mirror of
https://github.com/AdaCore/PolyORB.git
synced 2026-02-12 13:01:15 -08:00
26 lines
1.2 KiB
Plaintext
26 lines
1.2 KiB
Plaintext
This example describes how to use the DSM storage support. This
|
|
storage support runs one of the Li and Hudak's algorithms. Therefore,
|
|
a shared passive package using this storage support cannot be
|
|
configured on a passive partition but on an active
|
|
partition. Moreover, the user has to prevent GLADE to use a light PCS
|
|
in order to run this algorithm. This cannot be automatically detected
|
|
as it is specific to this storage support.
|
|
|
|
Shared passive package Vector and RCI package Scheduler are configured
|
|
on part1. We use dsm:// as a data storage support for Vector. Client
|
|
and Server make concurrent computations of a shared vector. We are
|
|
going to run one instance of part1 and three instances of part2. Each
|
|
of these four partitions increments a different block of vector with
|
|
its partition id. Then, each partition does the same computation on
|
|
the next block. Before moving to the next block, the partitions
|
|
synchronize using Scheduler. At the end, each block has been modified
|
|
the same way and vector has been massively and concurrently modified.
|
|
|
|
part1 and part2 are active partitions.
|
|
|
|
Run in four separate sessions, one with part1 and three with part2.
|
|
As a result, all the vector item values are set to 10.
|
|
|
|
|
|
|