This is the mail archive of the 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]
Other format: [Raw text]

Re: [PATCH] Overhaul __builtin_constant_p processing (take 2)

Hi Roger,

I mentioned previously the regression for gcc.c-torture/execute/20010119-1.c
on hppa2.0w-hp-hpux11* on the trunk.  I have tracked this regression to your

2003-01-21  Roger Sayle  <roger at eyesopen dot com>

        PR optimization/8423
	* cse.c (fold_rtx): Only eliminate a CONSTANT_P_RTX to 1 when
	its argument is constant, or 0 if !flag_gcse.
	* simplify-rtx.c (simplify_rtx): Convert CONSTANT_P_RTX to 1
	if it's argument is constant.
	* gcse.c (want_to_gcse_p): Ignore CONSTANT_P_RTX nodes.
	(hash_scan_set): Don't record CONSTANT_P_RTX expressions.
	(do_local_cprop): Don't propagate CONSTANT_P_RTX constants.
	* builtins.c (purge_builtin_constant_p): New function to force
	instantiation of any remaining CONSTANT_P_RTX nodes.
	* rtl.h (purge_builtin_constant_p): Prototype here.
	* toplev.c (rest_of_compilation): Invoke purge_builtin_constant_p
	pass after GCSE and before loop.
	(flag_gcse): No longer static.
	* flags.h (flag_gcse): Prototype here.

Possibly, some of the slow down that you saw was due to this optimization

J. David Anglin                                  dave dot anglin at nrc-cnrc dot gc dot ca
National Research Council of Canada              (613) 990-0752 (FAX: 952-6605)

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