Talk:Checkpoint synchronization: Difference between revisions
Communication phase
(No communication phase?) |
(Communication phase) |
||
Line 1:
I'm trying to understand the problem so I can adapt it into my language. Do I correctly understand that there is no point at which the tasks communicate with each other ''during'' the checkpoint? The ''only'' property produced by this concurrency mechanism is that no worker shall begin work iteration #n before everyone has finished #(n-1)? —[[User:Kevin Reid|Kevin Reid]] 22:59, 10 August 2010 (UTC)
: That's what it looks like to me. There are scenarios where that's the normal model (most notably GPGPU programming; the checkpoint is needed for hardware instruction resync.). In languages where data structure syncronization is implicit, I don't think a communication phase is necessary. Otherwise, (if I were writing in C++, for example), the communication phase might get used for work item distribution and condition checks. --[[User:Short Circuit|Michael Mol]] 00:37, 11 August 2010 (UTC)
| |||