sys 0.73 ( 2%) wall 26569 kB ( 3%) ggc |overload resolution : 1.17 ( 4%) usr 0.38 ( 7%) sys 1.37 ( 4%) wall 121708 kB (12%) ggc GCC 6: phase setup : 0.01 ( 0%) usr 0.00 ( 0%) sys 0.01 ( 0%) wall 1396 kB ( 0%) ggc phase parsing : 2.54 ( 7%) usr 1.23 (22%) sys 3.77 ( 9%) wall 247024 kB (20%) ggc phase lang. deferred : 3.18 ( 8%) usr 0.63 (11%) sys 3.81 ( 9%) wall 235161 kB (19%) ggc phase opt and generate : 32.05 (85%) usr 3.86 (67%) sys 35.90 (83%) wall 737728 kB (60%) ggc |name lookup : 0.72 ( 2%) usr 0.17 ( 3%) sys 0.93 ( 2%) wall 35248 kB ( 3%) ggc |overload resolution : 1.85 ( 5%) usr 0.48 ( 8%) sys 2.09 ( 5%) wall 155432 kB (13%) ggc So quite noticeable regression in parsing up.
