Index

A

affinity scheduling affinity smart priority scheduling algorithms barrier

ancestor tasks architectures CUDA, runtime Offload, runtime runtime Ayudame instrumentation plugins

B

barrier algorithms centralized plugins tree

base language breadth first scheduling build requirements Nanos++

C

centralized barrier child task cluster Point-to-point events options common parameters scheduling plugins compilation problems compile OmpSs OmpSs CUDA OmpSs OpenCL OmpSs shared-memory

configure flags Nanos++ construct CUDA compile OmpSs options runtime architectures

D

data environment declarative directive default scheduler dependence dependences plugins plugins examples, [1], [2] plugins perfect-regions plugins plain plugins regions regions troubleshooting

Dependency Graph instrumentation plugins descendant tasks device directive distributed breadth first scheduling dummy throttle throttling dynamic extent

E

empty trace instrumentation plugins Events Nanos State Events examples dependences plugins, [1], [2] throttle throttling

executable directive execute hybrid; MPI expression Extrae installation instrumentation plugins

F

FAQ NUMA; scheduling application crash bgq burst events macros performance regions tracking dependences

function task

G

general options

H

hardware counters histeresys throttle throttling

host device hybrid MPI OmpSs

I

idle threads throttle throttling immediate successor scheduling inline task installation Extrae Mercurium Nanos++ Ompss

instrument hybrid; MPI instrumentation plugins plugins Ayudame plugins Dependency Graph plugins Extrae plugins Task Sim plugins empty trace tdg

L

lexical scope

M

Mercurium common flags help installation vendor-specific flags

MPI OmpSs hybrid

N

Nanos++ build requirements configure flags installation nanox tool

NUMA scheduling number of tasks throttle throttling

O

Offload runtime architectures OmpSs CUDA, compile OpenCL, compile compile first program hybrid, MPI running shared-memory, compile

Ompss installation OpenCL compile OmpSs options CUDA cluster general runtime outline tasks

P

Paraver Point-to-point events (cluster), traces Point-to-point events, traces Paraver, Extrae, traces parent continuation scheduling parent task perfect-regions dependences plugins performance troubleshooting plain dependences plugins plugins Ayudame, instrumentation Dependency Graph, instrumentation Extrae, instrumentation Task Sim, instrumentation barrier common parameters, scheduling dependences empty trace, instrumentation examples, dependences, [1], [2] instrumentation perfect-regions, dependences plain, dependences regions, dependences scheduling throttle throttling

Point-to-point events cluster traces Paraver Point-to-point events (cluster) traces Paraver predecessor task priority scheduling problems compilation punctual events

R

ready task pool ready tasks throttle throttling regions dependences plugins

running OmpSs runtime architectures architectures CUDA architectures Offload options

S

scheduler default scheduling NUMA affinity affinity smart priority breadth first distributed breadth first immediate successor parent continuation plugins plugins common parameters priority threadmanager throttling versioning work first

shared-memory compile OmpSs sliceable task slicer policy structured block successor task

T

target device task task dependency graph task depth throttle throttling task generating code Task Sim instrumentation plugins tdg instrumentation thread threadmanager scheduling throttle dummy examples histeresys idle threads number of tasks plugins ready tasks task depth

throttling dummy examples histeresys idle threads number of tasks plugins ready tasks scheduling task depth tool nanox traces Paraver Point-to-point events Paraver Point-to-point events (cluster) tree barrier

V

versioning scheduling

W

work first scheduling