Commit 26f07e23 authored by THIBAULT Samuel's avatar THIBAULT Samuel

emphasize that MPI communications can be either explicit or implicit, and we...

emphasize that MPI communications can be either explicit or implicit, and we have master-slave support coming
parent fa5833a6
......@@ -217,16 +217,28 @@ negligible.
<h4>Clusters</h4>
<p>
To deal with clusters, StarPU can nicely integrate with <a href="doc/html/MPISupport.html">MPI</a> through
either explicit network communications, which will then be <b>automatically combined and
overlapped</b> with the intra-node data transfers and computation, or the application
can also just provide the whole task graph, a data distribution over MPI nodes, and StarPU
will automatically determine which MPI node should execute which task, and
<b>automatically generate all required MPI communications</b> accordingly (new in v0.9). We
have gotten excellent scaling on a 256-node cluster with GPUs, we have not yet
had the opportunity to test on a yet larger cluster. We have however measured
that with naive task submission, it should scale to a thousand nodes, and with
pruning-tuned task submission, it should scale to about a <b>million nodes</b>.
To deal with clusters, StarPU can nicely integrate with <a
href="doc/html/MPISupport.html">MPI</a>, through explicit or implicit
support, according to the application's preference.
<ul>
<li>Explicit network communication requests can be emitted, which will
then be <b>automatically combined and overlapped</b> with the intra-node data
transfers and computation,
<li>The application can also just provide the whole task graph, a
data distribution over MPI nodes, and StarPU will automatically determine which
MPI node should execute which task, and <b>automatically generate all required
MPI communications</b> accordingly (new in v0.9). We have gotten excellent
scaling on a 256-node cluster with GPUs, we have not yet had the opportunity
to test on a yet larger cluster. We have however measured that with naive task
submission, it should scale to a thousand nodes, and with pruning-tuned task
submission, it should scale to about a <b>million nodes</b>.
<li>Starting with v1.3, the application can also just provide the
whole task graph, and let StarPU decide the data distribution and task
distribution, thanks to a master-slave mechanism. This will however by nature
have a more limited scalability than the fully distributed paradigm mentioned
above.
</ul>
</p>
<h4>Out of core</h4>
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment