From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx0a-001b2d01.pphosted.com (mx0a-001b2d01.pphosted.com [148.163.156.1]) by sourceware.org (Postfix) with ESMTPS id 987093858D39 for ; Mon, 23 Oct 2023 01:29:52 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 987093858D39 Authentication-Results: sourceware.org; dmarc=none (p=none dis=none) header.from=linux.ibm.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=linux.ibm.com ARC-Filter: OpenARC Filter v1.0.0 sourceware.org 987093858D39 Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=148.163.156.1 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1698024594; cv=none; b=oqJGaltxaElyqSv40AWuicbtiXnzFh8bGl4aqpyKfKNqA5oAkK9srVGl2M7RPPtjpn/v7hn/5NhGJldVPljyO77MiCg/D4V9bqg4iVEIK3+EzINj/XKRWejn61DbOWQILXMykP1FpTmpzXdGTxDeCwV7g8ieQAi09+NyuUpTV/o= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1698024594; c=relaxed/simple; bh=U6fvDqU6s1ypivmM0h1seEG4fMfLpqc7OLiijg1Pohs=; h=DKIM-Signature:Message-ID:Date:Subject:To:From:MIME-Version; b=ecb5CsZTYQTA+Fm5I0qWjPYtg9Aa7010EpBDmsyPtopFDN6QaDfXwsLodGmYvFFx3nHL07eWQcQYHeP53yUZpkhBGwVRbPmaYost+5s3tLlCnQGrBIVoHlq0zIjPslSG85Al9yWl14kmiallEPkSRdzKKMocHs59d+mKJbrAiko= ARC-Authentication-Results: i=1; server2.sourceware.org Received: from pps.filterd (m0356517.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.17.1.19/8.17.1.19) with ESMTP id 39MNgOrp009514; Mon, 23 Oct 2023 01:29:49 GMT DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=ibm.com; h=message-id : date : subject : to : references : cc : from : in-reply-to : content-type : content-transfer-encoding : mime-version; s=pp1; bh=DlTT4m+jue+pmMWN9GLFl3xuxY+KWNwFmqikCTUHiwM=; b=n2TChfVl3/B9vHIcubAyVTW0iVZKc82/rGKCWV1DbmC8JkDrCV+cv/wlZeUJTTsqdr1s wmcLvWbV8A+QUKgwdIsZp+PcpONH7ogw8ioIPXqQQ4MQt6BFBH5oWJ1XN9skacH2mqAH Y6IXgxDVW4RLc4HBV/DlVXWqML+K+3aaJOJE/4Z6Z6/RQQNGj7bHUkAU9V2b66Qei7j8 IrCGECyO9NfmPrk8Z+pRZnhjdtLnI+sYGdI21NzgqmM8wDDCqJRabcp0qNfmhaB8RMu6 8gvN7nHDS/4G4Q56eDo5ogvIn3s/JypUd6Ujr4F7C4FpryqjGQEA/bYf/WIyQivCxppq QA== Received: from ppma13.dal12v.mail.ibm.com (dd.9e.1632.ip4.static.sl-reverse.com [50.22.158.221]) by mx0a-001b2d01.pphosted.com (PPS) with ESMTPS id 3twdn41ykx-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 23 Oct 2023 01:29:48 +0000 Received: from pps.filterd (ppma13.dal12v.mail.ibm.com [127.0.0.1]) by ppma13.dal12v.mail.ibm.com (8.17.1.19/8.17.1.19) with ESMTP id 39MLrX6B024365; Mon, 23 Oct 2023 01:29:47 GMT Received: from smtprelay01.fra02v.mail.ibm.com ([9.218.2.227]) by ppma13.dal12v.mail.ibm.com (PPS) with ESMTPS id 3tvu6jn8pm-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Mon, 23 Oct 2023 01:29:47 +0000 Received: from smtpav01.fra02v.mail.ibm.com (smtpav01.fra02v.mail.ibm.com [10.20.54.100]) by smtprelay01.fra02v.mail.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id 39N1TjtF18612950 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Mon, 23 Oct 2023 01:29:45 GMT Received: from smtpav01.fra02v.mail.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id 9692E20040; Mon, 23 Oct 2023 01:29:45 +0000 (GMT) Received: from smtpav01.fra02v.mail.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id A5C0C20043; Mon, 23 Oct 2023 01:29:43 +0000 (GMT) Received: from [9.197.224.185] (unknown [9.197.224.185]) by smtpav01.fra02v.mail.ibm.com (Postfix) with ESMTP; Mon, 23 Oct 2023 01:29:43 +0000 (GMT) Message-ID: <236f3854-2b52-10d9-08d4-df2d36bcfff2@linux.ibm.com> Date: Mon, 23 Oct 2023 09:29:42 +0800 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101 Thunderbird/102.15.1 Subject: Re: [PATCH-1v4, expand] Enable vector mode for compare_by_pieces [PR111449] To: Richard Sandiford References: <992fc509-51d6-3dbd-ee9d-393378227fc5@linux.ibm.com> Content-Language: en-US Cc: gcc-patches From: HAO CHEN GUI In-Reply-To: Content-Type: text/plain; charset=UTF-8 X-TM-AS-GCONF: 00 X-Proofpoint-GUID: WlipSTpYI8_hBqJFdcOvpLTZlol_Gvm1 X-Proofpoint-ORIG-GUID: WlipSTpYI8_hBqJFdcOvpLTZlol_Gvm1 Content-Transfer-Encoding: 8bit X-Proofpoint-UnRewURL: 0 URL was un-rewritten MIME-Version: 1.0 X-Proofpoint-Virus-Version: vendor=baseguard engine=ICAP:2.0.272,Aquarius:18.0.980,Hydra:6.0.619,FMLib:17.11.176.26 definitions=2023-10-22_23,2023-10-19_01,2023-05-22_02 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 clxscore=1015 mlxscore=0 phishscore=0 adultscore=0 spamscore=0 suspectscore=0 priorityscore=1501 impostorscore=0 bulkscore=0 mlxlogscore=999 malwarescore=0 lowpriorityscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.12.0-2310170001 definitions=main-2310230010 X-Spam-Status: No, score=-14.8 required=5.0 tests=BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_EF,GIT_PATCH_0,KAM_SHORT,NICE_REPLY_A,RCVD_IN_MSPIKE_H4,RCVD_IN_MSPIKE_WL,SPF_HELO_NONE,SPF_PASS,TXREP autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org List-Id: Committed as r14-4835. https://gcc.gnu.org/g:f08ca5903c7a02b450b93143467f70b9fd8e0085 Thanks Gui Haochen 在 2023/10/20 16:49, Richard Sandiford 写道: > HAO CHEN GUI writes: >> Hi, >> Vector mode instructions are efficient for compare on some targets. >> This patch enables vector mode for compare_by_pieces. Two help >> functions are added to check if vector mode is available for certain >> by pieces operations and if if optabs exists for the mode and certain >> by pieces operations. One member is added in class op_by_pieces_d to >> record the type of operations. >> >> The test case is in the second patch which is rs6000 specific. >> >> Compared to last version, the main change is to add a target hook >> check - scalar_mode_supported_p when retrieving the available scalar >> modes. The mode which is not supported for a target should be skipped. >> (e.g. TImode on ppc). Also some function names and comments are refined >> according to reviewer's advice. >> >> Bootstrapped and tested on x86 and powerpc64-linux BE and LE with no >> regressions. >> >> Thanks >> Gui Haochen >> >> ChangeLog >> Expand: Enable vector mode for by pieces compares >> >> Vector mode compare instructions are efficient for equality compare on >> rs6000. This patch refactors the codes of by pieces operation to enable >> vector mode for compare. >> >> gcc/ >> PR target/111449 >> * expr.cc (can_use_qi_vectors): New function to return true if >> we know how to implement OP using vectors of bytes. >> (qi_vector_mode_supported_p): New function to check if optabs >> exists for the mode and certain by pieces operations. >> (widest_fixed_size_mode_for_size): Replace the second argument >> with the type of by pieces operations. Call can_use_qi_vectors >> and qi_vector_mode_supported_p to do the check. Call >> scalar_mode_supported_p to check if the scalar mode is supported. >> (by_pieces_ninsns): Pass the type of by pieces operation to >> widest_fixed_size_mode_for_size. >> (class op_by_pieces_d): Remove m_qi_vector_mode. Add m_op to >> record the type of by pieces operations. >> (op_by_pieces_d::op_by_pieces_d): Change last argument to the >> type of by pieces operations, initialize m_op with it. Pass >> m_op to function widest_fixed_size_mode_for_size. >> (op_by_pieces_d::get_usable_mode): Pass m_op to function >> widest_fixed_size_mode_for_size. >> (op_by_pieces_d::smallest_fixed_size_mode_for_size): Call >> can_use_qi_vectors and qi_vector_mode_supported_p to do the >> check. >> (op_by_pieces_d::run): Pass m_op to function >> widest_fixed_size_mode_for_size. >> (move_by_pieces_d::move_by_pieces_d): Set m_op to MOVE_BY_PIECES. >> (store_by_pieces_d::store_by_pieces_d): Set m_op with the op. >> (can_store_by_pieces): Pass the type of by pieces operations to >> widest_fixed_size_mode_for_size. >> (clear_by_pieces): Initialize class store_by_pieces_d with >> CLEAR_BY_PIECES. >> (compare_by_pieces_d::compare_by_pieces_d): Set m_op to >> COMPARE_BY_PIECES. > > OK, thanks. And thanks for your patience. > > Richard > >> patch.diff >> diff --git a/gcc/expr.cc b/gcc/expr.cc >> index 2c9930ec674..ad5f9dd8ec2 100644 >> --- a/gcc/expr.cc >> +++ b/gcc/expr.cc >> @@ -988,18 +988,44 @@ alignment_for_piecewise_move (unsigned int max_pieces, unsigned int align) >> return align; >> } >> >> -/* Return the widest QI vector, if QI_MODE is true, or integer mode >> - that is narrower than SIZE bytes. */ >> +/* Return true if we know how to implement OP using vectors of bytes. */ >> +static bool >> +can_use_qi_vectors (by_pieces_operation op) >> +{ >> + return (op == COMPARE_BY_PIECES >> + || op == SET_BY_PIECES >> + || op == CLEAR_BY_PIECES); >> +} >> + >> +/* Return true if optabs exists for the mode and certain by pieces >> + operations. */ >> +static bool >> +qi_vector_mode_supported_p (fixed_size_mode mode, by_pieces_operation op) >> +{ >> + if ((op == SET_BY_PIECES || op == CLEAR_BY_PIECES) >> + && optab_handler (vec_duplicate_optab, mode) != CODE_FOR_nothing) >> + return true; >> + >> + if (op == COMPARE_BY_PIECES >> + && optab_handler (mov_optab, mode) != CODE_FOR_nothing >> + && can_compare_p (EQ, mode, ccp_jump)) >> + return true; >> >> + return false; >> +} >> + >> +/* Return the widest mode that can be used to perform part of an >> + operation OP on SIZE bytes. Try to use QI vector modes where >> + possible. */ >> static fixed_size_mode >> -widest_fixed_size_mode_for_size (unsigned int size, bool qi_vector) >> +widest_fixed_size_mode_for_size (unsigned int size, by_pieces_operation op) >> { >> fixed_size_mode result = NARROWEST_INT_MODE; >> >> gcc_checking_assert (size > 1); >> >> /* Use QI vector only if size is wider than a WORD. */ >> - if (qi_vector && size > UNITS_PER_WORD) >> + if (can_use_qi_vectors (op) && size > UNITS_PER_WORD) >> { >> machine_mode mode; >> fixed_size_mode candidate; >> @@ -1009,8 +1035,7 @@ widest_fixed_size_mode_for_size (unsigned int size, bool qi_vector) >> { >> if (GET_MODE_SIZE (candidate) >= size) >> break; >> - if (optab_handler (vec_duplicate_optab, candidate) >> - != CODE_FOR_nothing) >> + if (qi_vector_mode_supported_p (candidate, op)) >> result = candidate; >> } >> >> @@ -1019,9 +1044,14 @@ widest_fixed_size_mode_for_size (unsigned int size, bool qi_vector) >> } >> >> opt_scalar_int_mode tmode; >> + scalar_int_mode mode; >> FOR_EACH_MODE_IN_CLASS (tmode, MODE_INT) >> - if (GET_MODE_SIZE (tmode.require ()) < size) >> - result = tmode.require (); >> + { >> + mode = tmode.require (); >> + if (GET_MODE_SIZE (mode) < size >> + && targetm.scalar_mode_supported_p (mode)) >> + result = mode; >> + } >> >> return result; >> } >> @@ -1061,8 +1091,7 @@ by_pieces_ninsns (unsigned HOST_WIDE_INT l, unsigned int align, >> { >> /* NB: Round up L and ALIGN to the widest integer mode for >> MAX_SIZE. */ >> - mode = widest_fixed_size_mode_for_size (max_size, >> - op == SET_BY_PIECES); >> + mode = widest_fixed_size_mode_for_size (max_size, op); >> if (optab_handler (mov_optab, mode) != CODE_FOR_nothing) >> { >> unsigned HOST_WIDE_INT up = ROUND_UP (l, GET_MODE_SIZE (mode)); >> @@ -1076,8 +1105,7 @@ by_pieces_ninsns (unsigned HOST_WIDE_INT l, unsigned int align, >> >> while (max_size > 1 && l > 0) >> { >> - mode = widest_fixed_size_mode_for_size (max_size, >> - op == SET_BY_PIECES); >> + mode = widest_fixed_size_mode_for_size (max_size, op); >> enum insn_code icode; >> >> unsigned int modesize = GET_MODE_SIZE (mode); >> @@ -1317,8 +1345,8 @@ class op_by_pieces_d >> bool m_push; >> /* True if targetm.overlap_op_by_pieces_p () returns true. */ >> bool m_overlap_op_by_pieces; >> - /* True if QI vector mode can be used. */ >> - bool m_qi_vector_mode; >> + /* The type of operation that we're performing. */ >> + by_pieces_operation m_op; >> >> /* Virtual functions, overriden by derived classes for the specific >> operation. */ >> @@ -1331,7 +1359,7 @@ class op_by_pieces_d >> public: >> op_by_pieces_d (unsigned int, rtx, bool, rtx, bool, by_pieces_constfn, >> void *, unsigned HOST_WIDE_INT, unsigned int, bool, >> - bool = false); >> + by_pieces_operation); >> void run (); >> }; >> >> @@ -1349,11 +1377,11 @@ op_by_pieces_d::op_by_pieces_d (unsigned int max_pieces, rtx to, >> void *from_cfn_data, >> unsigned HOST_WIDE_INT len, >> unsigned int align, bool push, >> - bool qi_vector_mode) >> + by_pieces_operation op) >> : m_to (to, to_load, NULL, NULL), >> m_from (from, from_load, from_cfn, from_cfn_data), >> m_len (len), m_max_size (max_pieces + 1), >> - m_push (push), m_qi_vector_mode (qi_vector_mode) >> + m_push (push), m_op (op) >> { >> int toi = m_to.get_addr_inc (); >> int fromi = m_from.get_addr_inc (); >> @@ -1375,8 +1403,7 @@ op_by_pieces_d::op_by_pieces_d (unsigned int max_pieces, rtx to, >> { >> /* Find the mode of the largest comparison. */ >> fixed_size_mode mode >> - = widest_fixed_size_mode_for_size (m_max_size, >> - m_qi_vector_mode); >> + = widest_fixed_size_mode_for_size (m_max_size, m_op); >> >> m_from.decide_autoinc (mode, m_reverse, len); >> m_to.decide_autoinc (mode, m_reverse, len); >> @@ -1401,7 +1428,7 @@ op_by_pieces_d::get_usable_mode (fixed_size_mode mode, unsigned int len) >> if (len >= size && prepare_mode (mode, m_align)) >> break; >> /* widest_fixed_size_mode_for_size checks SIZE > 1. */ >> - mode = widest_fixed_size_mode_for_size (size, m_qi_vector_mode); >> + mode = widest_fixed_size_mode_for_size (size, m_op); >> } >> while (1); >> return mode; >> @@ -1414,7 +1441,7 @@ fixed_size_mode >> op_by_pieces_d::smallest_fixed_size_mode_for_size (unsigned int size) >> { >> /* Use QI vector only for > size of WORD. */ >> - if (m_qi_vector_mode && size > UNITS_PER_WORD) >> + if (can_use_qi_vectors (m_op) && size > UNITS_PER_WORD) >> { >> machine_mode mode; >> fixed_size_mode candidate; >> @@ -1427,8 +1454,7 @@ op_by_pieces_d::smallest_fixed_size_mode_for_size (unsigned int size) >> break; >> >> if (GET_MODE_SIZE (candidate) >= size >> - && (optab_handler (vec_duplicate_optab, candidate) >> - != CODE_FOR_nothing)) >> + && qi_vector_mode_supported_p (candidate, m_op)) >> return candidate; >> } >> } >> @@ -1451,7 +1477,7 @@ op_by_pieces_d::run () >> >> /* widest_fixed_size_mode_for_size checks M_MAX_SIZE > 1. */ >> fixed_size_mode mode >> - = widest_fixed_size_mode_for_size (m_max_size, m_qi_vector_mode); >> + = widest_fixed_size_mode_for_size (m_max_size, m_op); >> mode = get_usable_mode (mode, length); >> >> by_pieces_prev to_prev = { nullptr, mode }; >> @@ -1516,8 +1542,7 @@ op_by_pieces_d::run () >> else >> { >> /* widest_fixed_size_mode_for_size checks SIZE > 1. */ >> - mode = widest_fixed_size_mode_for_size (size, >> - m_qi_vector_mode); >> + mode = widest_fixed_size_mode_for_size (size, m_op); >> mode = get_usable_mode (mode, length); >> } >> } >> @@ -1543,7 +1568,7 @@ class move_by_pieces_d : public op_by_pieces_d >> move_by_pieces_d (rtx to, rtx from, unsigned HOST_WIDE_INT len, >> unsigned int align) >> : op_by_pieces_d (MOVE_MAX_PIECES, to, false, from, true, NULL, >> - NULL, len, align, PUSHG_P (to)) >> + NULL, len, align, PUSHG_P (to), MOVE_BY_PIECES) >> { >> } >> rtx finish_retmode (memop_ret); >> @@ -1632,15 +1657,16 @@ move_by_pieces (rtx to, rtx from, unsigned HOST_WIDE_INT len, >> class store_by_pieces_d : public op_by_pieces_d >> { >> insn_gen_fn m_gen_fun; >> + >> void generate (rtx, rtx, machine_mode) final override; >> bool prepare_mode (machine_mode, unsigned int) final override; >> >> public: >> store_by_pieces_d (rtx to, by_pieces_constfn cfn, void *cfn_data, >> unsigned HOST_WIDE_INT len, unsigned int align, >> - bool qi_vector_mode) >> + by_pieces_operation op) >> : op_by_pieces_d (STORE_MAX_PIECES, to, false, NULL_RTX, true, cfn, >> - cfn_data, len, align, false, qi_vector_mode) >> + cfn_data, len, align, false, op) >> { >> } >> rtx finish_retmode (memop_ret); >> @@ -1729,8 +1755,8 @@ can_store_by_pieces (unsigned HOST_WIDE_INT len, >> max_size = STORE_MAX_PIECES + 1; >> while (max_size > 1 && l > 0) >> { >> - fixed_size_mode mode >> - = widest_fixed_size_mode_for_size (max_size, memsetp); >> + auto op = memsetp ? SET_BY_PIECES : STORE_BY_PIECES; >> + auto mode = widest_fixed_size_mode_for_size (max_size, op); >> >> icode = optab_handler (mov_optab, mode); >> if (icode != CODE_FOR_nothing >> @@ -1793,7 +1819,7 @@ store_by_pieces (rtx to, unsigned HOST_WIDE_INT len, >> optimize_insn_for_speed_p ())); >> >> store_by_pieces_d data (to, constfun, constfundata, len, align, >> - memsetp); >> + memsetp ? SET_BY_PIECES : STORE_BY_PIECES); >> data.run (); >> >> if (retmode != RETURN_BEGIN) >> @@ -1814,7 +1840,7 @@ clear_by_pieces (rtx to, unsigned HOST_WIDE_INT len, unsigned int align) >> /* Use builtin_memset_read_str to support vector mode broadcast. */ >> char c = 0; >> store_by_pieces_d data (to, builtin_memset_read_str, &c, len, align, >> - true); >> + CLEAR_BY_PIECES); >> data.run (); >> } >> >> @@ -1832,12 +1858,13 @@ class compare_by_pieces_d : public op_by_pieces_d >> void generate (rtx, rtx, machine_mode) final override; >> bool prepare_mode (machine_mode, unsigned int) final override; >> void finish_mode (machine_mode) final override; >> + >> public: >> compare_by_pieces_d (rtx op0, rtx op1, by_pieces_constfn op1_cfn, >> void *op1_cfn_data, HOST_WIDE_INT len, int align, >> rtx_code_label *fail_label) >> : op_by_pieces_d (COMPARE_MAX_PIECES, op0, true, op1, true, op1_cfn, >> - op1_cfn_data, len, align, false) >> + op1_cfn_data, len, align, false, COMPARE_BY_PIECES) >> { >> m_fail_label = fail_label; >> }