JuliaParallel / JuliaParallel/DistributedArrays.jl

DArray should be lazy/futures

Aperta
#209 2 commenti 0 reazioni 1 assegnatario Rivendicata da @vchuravy Vedi su GitHub
enhancement help wanted speculative
Lingua principale
Julia
Stelle
205
Fork
34
Metriche di merge delle PR
Nessuna PR unita negli ultimi 30g

Descrizione

One of the current issues with DArray is that each operation is immediately synchronizing. Requiring a distributed operation to finish before we can carry on with scheduling new operations. This simplifies the design,
but limits the scalability. Ideally we would want operations on DArray to be async/lazy similar to how CuArray works, and only synchronize on `show` and `convert`.

The major design issue here is to guarantee consistency. Operations need to appear to have executed in-order,
even though we might want to be able to execute reads out-of-order, but we will have to deal with updating data in-place. One idea might be to use vector clocks or look into how Fractal handles this or to run a consensus protocol to establish operations that can commit.

Guida per i contributori

Nessuna guida per i contributori indicizzata per questo repository

Valutazione

Questa issue non è ancora stata valutata.

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.