[llvm-branch-commits] [clang] [clang-tools-extra] [compiler-rt] [flang] [libc] [libcxx] [lldb] [llvm] [LV] Reduce register usage for scaled reductions (PR #133090)

Sam Tebbs via llvm-branch-commits llvm-branch-commits at lists.llvm.org
Wed Apr 9 07:17:36 PDT 2025


Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,zcfh <1412805291 at qq.com>,Alexey
 Bataev <a.bataev at outlook.com>,Florian Hahn <flo at fhahn.com>,Alexey Bataev
 <a.bataev at outlook.com>,Hristo Hristov <hghristov.rmm at gmail.com>,Mircea
 Trofin <mtrofin at google.com>,Florian Hahn <flo at fhahn.com>,Jonas Devlieghere
 <jonas at devlieghere.com>,Henry Jiang <h243jian at uwaterloo.ca>,Alexander
 Yermolovich <43973793+ayermolo at users.noreply.github.com>,Andy Kaylor
 <akaylor at nvidia.com>,Andy Kaylor <akaylor at nvidia.com>,Florian Hahn
 <flo at fhahn.com>,Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Andre Kuhlenschmidt
 <andre.kuhlenschmidt at gmail.com>,Jan Svoboda <jan_svoboda at apple.com>,Jorge
 Gorbe Moya <jgorbe at google.com>,Sumit Agarwal <sumitagarwal330 at gmail.com>,Andre
 Kuhlenschmidt <andre.kuhlenschmidt at gmail.com>,Louis Dionne
 <ldionne.2 at gmail.com>,Jason Molenda <jmolenda at apple.com>,modiking
 <modiking213 at gmail.com>,Ian Anderson <iana at apple.com>,Aditya Tejpaul
 <97700214+hoarfrost32 at users.noreply.github.com>,Alex MacLean
 <amaclean at nvidia.com>,Jorge Gorbe Moya <jgorbe at google.com>,Michael Jones
 <michaelrj at google.com>,Slava Zakharin <szakharin at nvidia.com>,Jerry-Ge
 <jerry.ge at arm.com>,Jerry-Ge <jerry.ge at arm.com>,Sudharsan Veeravalli
 <quic_svs at quicinc.com>,Rodrigo Rocha <rcor at dcc.ufmg.br>,Mircea Trofin
 <mtrofin at google.com>,Mircea Trofin <mtrofin at google.com>,NAKAMURA Takumi
 <geek4civic at gmail.com>,Fangrui Song <i at maskray.me>,Phoebe Wang
 <phoebe.wang at intel.com>,Jacob Lalonde <jalalonde at fb.com>,Kareem Ergawy
 <kareem.ergawy at amd.com>,cmtice <cmtice at google.com>,Fangrui Song
 <i at maskray.me>,Iris <0.0 at owo.li>,Mats Jun Larsen <mats at jun.codes>,Mats Jun
 Larsen <mats at jun.codes>,Mariya Podchishchaeva
 <mariya.podchishchaeva at intel.com>,Pavel Labath <pavel at labath.sk>,Vladi Krapp
 <vladi.krapp at arm.com>,David Sherwood <david.sherwood at arm.com>,Igor Wodiany
 <igor.wodiany at imgtec.com>,Adrian Kuegel <akuegel at google.com>,Tobias Stadler
 <mail at stadler-tobias.de>,Florian Hahn <flo at fhahn.com>,gbMattN
 <matthew.nagy at sony.com>,Alaa Ali <alaa.leithy at gmail.com>,Durgadoss R
 <durgadossr at nvidia.com>,Jerry-Ge <jerry.ge at arm.com>,Florian Hahn
 <flo at fhahn.com>,Vladi Krapp <vladi.krapp at arm.com>,Sergio Afonso
 <safonsof at amd.com>,Paul Walker <paul.walker at arm.com>,JaydeepChauhan14
 <chauhan.jaydeep.ashwinbhai at intel.com>,Vy Nguyen <vyng at google.com>,Zahira
 Ammarguellat <zahira.ammarguellat at intel.com>,Baranov Victor
 <bar.victor.2002 at gmail.com>,Ilya Biryukov <ibiryukov at google.com>,Ilya
 Biryukov <ibiryukov at google.com>,Mariya Podchishchaeva
 <mariya.podchishchaeva at intel.com>,Nashe Mncube <nashe.mncube at arm.com>,Asher
 Mancinelli <ashermancinelli at gmail.com>,Matthias Springer <me at m-sp.org>,Justin
 Bogner <mail at justinbogner.com>,Aaron Ballman <aaron at aaronballman.com>,Ramkumar
 Ramachandra <ramkumar.ramachandra at codasip.com>,Nikita Popov
 <npopov at redhat.com>,Nashe Mncube <nashe.mncube at arm.com>,Nikita Popov
 <npopov at redhat.com>,David Spickett <david.spickett at linaro.org>,Florian Hahn
 <flo at fhahn.com>,=?utf-8?q?Gaëtan?= Bossu <gaetan.bossu at arm.com>,Peter
 Klausler <pklausler at nvidia.com>,Peter Klausler <pklausler at nvidia.com>,Peter
 Klausler <pklausler at nvidia.com>,Peter Klausler <pklausler at nvidia.com>,Peter
 Klausler <pklausler at nvidia.com>,Peter Klausler <pklausler at nvidia.com>,Peter
 Klausler <pklausler at nvidia.com>,Louis Dionne <ldionne.2 at gmail.com>,Craig
 Topper <craig.topper at sifive.com>,Evan Wilde <ewilde at apple.com>,Kevin Gleason
 <gleasonk at google.com>,Paschalis Mpeis <paschalis.mpeis at arm.com>,Matthias
 Springer <me at m-sp.org>,Snehasish Kumar <snehasishk at google.com>,Craig Topper
 <craig.topper at sifive.com>,Peter Klausler <pklausler at nvidia.com>,Felipe de
 Azevedo Piovezan <fpiovezan at apple.com>,Alexey Bataev <a.bataev at outlook.com>,Jan
 Svoboda <jan_svoboda at apple.com>,Jan Svoboda <jan_svoboda at apple.com>,Jan
 Svoboda <jan_svoboda at apple.com>,Nico Weber <thakis at chromium.org>,Nico Weber
 <thakis at chromium.org>,Nico Weber <thakis at chromium.org>,Aaron Ballman
 <aaron at aaronballman.com>,Florian Mayer <fmayer at google.com>,Luke Lau
 <luke at igalia.com>,Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Alexey
 Bataev <a.bataev at outlook.com>,Ryosuke Niwa <rniwa at webkit.org>,LLVM GN
 Syncbot <llvmgnsyncbot at gmail.com>,LLVM GN Syncbot <llvmgnsyncbot at gmail.com>,LLVM
 GN Syncbot <llvmgnsyncbot at gmail.com>,Maksim Panchenko <maks at fb.com>,Sirraide
 <aeternalmail at gmail.com>,Louis Dionne <ldionne.2 at gmail.com>,Aaron Ballman
 <aaron at aaronballman.com>,Ryosuke Niwa <rniwa at webkit.org>,Lei Huang
 <lei at ca.ibm.com>,Zahira Ammarguellat <zahira.ammarguellat at intel.com>,erichkeane
 <ekeane at nvidia.com>,Craig Topper <craig.topper at sifive.com>,Jonas Devlieghere
 <jonas at devlieghere.com>,Michael Jones <michaelrj at google.com>,Lei Huang
 <lei at ca.ibm.com>,Florian Hahn <flo at fhahn.com>,Eugene Epshteyn
 <eepshteyn at nvidia.com>,Craig Topper <craig.topper at sifive.com>,Jonas
 Devlieghere <jonas at devlieghere.com>,Craig Topper <craig.topper at sifive.com>,Finn
 Plummer <canadienfinn at gmail.com>,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Craig Topper
 <craig.topper at sifive.com>,Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Jorge
 Gorbe Moya <jgorbe at google.com>,Alex MacLean <amaclean at nvidia.com>,Jerry-Ge
 <jerry.ge at arm.com>,Ashley Coleman <ascoleman at microsoft.com>,Austin Schuh
 <AustinSchuh at users.noreply.github.com>,Andy Kaylor <akaylor at nvidia.com>,Un1q32
 <joey.t.reinhart at gmail.com>,Jonas Devlieghere <jonas at devlieghere.com>,alx32
 <103613512+alx32 at users.noreply.github.com>,Slava Zakharin
 <szakharin at nvidia.com>,Michael Jones <michaelrj at google.com>,Andy Kaylor
 <akaylor at nvidia.com>,Felipe de Azevedo Piovezan <fpiovezan at apple.com>,weiwei
 chen <weiwei.chen at modular.com>,Sudharsan Veeravalli <quic_svs at quicinc.com>,Fangrui
 Song <i at maskray.me>,jobhdez <lara2993 at proton.me>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Changpeng Fang <changpeng.fang at amd.com>,Reid Kleckner
 <rnk at google.com>,Fangrui Song <i at maskray.me>,Aiden Grossman
 <aidengrossman at google.com>,Aiden Grossman <aidengrossman at google.com>,Vlad
 Serebrennikov <serebrennikov.vladislav at gmail.com>,Mats Jun Larsen
 <mats at jun.codes>,David CARLIER <devnexen at gmail.com>,David CARLIER
 <devnexen at gmail.com>,Andreas Jonson <andjo403 at hotmail.com>,Andreas Jonson
 <andjo403 at hotmail.com>,Phoebe Wang <phoebe.wang at intel.com>,Florian Hahn
 <flo at fhahn.com>,James E T Smith <jamesETsmith at users.noreply.github.com>,Matthias
 Springer <me at m-sp.org>,Martin =?utf-8?q?Storsjö?= <martin at martin.st>,Antonio
 Frighetto <me at antoniofrighetto.com>,Florian Hahn <flo at fhahn.com>,Rahul Joshi
 <rjoshi at nvidia.com>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Louis
 Dionne <ldionne.2 at gmail.com>,Fangrui Song <i at maskray.me>,Fangrui Song
 <i at maskray.me>,Aiden Grossman <aidengrossman at google.com>,Fangrui Song
 <i at maskray.me>,Florian Hahn <flo at fhahn.com>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Owen Pan
 <owenpiano at gmail.com>,Fangrui Song <i at maskray.me>,Florian Hahn
 <flo at fhahn.com>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,junfengd-nv
 <junfengd at nvidia.com>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>
 =?utf-8?q?,?úngrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Alan <ahulambda at gmail.com>,weiwei
 chen <weiwei.chen at modular.com>,Owen Pan <owenpiano at gmail.com>,Fangrui Song
 <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Fangrui
 Song <i at maskray.me>,Fangrui Song <i at maskray.me>,Florian Hahn <flo at fhahn.com>,Hui
 <hui.xie1990 at gmail.com>,Sergei Barannikov <barannikov88 at gmail.com>,Benjamin
 Kramer <kramerb at google.com>,Florian Hahn <flo at fhahn.com>,Florian Hahn
 <flo at fhahn.com>,Benjamin Kramer <benny.kra at googlemail.com>,Florian Hahn
 <flo at fhahn.com>,Florian Hahn <flo at fhahn.com>,Sergei Barannikov
 <barannikov88 at gmail.com>,Fangrui Song <i at maskray.me>,Fangrui Song
 <i at maskray.me>,Florian Hahn <flo at fhahn.com>,Fangrui Song <i at maskray.me>,Florian
 Hahn <flo at fhahn.com>,Mats Jun Larsen <mats at jun.codes>,Phoebe Wang
 <phoebe.wang at intel.com>,Matheus Izvekov <mizvekov at gmail.com>,Alan
 <ahulambda at gmail.com>,Baranov Victor <bar.victor.2002 at gmail.com>,Zhen Wang
 <37195552+wangzpgi at users.noreply.github.com>,Matheus Izvekov
 <mizvekov at gmail.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Tobias Gysi
 <tobias.gysi at nextsilicon.com>,Fangrui Song <i at maskray.me>,Nikita Popov
 <npopov at redhat.com>,
=?utf-8?q?Balázs_Kéri?= <balazs.keri at ericsson.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Robert
 Imschweiler <robert.imschweiler at amd.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,
Juan Manuel Martinez =?utf-8?q?Caamaño?=,Mike <FruitClover at gmail.com>,Luke
 Lau <luke at igalia.com>,Simon Pilgrim <llvm-dev at redking.me.uk>,Simon Pilgrim
 <llvm-dev at redking.me.uk>,Simon Pilgrim <llvm-dev at redking.me.uk>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Matthias Springer <me at m-sp.org>,Han-Kuan
 Chen <hankuan.chen at sifive.com>,Abhishek Kaushik <abhishek.kaushik at intel.com>,Anatoly
 Trosinenko <atrosinenko at accesssoftek.com>,
Andrzej =?utf-8?q?Warzyński?= <andrzej.warzynski at arm.com>,Jack Frankland
 <jack.frankland at arm.com>,Anatoly Trosinenko <atrosinenko at accesssoftek.com>,Mel
 Chen <mel.chen at sifive.com>,Tom Eccles <tom.eccles at arm.com>,Aaron Ballman
 <aaron at aaronballman.com>,Jorn Tuyls <jtuyls at users.noreply.github.com>,Simon
 Pilgrim <llvm-dev at redking.me.uk>,Jay Foad <jay.foad at amd.com>,Zhaoxin Yang
 <yangzhaoxin at loongson.cn>,Uday Bondhugula <uday at polymagelabs.com>,Mats Jun
 Larsen <mats at jun.codes>,Christian Sigg <csigg at google.com>,Jay Foad
 <jay.foad at amd.com>,JaydeepChauhan14 <chauhan.jaydeep.ashwinbhai at intel.com>,Matthias
 Springer <me at m-sp.org>,
Andrzej =?utf-8?q?Warzyński?= <andrzej.warzynski at arm.com>,Krisztian
 Rugasi <Krisztian.Rugasi at hightec-rt.com>,Nashe Mncube <nashe.mncube at arm.com>,Farzon
 Lotfi <farzonlotfi at microsoft.com>,Pedro Lobo <pedro.lobo at tecnico.ulisboa.pt>,Asher
 Mancinelli <ashermancinelli at gmail.com>,Farzon Lotfi
 <farzonlotfi at microsoft.com>,Igor Wodiany <igor.wodiany at imgtec.com>,Farzon
 Lotfi <farzonlotfi at microsoft.com>,zhijian lin <zhijian at ca.ibm.com>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Justin Bogner <mail at justinbogner.com>,Michael
 Klemm <michael.klemm at amd.com>,Matheus Izvekov <mizvekov at gmail.com>,Joseph
 Huber <huberjn at outlook.com>,Julian Lettner <yln at users.noreply.github.com>,Alexandre
 Ganea <alex_toresh at yahoo.fr>,David Spickett <david.spickett at linaro.org>,Paul
 Kirth <paulkirth at google.com>,Rahul Joshi <rjoshi at nvidia.com>,Farzon Lotfi
 <farzonlotfi at microsoft.com>,Simon Pilgrim <llvm-dev at redking.me.uk>,Linux
 User <linuxusergd at outlook.com>,Mark de Wever <koraq at xs4all.nl>,Leandro
 Lupori <leandro.lupori at linaro.org>,Mark de Wever <koraq at xs4all.nl>,Lee Wei
 <lee10202013 at gmail.com>,Alexey Bataev <a.bataev at outlook.com>,Henry Jiang
 <h243jian at uwaterloo.ca>,Jerry-Ge <jerry.ge at arm.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Tai Ly <tai.ly at arm.com>,Jan Leyonberg
 <jan_sjodin at yahoo.com>,Simon Pilgrim <llvm-dev at redking.me.uk>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Rahul Joshi <rjoshi at nvidia.com>,vdonaldson
 <37090318+vdonaldson at users.noreply.github.com>,Nico Weber
 <thakis at chromium.org>,Pranav Kant <prka at google.com>,Jason Molenda
 <jmolenda at apple.com>,Jason Molenda <jmolenda at apple.com>,YLChenZ
 <chentongyongcz at gmail.com>,Piotr Fusik <p.fusik at samsung.com>,Brad Smith
 <brad at comstyle.com>,Leandro Lupori <leandro.lupori at linaro.org>,Drew Lewis
 <drew90 at vt.edu>,Damien L-G <dalg24 at gmail.com>,Lei Huang <lei at ca.ibm.com>,Rahul
 Joshi <rjoshi at nvidia.com>,Brad Smith <brad at comstyle.com>,LLVM GN Syncbot
 <llvmgnsyncbot at gmail.com>,LLVM GN Syncbot <llvmgnsyncbot at gmail.com>,LLVM GN
 Syncbot <llvmgnsyncbot at gmail.com>,LLVM GN Syncbot <llvmgnsyncbot at gmail.com>,Florian
 Hahn <flo at fhahn.com>,Shafik Yaghmour <shafik.yaghmour at intel.com>,Peter
 Collingbourne <pcc at google.com>,Nicolas van Kempen <nvankemp at gmail.com>,Krzysztof
 Drewniak <Krzysztof.Drewniak at amd.com>,vporpo <vporpodas at google.com>,Jason
 Molenda <jmolenda at apple.com>,Ramkumar Ramachandra
 <ramkumar.ramachandra at codasip.com>,Pedro Lobo <pedro.lobo at tecnico.ulisboa.pt>
 =?utf-8?q?,?=amansharma612 <amansharma6122002 at gmail.com>,Steven Hedges
 <steven.hedges34 at gmail.com>,Florian Hahn <flo at fhahn.com>,Sarah Spall
 <sarahspall at microsoft.com>,Philip Reames <preames at rivosinc.com>,Andres
 Chavarria <84650073+chavandres at users.noreply.github.com>,Mingming Liu
 <mingmingl at google.com>,Jerry-Ge <jerry.ge at arm.com>,Tom Yang
 <zhenyutyang at gmail.com>,Justin Bogner <mail at justinbogner.com>,Jorge Gorbe
 Moya <jgorbe at google.com>,Han-Kuan Chen <hankuan.chen at sifive.com>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Matheus Izvekov <mizvekov at gmail.com>,Rahul
 Joshi <rjoshi at nvidia.com>,Rahul Joshi <rjoshi at nvidia.com>,Rahul Joshi
 <rjoshi at nvidia.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Bruno
 Cardoso Lopes <bruno.cardoso at gmail.com>,Rahul Joshi <rjoshi at nvidia.com>,Rahul
 Joshi <rjoshi at nvidia.com>,Rahul Joshi <rjoshi at nvidia.com>,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Matheus Izvekov <mizvekov at gmail.com>,Peter
 Collingbourne <pcc at google.com>,Congcong Cai <congcongcai0907 at 163.com>,Mircea
 Trofin <mtrofin at google.com>,Mircea Trofin <mtrofin at google.com>,Naveen Seth
 Hanig <naveen.hanig at outlook.com>,Shoreshen <372660931 at qq.com>,Timm Baeder
 <tbaeder at redhat.com>,Timm Baeder <tbaeder at redhat.com>,Timm Baeder
 <tbaeder at redhat.com>,Aniket Lal <lalaniket8 at gmail.com>,Mingming Liu
 <mingmingl at google.com>,
=?utf-8?q?Csanád_Hajdú?= <csanad.hajdu at arm.com>,Nikita Popov
 <npopov at redhat.com>,Shoreshen <372660931 at qq.com>,Orlando Cazalet-Hyams
 <orlando.hyams at sony.com>,lorenzo chelini <l.chelini at icloud.com>,lorenzo
 chelini <l.chelini at icloud.com>,tianleliu <tianle.l.liu at intel.com>,Jerry-Ge
 <jerry.ge at arm.com>,Jerry-Ge <jerry.ge at arm.com>,Jerry-Ge <jerry.ge at arm.com>,Pedro
 Lobo <pedro.lobo at tecnico.ulisboa.pt>,Benjamin Maxwell
 <benjamin.maxwell at arm.com>,Stephen Tozer <stephen.tozer at sony.com>,David
 Green <david.green at arm.com>,David Green <david.green at arm.com>,
Nathan =?utf-8?q?Gauër?= <brioche at google.com>,Tom Eccles
 <tom.eccles at arm.com>,Tom Eccles <tom.eccles at arm.com>,Jakub Ficek
 <jakub.ficek at intel.com>,Jay Foad <jay.foad at amd.com>,Ramkumar Ramachandra
 <ramkumar.ramachandra at codasip.com>,MisakaVan
 <102937903+MisakaVan at users.noreply.github.com>,Anatoly Trosinenko
 <atrosinenko at accesssoftek.com>,Simon Pilgrim <llvm-dev at redking.me.uk>,Paul
 Walker <paul.walker at arm.com>,Simon Tatham <simon.tatham at arm.com>,Jonathan
 Thackray <jonathan.thackray at arm.com>,
Nathan =?utf-8?q?Gauër?= <brioche at google.com>,Nikolas Klauser
 <nikolasklauser at berlin.de>,Paul Walker <paul.walker at arm.com>,Kajetan
 Puchalski <kajetan.puchalski at arm.com>,Omair Javaid <omair.javaid at linaro.org>,Michael
 Klemm <michael.klemm at amd.com>,Florian Hahn <flo at fhahn.com>,Kajetan Puchalski
 <kajetan.puchalski at arm.com>,David Spickett <david.spickett at linaro.org>,Joseph
 Huber <huberjn at outlook.com>,Akshat Oke <Akshat.Oke at amd.com>,TatWai Chong
 <tatwai.chong at arm.com>,Kajetan Puchalski <kajetan.puchalski at arm.com>,Jay
 Foad <jay.foad at amd.com>,Nico Weber <thakis at chromium.org>,Alan Li
 <me at alanli.org>,wldfngrs <wldfngrs at gmail.com>,Hans Wennborg <hans at hanshq.net>
 =?utf-8?q?,?=Christian Sigg <csigg at google.com>,Romaric Jodin
 <rjodin at google.com>,Shilei Tian <i at tianshilei.me>,Mircea Trofin
 <mtrofin at google.com>,Michael Kruse <llvm-project at meinersbur.de>,Philip
 Reames <preames at rivosinc.com>,Han-Kuan Chen <hankuan.chen at sifive.com>,Han-Kuan
 Chen <hankuan.chen at sifive.com>,Michael Liao <michael.hliao at gmail.com>,Alan
 Li <me at alanli.org>,Christopher McGirr
 <7071833+chrsmcgrr at users.noreply.github.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Adrian Prantl <aprantl at apple.com>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Matt
 Arsenault <Matthew.Arsenault at amd.com>,Sirraide <aeternalmail at gmail.com>,Christian
 Sigg <csigg at google.com>,tdanyluk <tdanyluk at nvidia.com>,Congcong Cai
 <congcongcai0907 at 163.com>,Matt Arsenault <Matthew.Arsenault at amd.com>,Nuno
 Lopes <nuno.lopes at tecnico.ulisboa.pt>,Krzysztof Drewniak
 <Krzysztof.Drewniak at amd.com>,Stephen Tozer <stephen.tozer at sony.com>,Nico
 Weber <thakis at chromium.org>,Fangrui Song <i at maskray.me>,Matt Arsenault
 <Matthew.Arsenault at amd.com>,Nikolas Klauser <nikolasklauser at berlin.de>,Thurston
 Dang <thurston at google.com>,Dmitry Chestnykh <dm.chestnykh at gmail.com>,Simon
 Pilgrim <llvm-dev at redking.me.uk>,Stephen Tozer <stephen.tozer at sony.com>,Jacob
 Lalonde <jalalonde at fb.com>,Adrian Prantl <aprantl at apple.com>,k-kashapov
 <52855633+k-kashapov at users.noreply.github.com>,Morris Hafner
 <mmha at users.noreply.github.com>,Alexey Bataev <a.bataev at outlook.com>,Alexey
 Bataev <a.bataev at outlook.com>,Erich Keane <ekeane at nvidia.com>,Mark de Wever
 <koraq at xs4all.nl>,Fangrui Song <i at maskray.me>,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Matthias Springer <me at m-sp.org>,Min-Yih
 Hsu <min.hsu at sifive.com>,$x3 = ADDXrr $x0, $x0,Andy Kaylor
 <akaylor at nvidia.com>,Mark de Wever <koraq at xs4all.nl>,Louis Dionne
 <ldionne.2 at gmail.com>,Peter Collingbourne <pcc at google.com>,Stephen Tozer
 <stephen.tozer at sony.com>,Jonas Devlieghere <jonas at devlieghere.com>,Yijia Gu
 <yijiagu at google.com>,Matheus Izvekov <mizvekov at gmail.com>,Yijia Gu
 <yijiagu at google.com>,Dmitry Vasilyev <dvassiliev at accesssoftek.com>,Aaron
 Ballman <aaron at aaronballman.com>,Matthias Springer <me at m-sp.org>,Yijia Gu
 <yijiagu at google.com>,cmtice <cmtice at google.com>,Alexandre Ganea
 <alex_toresh at yahoo.fr>,Alexandre Ganea <alex_toresh at yahoo.fr>,Alexandre
 Ganea <alex_toresh at yahoo.fr>,Philip Reames <preames at rivosinc.com>,Adrian
 Prantl <aprantl at apple.com>,Michael Jones <michaelrj at google.com>,Florian Hahn
 <flo at fhahn.com>,Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,Jerry-Ge <jerry.ge at arm.com>,Adrian
 Prantl <aprantl at apple.com>,Jann <jannh at google.com>,lntue <lntue at google.com>,Douglas
 <Douglas.Gliner at sony.com>,"Oleksandr T." <oleksandr.tarasiuk at outlook.com>,Alexey
 Bataev <a.bataev at outlook.com>,Kevin McAfee <kmcafee at nvidia.com>,Alex MacLean
 <amaclean at nvidia.com>,Andy Kaylor <akaylor at nvidia.com>,Jason Molenda
 <jmolenda at apple.com>,Mircea Trofin <mtrofin at google.com>,
Valentin Clement =?utf-8?b?KOODkOODrOODsw=?=,k-kashapov
 <52855633+k-kashapov at users.noreply.github.com>,Cyndy Ishida
 <cyndy_ishida at apple.com>,Georgios Pinitas <georgios.pinitas at arm.com>,Peter
 Collingbourne <pcc at google.com>,Peter Collingbourne <pcc at google.com>,Peter
 Collingbourne <peter at pcc.me.uk>,Peter Collingbourne <pcc at google.com>,Peter
 Collingbourne <peter at pcc.me.uk>,"A. Jiang" <de34 at live.cn>,Sam Elliott
 <quic_aelliott at quicinc.com>,Mircea Trofin <mtrofin at google.com>,Mircea Trofin
 <mtrofin at google.com>,Chao Chen <chao.chen at intel.com>,Andres-Salamanca
 <andrealebarbaritos at gmail.com>,Craig Topper <craig.topper at sifive.com>,vdonaldson
 <37090318+vdonaldson at users.noreply.github.com>,brandb97
 <77328395+brandb97 at users.noreply.github.com>,lntue <lntue at google.com>,Dirk
 Pranke <dpranke at gmail.com>,Fangrui Song <i at maskray.me>,lntue
 <lntue at google.com>,Fangrui Song <i at maskray.me>,Fangrui Song <i at maskray.me>,tangaac
 <tangyan01 at loongson.cn>,Fangrui Song <i at maskray.me>,Liao Chunyu
 <chunyu at iscas.ac.cn>,Cyndy Ishida <cyndy_ishida at apple.com>,lntue
 <lntue at google.com>,Jason Molenda <jmolenda at apple.com>,Lakshay Kumar
 <lakshayk at nvidia.com>,Craig Topper <craig.topper at sifive.com>,Dmitry Vasilyev
 <dvassiliev at accesssoftek.com>,Jim Lin <jim at andestech.com>,Craig Topper
 <craig.topper at sifive.com>,Sudharsan Veeravalli <quic_svs at quicinc.com>,Sudharsan
 Veeravalli <quic_svs at quicinc.com>,Piotr Fusik <p.fusik at samsung.com>,Pavel
 Labath <pavel at labath.sk>,Chaitanya <Krishna.Sankisa at amd.com>,Madhur
 Amilkanthwar <madhura at nvidia.com>,Fangrui Song <i at maskray.me>,LLVM GN
 Syncbot <llvmgnsyncbot at gmail.com>,Steffen Larsen <steffen.larsen at intel.com>,yingopq
 <115543042+yingopq at users.noreply.github.com>,Luke Hutton
 <luke.hutton at arm.com>,Pavel Labath <pavel at labath.sk>,Stefan Schmidt
 <thrimbor at users.noreply.github.com>,David Sherwood <david.sherwood at arm.com>,NimishMishra
 <42909663+NimishMishra at users.noreply.github.com>,Simon Pilgrim
 <llvm-dev at redking.me.uk>,Samuel Tebbs <samuel.tebbs at arm.com>,Samuel Tebbs
 <samuel.tebbs at arm.com>,Samuel Tebbs <samuel.tebbs at arm.com>,Samuel Tebbs
 <samuel.tebbs at arm.com>,Samuel Tebbs <samuel.tebbs at arm.com>,Samuel Tebbs
 <samuel.tebbs at arm.com>,Samuel Tebbs <samuel.tebbs at arm.com>,Sam Tebbs
 <samuel.tebbs at arm.com>,Sam Tebbs <samuel.tebbs at arm.com>,Sam Tebbs
 <samuel.tebbs at arm.com>,Samuel Tebbs <samuel.tebbs at arm.com>
Message-ID:
In-Reply-To: <llvm.org/llvm/llvm-project/pull/133090 at github.com>


================
@@ -5039,10 +5039,26 @@ calculateRegisterUsage(VPlan &Plan, ArrayRef<ElementCount> VFs,
             // even in the scalar case.
             RegUsage[ClassID] += 1;
           } else {
+            ElementCount VF = VFs[J];
+            // The output from scaled phis and scaled reductions actually has
+            // fewer lanes than the VF.
+            if (isa<VPPartialReductionRecipe, VPReductionPHIRecipe>(R)) {
+              auto *ReductionR = dyn_cast<VPReductionPHIRecipe>(R);
+              auto *PartialReductionR = ReductionR ? nullptr : dyn_cast<VPPartialReductionRecipe>(R);
+              unsigned ScaleFactor = ReductionR ? ReductionR->getVFScaleFactor() : PartialReductionR->getVFScaleFactor();
+              VF = VF.divideCoefficientBy(ScaleFactor);
+            }
----------------
SamTebbs33 wrote:

Done.

https://github.com/llvm/llvm-project/pull/133090


More information about the llvm-branch-commits mailing list