[Beowulf] The computing Crunch for CFD at Exascale

Joshua Mora joshua_mora at usa.net
Tue Jun 10 10:58:19 PDT 2014

My 2 cents.

For CFD: From math point of view: hybrid Eulerian-Lagrangian formulation,
hybrid numerical + analytic models, automatic differentiation, interval
arithmetic for sensitive analysis, machine learning (artificial

For HPC (that includes CFD): From sw implementation point of view:
asynchronous computation, symbolic computation, fault tolerant, DSLs, dynamic
and hierarchical data structures.

For HPC: From hw point of view: non coherent SOC (others called it relaxed
coherency), transaction based. The network holds the data, or the processing
happens in the network (eg. like packet inspection/enrichment). Very much in
the same way as air is held in the pipes for pneumatic infrastructures.

------ Original Message ------
Received: 09:45 AM PDT, 06/10/2014
From: John Hearns <hearnsj at googlemail.com>
To: Beowulf Mailing List <beowulf at beowulf.org>
Subject: [Beowulf] The computing Crunch for CFD at Exascale

> We have a subscription to Aviation Week and Space Technology at McLaren,
> which I read at cofee break. Who knows when our rivals might launch a
> missile strike on us, or send a heat seeking missile up out exhausts? Best
> be ready I say.
> I found this very perceptive article on CFD algorithms and how things will
> have to change for exascale. Well digested, and well worth a read.

> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit

More information about the Beowulf mailing list