This is the mail archive of the gcc-bugs@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]

Re: big slowdown in egcs-1.1.2->gcc-2.95 on alpha


> I think we could fix this by using pseudo register birth / death lists
> instead of complete register sets.

That would be nice; here are times comparing egcs-1.1.2 and gcc-2.9.5 on
this alpha-ev6 box, without the profiling overhead (0 times have been elided):

popov-3% /usr/lib/gcc-lib/alpha-redhat-linux/egcs-2.91.66/cc1 g0-1.i
 __copysignf copysignf __copysign copysign __fabsf fabsf __fabs fabs __floorf __floor floorf floor __fdimf fdimf __fdim fdim ___H__20_g0_2d_1 ___init_proc ____20_g0_2d_1
time in parse: 10.890208
time in jump: 0.900848
time in flow: 2.382416
time in global-alloc: 10.848240
time in shorten-branch: 0.927200
time in final: 5.018592
time in varconst: 0.007808

popov-4% /export/u10/gcc-2.95/lib/gcc-lib/alphaev6-unknown-linux-gnu/2.95/cc1 g0-1.i
 __copysignf copysignf __copysign copysign __fabsf fabsf __fabs fabs __floorf __floor floorf floor __fdimf fdimf __fdim fdim ___H__20_g0_2d_1 ___init_proc ____20_g0_2d_1
time in parse: 11.073696
time in jump: 0.923296
time in flow: 3.405264
time in global-alloc: 17.851040
time in shorten-branch: 0.829600
time in final: 5.183536

Brad Lucier    lucier@math.purdue.edu


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]