CFD Direct is pleased to announce the release of version 7.0 of CFD Direct From the Cloud™ (CFDDFC), the leading marketplace product for computational fluid dynamic (CFD) on Amazon Web Services (AWS). CFDDFC v7.0 includes:
- OpenFOAM v7 released by the OpenFOAM Foundation, produced and maintained by CFD Direct;
- ParaView v5.6.0 for data visualization;
- OpenMPI v3.1.4 with support for Elastic Fabric Adapter (EFA) for HPC workloads;
- FreeCAD v0.18.3 parametric 3D CAD modeler;
- … running on Ubuntu 18.04 LTS GNU/Linux operating system.
Support for C5 and C5n Instances
Amazon Elastic Cloud Compute (EC2) includes Compute Optimized instances for compute-intensive workloads. CFDDFC supports Elastic Network Adapter (ENA) and NVM Express (NVMe) used by C5 instances to target fast networking. C5n instances are a variant of C5 that provide significantly higher network performance across all instance sizes. The largest C5n instance,
c5n.18xlarge, contains 36 physical cores and has 100 Gbps of network bandwidth, compared to 25 Gbps with
HPC with C5n Instances
C5n instances can be clustered for HPC, delivering 70%-90% scaling at 504 cores. This performance is significantly better than standard C5 instances, which exhibit scaling below 50% for the same number of cores. Even with the higher price of ~25%, C5n instances are more cost-effective than C5 when clustering multiple instances.
New C5 Hardware
The standard C5 instances now include some newer custom 2nd generation Intel Xeon (Cascade Lake) processors with an higher clock speed than the previous processors. The new processors contain more cores, so new instance sizes of
c5.24xlarge have been introduced, containing 24 and 48 cores respectively. (If the user selects one of these new sizes, they are guaranteed the new hardware.) For single instance use, we recommend the standard C5 instances.
Support for Elastic Fabric Adapter
In November 2018, AWS announced Elastic Fabric Adapter (EFA), a network interface for HPC applications running on EC2. We previewed the EFA technology, which showed ~70% scaling at 1008 cores for meshes of 100 k cells per core and linear scaling at 1008 cores for a fixed mesh size of 97 m cells. EFA is now publicly available and CFDDFC v7.0 supports EFA, but currently only in the
Command Line Interface
cfddfc command line interface (CLI) manages instances of CFDDFC on AWS, transfers data to and from instances and runs provides convenient access to the instance by remote desktop (see above). It now includes a new
cluster sub-command to launch a cluster of instances. The user first launches a master instance, e.g.
c5n.18xlarge, that includes the required storage for the entire cluster, e.g. 300 GB:
cfddfc launch -instance c5n.18xlarge -volume 300
The user then creates a cluster by adding a specified number, e.g. 2, of slave instances of the same instance type, e.g.:
cfddfc cluster -slaves 2
This would create a cluster of 3 × 36 = 108 cores. The
launch command includes an
-efa option to attach the EFA to instances.
Monitoring Cost using the CLI
cfddfc CLI now includes a
cost sub-command to report all costs associated with running CFDDFC. It requires a short configuration through the AWS console, which is described by the instructions provided by the following command:
cfddfc config -cost
Once configured, the user can report previous costs by running:
which produces output containing the instance “key code”, costs and dates of deployment, e.g.:
Code Cost Period eo1MNXXX 0.04 2019-06-24 to 2019-06-26 jU2xHyaa 10.56 2019-06-27 to 2019-07-01 e2cXK1i9 18.32 2019-06-26 to 2019-07-04 i3TlNLpm 50.71 2019-07-03 to 2019-07-15