| Message ID | SN2PR0701MB10248A6A4DCCC536CFA6E0538E610@SN2PR0701MB1024.namprd07.prod.outlook.com |
|---|---|
| State | New |
| Headers | show |
On Tue, Aug 25, 2015 at 5:29 AM, Hurugalawadi, Naveen <Naveen.Hurugalawadi@caviumnetworks.com> wrote: > Hi, > > Please find attached the remaining part of patch. > > Tested the patch on AArch64 and X86 without any regressions. > > Please review the patch and let me know if any modifications are required. Ok. Thanks, Richard. > Thanks, > Naveen > > ChangeLog > > 2015-08-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com> > > * fold-const.c (fold_binary_loc) : Move Optimize > root(x)*root(y) as root(x*y) to match.pd. > Move Optimize expN(x)*expN(y) as expN(x+y) to match.pd. > Move Optimize pow(x,y)*pow(x,z) as pow(x,y+z) to match.pd. > Move Optimize a/root(b/c) into a*root(c/b) to match.pd. > Move Optimize x/expN(y) into x*expN(-y) to match.pd. > > * match.pd (mult (root:s @0) (root:s @1)): New simplifier. > (mult (POW:s @0 @1) (POW:s @0 @2)) : New simplifier. > (mult (exps:s @0) (exps:s @1)) : New simplifier. > (rdiv @0 (root:s (rdiv:s @1 @2))) : New simplifier. > (rdiv @0 (exps:s @1)) : New simplifier.
"Hurugalawadi, Naveen" <Naveen.Hurugalawadi@caviumnetworks.com> writes: > * fold-const.c (fold_binary_loc) : Move Optimize > root(x)*root(y) as root(x*y) to match.pd. > Move Optimize expN(x)*expN(y) as expN(x+y) to match.pd. > Move Optimize pow(x,y)*pow(x,z) as pow(x,y+z) to match.pd. > Move Optimize a/root(b/c) into a*root(c/b) to match.pd. > Move Optimize x/expN(y) into x*expN(-y) to match.pd. > > * match.pd (mult (root:s @0) (root:s @1)): New simplifier. > (mult (POW:s @0 @1) (POW:s @0 @2)) : New simplifier. > (mult (exps:s @0) (exps:s @1)) : New simplifier. > (rdiv @0 (root:s (rdiv:s @1 @2))) : New simplifier. > (rdiv @0 (exps:s @1)) : New simplifier. FAIL: gcc.dg/builtins-11.c (test for excess errors) Excess errors: builtins-11.c:(.text+0x52): undefined reference to `link_error' Andreas.
On Thu, 27 Aug 2015, Andreas Schwab wrote: > "Hurugalawadi, Naveen" <Naveen.Hurugalawadi@caviumnetworks.com> writes: > >> * fold-const.c (fold_binary_loc) : Move Optimize >> root(x)*root(y) as root(x*y) to match.pd. >> Move Optimize expN(x)*expN(y) as expN(x+y) to match.pd. >> Move Optimize pow(x,y)*pow(x,z) as pow(x,y+z) to match.pd. >> Move Optimize a/root(b/c) into a*root(c/b) to match.pd. >> Move Optimize x/expN(y) into x*expN(-y) to match.pd. >> >> * match.pd (mult (root:s @0) (root:s @1)): New simplifier. >> (mult (POW:s @0 @1) (POW:s @0 @2)) : New simplifier. >> (mult (exps:s @0) (exps:s @1)) : New simplifier. >> (rdiv @0 (root:s (rdiv:s @1 @2))) : New simplifier. >> (rdiv @0 (exps:s @1)) : New simplifier. > > FAIL: gcc.dg/builtins-11.c (test for excess errors) > Excess errors: > builtins-11.c:(.text+0x52): undefined reference to `link_error' Indeed, generic-match ends up testing for sqrt(x)*sqrt(y) before testing for sqrt(x)@1*@1, so it simplifies it to sqrt(x*x)->abs(x) instead of plain x. Changing the genmatch machinery to better respect the order of patterns in match.pd might not be trivial without sacrificing performance, a workaround might be to add a special pattern sqrt(x)*sqrt(x), or to disable some patterns for GENERIC so CSE has a chance to run first.
On Fri, Aug 28, 2015 at 7:33 AM, Marc Glisse <marc.glisse@inria.fr> wrote: > On Thu, 27 Aug 2015, Andreas Schwab wrote: > >> "Hurugalawadi, Naveen" <Naveen.Hurugalawadi@caviumnetworks.com> writes: >> >>> * fold-const.c (fold_binary_loc) : Move Optimize >>> root(x)*root(y) as root(x*y) to match.pd. >>> Move Optimize expN(x)*expN(y) as expN(x+y) to match.pd. >>> Move Optimize pow(x,y)*pow(x,z) as pow(x,y+z) to match.pd. >>> Move Optimize a/root(b/c) into a*root(c/b) to match.pd. >>> Move Optimize x/expN(y) into x*expN(-y) to match.pd. >>> >>> * match.pd (mult (root:s @0) (root:s @1)): New simplifier. >>> (mult (POW:s @0 @1) (POW:s @0 @2)) : New simplifier. >>> (mult (exps:s @0) (exps:s @1)) : New simplifier. >>> (rdiv @0 (root:s (rdiv:s @1 @2))) : New simplifier. >>> (rdiv @0 (exps:s @1)) : New simplifier. >> >> >> FAIL: gcc.dg/builtins-11.c (test for excess errors) >> Excess errors: >> builtins-11.c:(.text+0x52): undefined reference to `link_error' > > > Indeed, generic-match ends up testing for sqrt(x)*sqrt(y) before testing for > sqrt(x)@1*@1, so it simplifies it to sqrt(x*x)->abs(x) instead of plain x. > Changing the genmatch machinery to better respect the order of patterns in > match.pd might not be trivial without sacrificing performance, a workaround > might be to add a special pattern sqrt(x)*sqrt(x), or to disable some > patterns for GENERIC so CSE has a chance to run first. Hum, it _does_ respect ordering (well, it is supposed to). But indeed I can see it does not. Bah. Mind opening a bugreport for this? Thanks, Richard. > -- > Marc Glisse
diff --git a/gcc/fold-const.c b/gcc/fold-const.c index 1e01726..c826e67 100644 --- a/gcc/fold-const.c +++ b/gcc/fold-const.c @@ -9947,51 +9947,6 @@ fold_binary_loc (location_t loc, if (flag_unsafe_math_optimizations) { - enum built_in_function fcode0 = builtin_mathfn_code (arg0); - enum built_in_function fcode1 = builtin_mathfn_code (arg1); - - /* Optimizations of root(...)*root(...). */ - if (fcode0 == fcode1 && BUILTIN_ROOT_P (fcode0)) - { - tree rootfn, arg; - tree arg00 = CALL_EXPR_ARG (arg0, 0); - tree arg10 = CALL_EXPR_ARG (arg1, 0); - - /* Optimize root(x)*root(y) as root(x*y). */ - rootfn = TREE_OPERAND (CALL_EXPR_FN (arg0), 0); - arg = fold_build2_loc (loc, MULT_EXPR, type, arg00, arg10); - return build_call_expr_loc (loc, rootfn, 1, arg); - } - - /* Optimize expN(x)*expN(y) as expN(x+y). */ - if (fcode0 == fcode1 && BUILTIN_EXPONENT_P (fcode0)) - { - tree expfn = TREE_OPERAND (CALL_EXPR_FN (arg0), 0); - tree arg = fold_build2_loc (loc, PLUS_EXPR, type, - CALL_EXPR_ARG (arg0, 0), - CALL_EXPR_ARG (arg1, 0)); - return build_call_expr_loc (loc, expfn, 1, arg); - } - - /* Optimizations of pow(...)*pow(...). */ - if ((fcode0 == BUILT_IN_POW && fcode1 == BUILT_IN_POW) - || (fcode0 == BUILT_IN_POWF && fcode1 == BUILT_IN_POWF) - || (fcode0 == BUILT_IN_POWL && fcode1 == BUILT_IN_POWL)) - { - tree arg00 = CALL_EXPR_ARG (arg0, 0); - tree arg01 = CALL_EXPR_ARG (arg0, 1); - tree arg10 = CALL_EXPR_ARG (arg1, 0); - tree arg11 = CALL_EXPR_ARG (arg1, 1); - - /* Optimize pow(x,y)*pow(x,z) as pow(x,y+z). */ - if (operand_equal_p (arg00, arg10, 0)) - { - tree powfn = TREE_OPERAND (CALL_EXPR_FN (arg0), 0); - tree arg = fold_build2_loc (loc, PLUS_EXPR, type, - arg01, arg11); - return build_call_expr_loc (loc, powfn, 2, arg00, arg); - } - } /* Canonicalize x*x as pow(x,2.0), which is expanded as x*x. */ if (!in_gimple_form @@ -10403,40 +10358,6 @@ fold_binary_loc (location_t loc, TREE_OPERAND (arg1, 0)); } - if (flag_unsafe_math_optimizations) - { - enum built_in_function fcode1 = builtin_mathfn_code (arg1); - - /* Optimize a/root(b/c) into a*root(c/b). */ - if (BUILTIN_CBRT_P (fcode1)) - { - tree rootarg = CALL_EXPR_ARG (arg1, 0); - - if (TREE_CODE (rootarg) == RDIV_EXPR) - { - tree rootfn = TREE_OPERAND (CALL_EXPR_FN (arg1), 0); - tree b = TREE_OPERAND (rootarg, 0); - tree c = TREE_OPERAND (rootarg, 1); - - tree tmp = fold_build2_loc (loc, RDIV_EXPR, type, c, b); - - tmp = build_call_expr_loc (loc, rootfn, 1, tmp); - return fold_build2_loc (loc, MULT_EXPR, type, arg0, tmp); - } - } - - /* Optimize x/expN(y) into x*expN(-y). */ - if (BUILTIN_EXPONENT_P (fcode1)) - { - tree expfn = TREE_OPERAND (CALL_EXPR_FN (arg1), 0); - tree arg = negate_expr (CALL_EXPR_ARG (arg1, 0)); - arg1 = build_call_expr_loc (loc, - expfn, 1, - fold_convert_loc (loc, type, arg)); - return fold_build2_loc (loc, MULT_EXPR, type, arg0, arg1); - } - - } return NULL_TREE; case TRUNC_DIV_EXPR: diff --git a/gcc/match.pd b/gcc/match.pd index eb0ba9d..289bc5c 100644 --- a/gcc/match.pd +++ b/gcc/match.pd @@ -2014,11 +2014,28 @@ along with GCC; see the file COPYING3. If not see (if (!HONOR_SNANS (type)) @0)) + /* Simplify sqrt(x) * sqrt(y) -> sqrt(x*y). */ + (for root (SQRT CBRT) + (simplify + (mult (root:s @0) (root:s @1)) + (root (mult @0 @1)))) + + /* Simplify pow(x,y) * pow(x,z) -> pow(x,y+z). */ + (simplify + (mult (POW:s @0 @1) (POW:s @0 @2)) + (POW @0 (plus @1 @2))) + /* Simplify pow(x,y) * pow(z,y) -> pow(x*z,y). */ (simplify (mult (POW:s @0 @1) (POW:s @2 @1)) (POW (mult @0 @2) @1)) + /* Simplify expN(x) * expN(y) -> expN(x+y). */ + (for exps (EXP EXP2 EXP10 POW10) + (simplify + (mult (exps:s @0) (exps:s @1)) + (exps (plus @0 @1)))) + /* Simplify tan(x) * cos(x) -> sin(x). */ (simplify (mult:c (TAN:s @0) (COS:s @0)) @@ -2061,9 +2078,16 @@ along with GCC; see the file COPYING3. If not see (POW @0 (minus @1 { build_one_cst (type); })))) /* Simplify a/root(b/c) into a*root(c/b). */ - (simplify - (rdiv @0 (SQRT:s (rdiv:s @1 @2))) - (mult @0 (SQRT (rdiv @2 @1)))) + (for root (SQRT CBRT) + (simplify + (rdiv @0 (root:s (rdiv:s @1 @2))) + (mult @0 (root (rdiv @2 @1))))) + + /* Simplify x/expN(y) into x*expN(-y). */ + (for exps (EXP EXP2 EXP10 POW10) + (simplify + (rdiv @0 (exps:s @1)) + (mult @0 (exps (negate @1))))) /* Simplify x / pow (y,z) -> x * pow(y,-z). */ (simplify