From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 1005) id 56B713858C83; Fri, 21 Apr 2023 19:01:33 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 56B713858C83 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1682103693; bh=HSYmsn3RHF4EIiIHUIKPDsDNdviCgu7NM5W3uSxRh8M=; h=From:To:Subject:Date:From; b=CEPyXCMzjYjaZg/vBxF6+sQ/y4ATHmHEwfRzU3yutrfjyM3Ijya2Ru5sJ/MsLgTZZ lyk7QIKkT6JJojo8GS8leEkxjHATubjfYZmb+l52PBZjfM0OhFejFkmzc7ZscAzz5/ Pe/Z6nuqSNp7I58+DMbKSF1wxy+NjXEITt0qJ/PA= Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit From: Michael Meissner To: gcc-cvs@gcc.gnu.org Subject: [gcc(refs/users/meissner/heads/work119)] Update ChangeLog.meissner. X-Act-Checkin: gcc X-Git-Author: Michael Meissner X-Git-Refname: refs/users/meissner/heads/work119 X-Git-Oldrev: db1da1abcc562f3f2cf109f37f4d932bff0571b9 X-Git-Newrev: 9c4f9bcb321c704b7a24ec1f101f3e2df2a5a0ce Message-Id: <20230421190133.56B713858C83@sourceware.org> Date: Fri, 21 Apr 2023 19:01:33 +0000 (GMT) List-Id: https://gcc.gnu.org/g:9c4f9bcb321c704b7a24ec1f101f3e2df2a5a0ce commit 9c4f9bcb321c704b7a24ec1f101f3e2df2a5a0ce Author: Michael Meissner Date: Fri Apr 21 15:00:58 2023 -0400 Update ChangeLog.meissner. 2023-04-21 Michael Meissner gcc/ * ChangeLog.meissner: Update. Diff: --- gcc/ChangeLog.meissner | 144 +++++++++---------------------------------------- 1 file changed, 25 insertions(+), 119 deletions(-) diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner index b16a81f8d58..82745912b34 100644 --- a/gcc/ChangeLog.meissner +++ b/gcc/ChangeLog.meissner @@ -1,151 +1,57 @@ -==================== Branch work119, patch #49 ==================== +==================== Branch work119, patch #51 ==================== -Fold sign or zero extension into vsx_extract from memory with variable element. - -This patch folds conversion to floating point of vsx_extract from memory of V4SI -elements where the element number is constant. This code optimizes things so it -will load the integer with LFIWAX or LFIWZX directly into a vector register -rather than loading it into a GPR and doing a direct move operation. - -2023-04-21 Michael Meissner - -gcc/ - - * config/rs6000/vsx.md (vsx_extract_v4si_var_load_to_): New - insn. - -gcc/testsuite/ - - * gcc.target/powerpc/vec-extract-mem-int-4.c: New file. - -==================== Branch work119, patch #48 ==================== +Combine vec_extract of V4SF with DF convert. -Fold sign or zero convert into variable vsx_extract from memory. +This patch adds a combine insn that merges loading up a vec_extract of V4SFmode +where the element number is constant combined with a conversion to DFmode. -This patch folds sign or zero convert operations into vsx_extract from memory -where the element number is constant. +In addition, I changed the vec_extract of V4SFmode where the element number is +constant without conversion to do the split before register allocation. 2023-04-21 Michael Meissner gcc/ - * config/rs6000/vsx.md (vsx_extract__var_load_to_udi): New insn. - (vsx_extract__var_load_to_sdi): New insn. - (vsx_extract_v8hi_var_load_to_si): New insn. + * config/rs6000/vsx.md (vsx_extract_v4sf_load): Allow split before + register allocation. + (vsx_extract_v4sf_to_df_load): New insn. gcc/testsuite/ - * gcc.target/powerpc/vec-extract-mem-char-2.c: New file. - * gcc.target/powerpc/vec-extract-mem-int-3.c: New file. - * gcc.target/powerpc/vec-extract-mem-short-2.c: New file. - -==================== Branch work119, patch #47 ==================== - -Allow vec_extract with variable element number to load vector registers. - -2023-04-21 Michael Meissner - -gcc/ - - * config/rs6000/vsx.md (vsx_extract__var_load): Allow vec_extract - of integer types with a variable element number to load into vector - registers. Allow splitting before register allocation. - -==================== Branch work119, patch #46 ==================== + * gcc.target/powerpc/vec-extract-mem-float-1.c: New test. -Combine variable element vec_extract of V4SF with DF convert. +==================== Branch work119, patch #50 ==================== -This patch adds a combine insn that merges loading up a vec_extract of V4SFmode -where the element number is variable combined with a conversion to DFmode. +Allow vec_extract support functions to be called before reload. 2023-04-21 Michael Meissner gcc/ - * config/rs6000/vsx.md (vsx_extract_v4sf_var_load_to_df): New insn. - -==================== Branch work119, patch #45 ==================== - -Fold conversion to float into V4SI vsx_extract from memory. + * config/rs6000/rs6000.cc (get_vector_offset): Allow being called before + register allocation. + (adjust_vec_address_pcrel): Likewise. + (rs6000_adjust_vec_address): Likewise. -This patch folds conversion to floating point of vsx_extract from memory of V4SI -elements where the element number is constant. This code optimizes things so it -will load the integer with LFIWAX or LFIWZX directly into a vector register -rather than loading it into a GPR and doing a direct move operation. +==================== Branch work119, patch #49 was reverted ==================== -2023-04-18 Michael Meissner +==================== Branch work119, patch #48 was reverted ==================== -gcc/ - - * config/rs6000/vsx.md (SIGN_ZERO_EXTEND): New mode attribute. - (vsx_extract_v4si_load_to_): New insn. +==================== Branch work119, patch #47 was reverted ==================== -gcc/testsuite/ +==================== Branch work119, patch #46 was reverted ==================== - * gcc.target/powerpc/vec-extract-mem-int-2.c: New file. +==================== Branch work119, patch #45 was reverted ==================== ==================== Branch work119, patch #44 was reverted ==================== -==================== Branch work119, patch #43 ==================== +==================== Branch work119, patch #43 was reverted ==================== -Fold sign or zero convert into vsx_extract from memory. +==================== Branch work119, patch #42 was reverted ==================== -This patch folds sign or zero convert operations into vsx_extract from memory -where the element number is constant. +==================== Branch work119, patch #41 was reverted ==================== -2023-04-21 Michael Meissner - -gcc/ - - * config/rs6000/vsx.md (VSX_EXTRACT_ISIGN): New mode attribute. - (vsx_extract__load_to_udi): New insn. - (vsx_extract__load_to_sdi): New insn. - (vsx_extract_v8hi_load_to_si): New insn. - -gcc/testsuite/ - - * gcc.target/powerpc/vec-extract-mem-char-1.c: New file. - * gcc.target/powerpc/vec-extract-mem-int-1.c: New file. - * gcc.target/powerpc/vec-extract-mem-short-1.c: New file. - -==================== Branch work119, patch #42 ==================== - -Allow vec_extract to load vector registers. - -2023-04-18 Michael Meissner - -gcc/ - - * config/rs6000/vsx.md (VSX_EX_ISA): New mode attribute - (vsx_extract__load): Allow vec_extract of integer types with a - constant element number to load into vector registers. Allow splitting - before register allocation. - -==================== Branch work119, patch #41 ==================== - -Combine vec_extract of V4SF with DF convert. - -This patch adds a combine insn that merges loading up a vec_extract of V4SFmode -where the element number is constant combined with a conversion to DFmode. - -2023-04-18 Michael Meissner - -gcc/ - - * config/rs6000/vsx.md (vsx_extract_v4sf_to_df_load): New insn. - -==================== Branch work119, patch #40 ==================== - -Allow vec_extract support functions to be called before reload. - -2023-04-19 Michael Meissner - -gcc/ - - * config/rs6000/rs6000.cc (get_vector_offset): Allow being called before - register allocation. - (adjust_vec_address_pcrel): Likewise. - (rs6000_adjust_vec_address): Likewise. +==================== Branch work119, patch #40 was reverted ==================== ==================== Branch work119, patch #33 was reverted ====================