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]: PR29335 use MPFR for two-argument math builtins (pow, atan2, hypot)

Hi Kaveh,

On Sat, 28 Oct 2006, Kaveh R. GHAZI wrote:
> hypot(x,0) -> fabs(x)
> hypot(0,y) -> fabs(y)
> hypot(x,x) -> x*sqrt(2)
> hypot(-x,y) -> hypot(x,y)
> hypot(x,-y) -> hypot(x,y)
> hypot(-x,-y) -> hypot(x,y)


> 2006-10-27  Kaveh R. Ghazi  <>
> 	PR middle-end/29335
> 	* builtins.c (do_mpfr_arg2, fold_builtin_hypot): New.
> 	(fold_builtin_pow): Evaluate constant arguments at compile-time
> 	using MPFR.
> 	(fold_builtin_1): Handle BUILT_IN_ATAN2 and BUILT_IN_HYPOT.
> 	(do_mpfr_ckconv): New helper function.
> 	(do_mpfr_arg1): Use do_mpfr_ckconv.
> 	(do_mpfr_arg2): New.
> testsuite:
> 	* gcc.dg/builtins-20.c: Add tests for hypot.
> 	* gcc.dg/torture/builtin-math-2.c (TESTIT2): New.  Add tests for
> 	two-argument builtins.
> 	* gcc.dg/torture/builtin-math-3.c (TESTIT_R): Renamed from
> 	TESTIT2.  Update all callers.
> 	(TESTIT2, TESTIT2_R): New helper macros.
> 	Add testcases for pow, hypot and atan2.

This is OK for mainline.

> +      tree narg0 = TREE_CODE (arg0) == NEGATE_EXPR ? TREE_OPERAND (arg0, 0) : arg0;
> +      tree narg1 = TREE_CODE (arg1) == NEGATE_EXPR ? TREE_OPERAND (arg1, 0) : arg1;

Though these long lines might need wrapping.  In fact, you could
probably also generalize this code to also handle:

hypot(fabs(x),y) = hypot(x,y)
hypot(x,fabs(y)) = hypot(x,y)
hypot(-fabs(x),-fabs(y)) = hypot(x,y)

by also testing for ABS_EXPR.  This would require the above expressions
to be slightly rewriten anyway.  Perhaps best left as a follow-up patch,
as you appear to be on a roll.

Nice clean-up with do_mpfr_ckconv.  Thanks.


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