From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1725) id B58EA385041E; Tue, 15 Jun 2021 17:19:42 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B58EA385041E Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: William Schmidt To: gcc-cvs@gcc.gnu.org Subject: [gcc(refs/users/wschmidt/heads/builtins10)] rs6000: Support for vectorizing built-in functions X-Act-Checkin: gcc X-Git-Author: Bill Schmidt X-Git-Refname: refs/users/wschmidt/heads/builtins10 X-Git-Oldrev: 42754a4c2e4e8d86f042000cfb58da98b78937b1 X-Git-Newrev: 000aba9eb77810a73e54764f781eef265afcc7cc Message-Id: <20210615171942.B58EA385041E@sourceware.org> Date: Tue, 15 Jun 2021 17:19:42 +0000 (GMT) X-BeenThere: gcc-cvs@gcc.gnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gcc-cvs mailing list List-Unsubscribe: , List-Archive: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Jun 2021 17:19:42 -0000 https://gcc.gnu.org/g:000aba9eb77810a73e54764f781eef265afcc7cc commit 000aba9eb77810a73e54764f781eef265afcc7cc Author: Bill Schmidt Date: Fri Mar 5 10:45:56 2021 -0600 rs6000: Support for vectorizing built-in functions 2021-03-05 Bill Schmidt gcc/ * config/rs6000/rs6000.c (rs6000-builtins.h): New include. (rs6000_new_builtin_vectorized_function): New function. (rs6000_new_builtin_md_vectorized_function): Likewise. (rs6000_builtin_vectorized_function): Call rs6000_new_builtin_vectorized_function. (rs6000_builtin_md_vectorized_function): Call rs6000_new_builtin_md_vectorized_function. Diff: --- gcc/config/rs6000/rs6000.c | 200 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 200 insertions(+) diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c index 75c2cc4f8b7..37eb05ddf4f 100644 --- a/gcc/config/rs6000/rs6000.c +++ b/gcc/config/rs6000/rs6000.c @@ -77,6 +77,7 @@ #include "case-cfn-macros.h" #include "ppc-auxv.h" #include "rs6000-internal.h" +#include "rs6000-builtins.h" #include "opts.h" /* This file should be included last. */ @@ -5488,6 +5489,198 @@ rs6000_loop_unroll_adjust (unsigned nunroll, struct loop *loop) return nunroll; } +/* Returns a function decl for a vectorized version of the builtin function + with builtin function code FN and the result vector type TYPE, or NULL_TREE + if it is not available. */ + +static tree +rs6000_new_builtin_vectorized_function (unsigned int fn, tree type_out, + tree type_in) +{ + machine_mode in_mode, out_mode; + int in_n, out_n; + + if (TARGET_DEBUG_BUILTIN) + fprintf (stderr, "rs6000_new_builtin_vectorized_function (%s, %s, %s)\n", + combined_fn_name (combined_fn (fn)), + GET_MODE_NAME (TYPE_MODE (type_out)), + GET_MODE_NAME (TYPE_MODE (type_in))); + + if (TREE_CODE (type_out) != VECTOR_TYPE + || TREE_CODE (type_in) != VECTOR_TYPE) + return NULL_TREE; + + out_mode = TYPE_MODE (TREE_TYPE (type_out)); + out_n = TYPE_VECTOR_SUBPARTS (type_out); + in_mode = TYPE_MODE (TREE_TYPE (type_in)); + in_n = TYPE_VECTOR_SUBPARTS (type_in); + + switch (fn) + { + CASE_CFN_COPYSIGN: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_CPSGNDP]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_CPSGNSP]; + if (VECTOR_UNIT_ALTIVEC_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_COPYSIGN_V4SF]; + break; + CASE_CFN_CEIL: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVRDPIP]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVRSPIP]; + if (VECTOR_UNIT_ALTIVEC_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VRFIP]; + break; + CASE_CFN_FLOOR: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVRDPIM]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVRSPIM]; + if (VECTOR_UNIT_ALTIVEC_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VRFIM]; + break; + CASE_CFN_FMA: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVMADDDP]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVMADDSP]; + if (VECTOR_UNIT_ALTIVEC_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VMADDFP]; + break; + CASE_CFN_TRUNC: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVRDPIZ]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVRSPIZ]; + if (VECTOR_UNIT_ALTIVEC_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VRFIZ]; + break; + CASE_CFN_NEARBYINT: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && flag_unsafe_math_optimizations + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVRDPI]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && flag_unsafe_math_optimizations + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVRSPI]; + break; + CASE_CFN_RINT: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && !flag_trapping_math + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_XVRDPIC]; + if (VECTOR_UNIT_VSX_P (V4SFmode) + && !flag_trapping_math + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_XVRSPIC]; + break; + default: + break; + } + + /* Generate calls to libmass if appropriate. */ + if (rs6000_veclib_handler) + return rs6000_veclib_handler (combined_fn (fn), type_out, type_in); + + return NULL_TREE; +} + +/* Implement TARGET_VECTORIZE_BUILTIN_MD_VECTORIZED_FUNCTION. */ + +static tree +rs6000_new_builtin_md_vectorized_function (tree fndecl, tree type_out, + tree type_in) +{ + machine_mode in_mode, out_mode; + int in_n, out_n; + + if (TARGET_DEBUG_BUILTIN) + fprintf (stderr, + "rs6000_new_builtin_md_vectorized_function (%s, %s, %s)\n", + IDENTIFIER_POINTER (DECL_NAME (fndecl)), + GET_MODE_NAME (TYPE_MODE (type_out)), + GET_MODE_NAME (TYPE_MODE (type_in))); + + if (TREE_CODE (type_out) != VECTOR_TYPE + || TREE_CODE (type_in) != VECTOR_TYPE) + return NULL_TREE; + + out_mode = TYPE_MODE (TREE_TYPE (type_out)); + out_n = TYPE_VECTOR_SUBPARTS (type_out); + in_mode = TYPE_MODE (TREE_TYPE (type_in)); + in_n = TYPE_VECTOR_SUBPARTS (type_in); + + enum rs6000_gen_builtins fn + = (enum rs6000_gen_builtins) DECL_MD_FUNCTION_CODE (fndecl); + switch (fn) + { + case RS6000_BIF_RSQRTF: + if (VECTOR_UNIT_ALTIVEC_OR_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VRSQRTFP]; + break; + case RS6000_BIF_RSQRT: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_RSQRT_2DF]; + break; + case RS6000_BIF_RECIPF: + if (VECTOR_UNIT_ALTIVEC_OR_VSX_P (V4SFmode) + && out_mode == SFmode && out_n == 4 + && in_mode == SFmode && in_n == 4) + return rs6000_builtin_decls_x[RS6000_BIF_VRECIPFP]; + break; + case RS6000_BIF_RECIP: + if (VECTOR_UNIT_VSX_P (V2DFmode) + && out_mode == DFmode && out_n == 2 + && in_mode == DFmode && in_n == 2) + return rs6000_builtin_decls_x[RS6000_BIF_RECIP_V2DF]; + break; + default: + break; + } + return NULL_TREE; +} + /* Handler for the Mathematical Acceleration Subsystem (mass) interface to a library with vectorized intrinsics. */ @@ -5607,6 +5800,9 @@ rs6000_builtin_vectorized_function (unsigned int fn, tree type_out, machine_mode in_mode, out_mode; int in_n, out_n; + if (new_builtins_are_live) + return rs6000_new_builtin_vectorized_function (fn, type_out, type_in); + if (TARGET_DEBUG_BUILTIN) fprintf (stderr, "rs6000_builtin_vectorized_function (%s, %s, %s)\n", combined_fn_name (combined_fn (fn)), @@ -5738,6 +5934,10 @@ rs6000_builtin_md_vectorized_function (tree fndecl, tree type_out, machine_mode in_mode, out_mode; int in_n, out_n; + if (new_builtins_are_live) + return rs6000_new_builtin_md_vectorized_function (fndecl, type_out, + type_in); + if (TARGET_DEBUG_BUILTIN) fprintf (stderr, "rs6000_builtin_md_vectorized_function (%s, %s, %s)\n", IDENTIFIER_POINTER (DECL_NAME (fndecl)),