triton/searchindex.js

Search.setIndex({docnames:["index","installation/from-source","installation/packaged-binaries","tutorials/01-vector-add","tutorials/02-fused-softmax"],envversion:{"sphinx.domains.c":2,"sphinx.domains.changeset":1,"sphinx.domains.citation":1,"sphinx.domains.cpp":3,"sphinx.domains.index":1,"sphinx.domains.javascript":2,"sphinx.domains.math":2,"sphinx.domains.python":2,"sphinx.domains.rst":2,"sphinx.domains.std":1,nbsphinx:3,sphinx:56},filenames:["index.rst","installation/from-source.rst","installation/packaged-binaries.rst","tutorials/01-vector-add.ipynb","tutorials/02-fused-softmax.ipynb"],objects:{},objnames:{},objtypes:{},terms:{"0001":4,"0002":4,"0003":4,"0004":4,"0005":4,"0006":4,"0007":4,"0009":4,"0010":4,"0012":4,"0014":4,"0017":4,"0028":4,"0029":4,"003":3,"0037":4,"004":3,"016":3,"019":3,"020":3,"033":3,"070":3,"071":3,"1024":3,"1048576":3,"110":3,"128":4,"131072":3,"139":3,"142":3,"144":3,"16777216":3,"1823":4,"1984":3,"2019":3,"2097152":3,"20u":3,"262144":3,"2696":3,"286":3,"287":3,"2mn":4,"3076":3,"33554432":3,"3713":3,"3mn":4,"3us":3,"4096":4,"4194304":3,"4940":3,"524288":3,"568":3,"572":3,"6682":3,"781":4,"7mn":4,"8388608":3,"98432":3,"byte":4,"case":4,"class":[3,4],"float":[3,4],"function":[3,4],"import":[3,4],"int":[3,4],"return":[3,4],"true":[3,4],"void":[3,4],"while":1,For:3,One:3,The:[3,4],__expf:4,__global__:[3,4],_add:3,_kernel:[3,4],_softmax:4,_src:[3,4],abov:3,abs:3,access:3,achiev:3,activ:3,add:3,addit:[0,4],address:4,advantag:3,after:3,against:3,all:4,allclos:4,alloc:3,allow:3,also:[1,3],ani:4,api:[1,3],appli:[3,4],approxim:4,apt:1,around:4,arrai:3,assert:[3,4],attribut:3,autograd:[3,4],axi:4,back:[3,4],backward:3,bandwidth:[3,4],base:3,basic:3,becaus:4,below:3,bench:1,benchmark:1,best:3,between:3,binari:[0,1],block:[3,4],blockidx:3,bool:[3,4],bound:[3,4],build:1,cach:3,callabl:3,can:[1,2,3,4],caveat:3,cdiv:3,ceil:3,check:[3,4],chip:4,choos:3,chunk:3,citizen:3,clone:1,cmake:1,code:[3,4],codegen:1,cold:3,column:4,com:1,come:3,command:[1,2],comment:3,compil:3,cond:3,condit:3,config:1,consid:4,constraint:[3,4],construct:3,constructor:3,context:3,contigu:3,contrari:3,control:3,coordin:3,copi:3,could:4,creat:[3,4],ctx:[3,4],cuda:[3,4],current:4,custom:[1,3,4],data:[3,4],data_ptr:[3,4],dblock:3,declar:3,def:[3,4],def_m:4,defin:[3,4],denom:4,denomin:4,dereferenc:3,detail:3,dev:1,devic:[3,4],dict:[3,4],differ:[3,4],dir:1,direct:1,directli:[1,2],divis:3,dllvm_config:1,do_bench:[3,4],doe:[3,4],done:1,dram:4,driver:1,dtype:[3,4],dynam:3,each:[3,4],education:4,elapsed_tim:3,element:[3,4],element_s:4,elementwis:4,empty_lik:[3,4],enable_tim:3,end_ev:3,enough:3,equival:3,event:3,execut:3,exist:3,exp:4,expect:[3,4],exponenti:4,extens:3,f32_infin:4,far:4,fast:4,few:3,file:3,first:3,float32:[3,4],flow:3,follow:[1,2,4],forward:[3,4],found:1,from:[0,2,3,4],fuse:0,gbp:4,get:[1,3,4],get_program_id:[3,4],git:1,github:1,gpu:[3,4],greater:4,grid:[3,4],guard:4,handl:4,has:4,have:4,here:[1,3,4],high:3,highlight:3,how:3,http:1,implement:[3,4],increas:3,index:4,indic:4,initi:3,input:[3,4],instal:[0,1,2],instead:[3,4],instruct:0,integr:1,interest:1,intermedi:1,intern:1,just:[3,4],kei:4,label:4,lambda:[3,4],languag:3,larg:3,latenc:3,launch:[3,4],lazili:3,learn:3,legend:4,less:4,let:[3,4],like:[3,4],limit:4,line:3,llvm:1,load:[3,4],macro:3,mai:[1,3],maintain:3,make:[1,3,4],make_add_kernel:3,make_kernel:4,mani:3,manual_se:3,mapl:3,matplotlib:4,matrix:4,matter:3,max:[3,4],maximum:3,mean:4,memori:[3,4],method:3,min:4,mkdir:1,model:3,modifi:3,more:3,move:4,much:3,multipl:3,must:4,naiv:4,naive_softmax:4,nativ:3,necessari:4,need:[3,4],nelement:4,next_power_of_2:4,none:4,normal:4,note:[1,3,4],now:[3,4],num:4,number:[3,4],numer:4,object:3,offici:1,offset:3,onc:4,one:4,onli:[3,4],oper:[3,4],ops:3,opt:[3,4],other:3,our:[3,4],out:[3,4],output:3,overach:3,overhead:3,packag:0,paper:3,par:3,parallel:3,paramet:3,parser:1,past:3,path:1,per:4,perform:3,person:3,pid:3,pip:[1,2],plot:[1,4],plt:4,point:3,pointer:3,portabl:3,possibl:[3,4],power:4,practic:[3,4],pre:3,predic:4,prefer:3,primit:3,print:[3,4],process:3,processor:3,program:[3,4],properli:4,provid:1,ptillet:1,ptr:3,pyplot:4,pytest:1,python:3,pytorch:[3,4],rand:3,randn:4,rang:[3,4],read:4,reason:3,record:3,reduc:3,reduct:4,ref_m:4,refer:3,rep:3,represent:1,requir:4,result:[1,4],ret:[3,4],retriev:3,row:4,run:[1,3],runtim:1,see:3,seem:[3,4],sens:3,set:3,shape:[3,4],share:4,shortcut:3,show:4,similar:3,simpl:[3,4],singl:3,size:3,small:3,smaller:3,smallest:4,softmax:0,some:[3,4],someth:3,sourc:[0,3,4],spec:4,speed:4,spend:4,spmd:3,stabil:4,standard:3,start_ev:3,staticmethod:[3,4],store:[3,4],stride:4,stride_xm:4,stride_ym:4,string:3,sudo:1,sum:4,support:[1,4],sure:[3,4],synchron:3,syntax:[3,4],take:3,tensor:[3,4],ternari:3,test:1,than:4,thei:3,theoret:4,thi:[1,3,4],thing:3,those:[1,3],though:4,three:3,time:[3,4],time_m:3,tmp:1,torch_m:3,total:4,transform:3,tri_m:4,triton:[1,2,3,4],triton_m:3,tupl:3,tutori:[0,1,3],two:4,type:3,unit:1,usag:1,use:[1,3],used:3,using:3,val_fals:3,val_tru:3,valid:3,valu:[3,4],valuabl:4,variabl:3,vector:0,veri:4,want:[1,3,4],warmup:3,well:3,what:3,when:[3,4],which:[3,4],whose:3,wise:[3,4],without:3,won:4,word:3,work:[3,4],would:4,wrapper:3,wrote:4,x_max:4,xlabel:4,y_ref:4,y_tri:4,ylabel:4,you:[1,3,4],your:1},titles:["Welcome to Triton\u2019s documentation!","From Source","Packaged Binaries","Vector Addition","Fused Softmax"],titleterms:{addit:3,benchmark:[3,4],binari:2,bind:[3,4],comput:[3,4],document:0,from:1,fuse:4,kernel:[3,4],packag:[1,2],python:1,softmax:4,sourc:1,test:[3,4],torch:[3,4],triton:0,unit:[3,4],vector:3,welcom:0,write:[3,4]}})