ࡱ> +-**bjbjUU G??*-----AAAA M A YYYYYYo {$-YY--YY-Y-YYP?=A0 HHH-, H :Abstract GPUs have established a new baseline for power efficiency and computing power, delivering larger bandwidth and more computing units in each new generation. Modern GPUs support the concurrent execution of kernels to maximize resource utilization, allowing kernels to better exploit idle resources. However, the decision on the simultaneous execution of different kernels is made by the hardware, and sometimes GPUs do not allow the execution of remaining blocks of other kernels, even with the availability of resources. In this work, we present an in-depth study on the simultaneous execution of kernels in the GPU. We present the necessary conditions for executing kernels simultaneously, we list the factors that influence competition and propose a model that describes performance degradation. Finally, we validate the model using kernels of real-world applications with different use intensities of computation and memory. Keywords: Concurrent Kernels; Multiprogramming; GPU computing. Resumo As GPUs estabeleceram uma nova linha de base em relao eficincia de energia e capacidade de computao, oferecendo maiores larguras de banda e mais unidades de computao em cada nova gerao. As GPUs modernas suportam a execuo simultnea de kernels para maximizar a utilizao dos recursos, permitindo que kernels possam explorar melhor os recursos ociosos. Entretanto, a deciso da execuo simultnea de kernels diferentes realizada pelo hardware e muitas vezes as GPUs no permitem a execuo de blocos remanescentes de outros kernels, mesmo com a disponiblidade de recursos. Neste trabalho, realizamos um estudo aprofundado sobre a execuo simultnea de kernels na GPU. Apresentamos as condies necessrias para executar kernels simultaneamente, listamos os fatores que influenciam a concorrncia e propomos um modelo que pode determinar a reduo de desempenho. Finalmente, validamos o modelo utilizando kernels de aplicaes reais com diferentes intensidades de computao e uso de memria. Palavras-chave: Kernels concorrentes; Multiprogramao; computao em GPU.   *置"h*h:v5CJ\aJmHsHh*h:vCJaJmHsH%h*h:vB*CJaJmHph"""sHh:vCJaJh*h:vB*CJaJph"""h*h:vCJaJh*h:v5CJ\aJ   *dhgd* $dha$gd* $dha$gd*E0P0p1h:p}[/ =!"#$% Dpn  666666666666666 666666666666666666666666666 6666666666 666666666666 6666666666666666666666666666666666666666666666666666666666666666662 0@P`p2( 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p8XV~$OJPJQJ^J_HmHnHsHtHT`T }[Normal d!B* CJ_HaJmH ph sH tH DA D 0Default Paragraph FontRiR 0 Table Normal4 l4a (k ( 0No List JJ }[0Heading $xCJOJQJ^JaJ8B8 }[0 Body Text d J/J nF*0Body Text CharB* mH ph sH tH $/"$ }[0ListD"2D }[0Caption  $xx6CJ]aJ*B* }[0Index $J>RJ }[0Title $xCJOJPJQJ^JaJd/ad nF* Title Char75B* CJ KHOJPJQJ\^JaJ mH ph sH tH ,r, }[0ndice $PK![Content_Types].xmlj0Eжr(΢Iw},-j4 wP-t#bΙ{UTU^hd}㨫)*1P' ^W0)T9<l#$yi};~@(Hu* Dנz/0ǰ $ X3aZ,D0j~3߶b~i>3\`?/[G\!-Rk.sԻ..a濭?PK!֧6 _rels/.relsj0 }Q%v/C/}(h"O = C?hv=Ʌ%[xp{۵_Pѣ<1H0ORBdJE4b$q_6LR7`0̞O,En7Lib/SeеPK!kytheme/theme/themeManager.xml M @}w7c(EbˮCAǠҟ7՛K Y, e.|,H,lxɴIsQ}#Ր ֵ+!,^$j=GW)E+& 8PK!Ptheme/theme/theme1.xmlYOo6w toc'vuر-MniP@I}úama[إ4:lЯGRX^6؊>$ !)O^rC$y@/yH*񄴽)޵߻UDb`}"qۋJחX^)I`nEp)liV[]1M<OP6r=zgbIguSebORD۫qu gZo~ٺlAplxpT0+[}`jzAV2Fi@qv֬5\|ʜ̭NleXdsjcs7f W+Ն7`g ȘJj|h(KD- dXiJ؇(x$( :;˹! I_TS 1?E??ZBΪmU/?~xY'y5g&΋/ɋ>GMGeD3Vq%'#q$8K)fw9:ĵ x}rxwr:\TZaG*y8IjbRc|XŻǿI u3KGnD1NIBs RuK>V.EL+M2#'fi ~V vl{u8zH *:(W☕ ~JTe\O*tHGHY}KNP*ݾ˦TѼ9/#A7qZ$*c?qUnwN%Oi4 =3ڗP 1Pm \\9Mؓ2aD];Yt\[x]}Wr|]g- eW )6-rCSj id DЇAΜIqbJ#x꺃 6k#ASh&ʌt(Q%p%m&]caSl=X\P1Mh9MVdDAaVB[݈fJíP|8 քAV^f Hn- "d>znNJ ة>b&2vKyϼD:,AGm\nziÙ.uχYC6OMf3or$5NHT[XF64T,ќM0E)`#5XY`פ;%1U٥m;R>QD DcpU'&LE/pm%]8firS4d 7y\`JnίI R3U~7+׸#m qBiDi*L69mY&iHE=(K&N!V.KeLDĕ{D vEꦚdeNƟe(MN9ߜR6&3(a/DUz<{ˊYȳV)9Z[4^n5!J?Q3eBoCM m<.vpIYfZY_p[=al-Y}Nc͙ŋ4vfavl'SA8|*u{-ߟ0%M07%<ҍPK! ѐ'theme/theme/_rels/themeManager.xml.relsM 0wooӺ&݈Э5 6?$Q ,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-![Content_Types].xmlPK-!֧6 +_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!Ptheme/theme/theme1.xmlPK-! ѐ' theme/theme/_rels/themeManager.xml.relsPK] ** * _GoBack,,}[:v(*@*,@*@@UnknownG*Ax Times New Roman5Symbol3. *Cx Arial7.@CalibriS& xP!Liberation SansArialG.R<(Microsoft YaHei7@CambriaACambria Math"#Zg#ZgF!0$P*}[!xxAbstractRommel Anatoli Quintanilla CruzHelioOh+'0 , P \ ht| Abstract Rommel Anatoli Quintanilla CruzNormal_WordconvHelio2Microsoft Office Outlook@0@t?=@t?=F՜.+,D՜.+,4 hp|    Abstract TitleH AppVersion DocSecurityHyperlinksChangedLinksUpToDate ScaleCrop ShareDoc00.0001   !#$%&'(),Root Entry F/W?=.1Table HWordDocumentGSummaryInformation(DocumentSummaryInformation8"CompObjy  F'Microsoft Office Word 97-2003 Document MSWordDocWord.Document.89q  F#Documento do Microsoft Office Word MSWordDocWord.Document.89q