TotalCAE sponsored and presented at the CONVERGE Global User Conference September 19th-23rd 2022.

We had a client present on their work, and had a great time seeing our other CONVERGE clients again, including friends at Convergent Science.

As part of the conference we have added a free course to TotalCAE University on how to submit, monitor, and manage CONVERGE jobs to TotalCAE HPC clusters, and TotalCAE on Cloud Infrastructure from AWS and Azure:

https://courses.totalcae.com/courses/submitting-to-totalcae-with-converge

The biggest news from the conference was Convergent Science announced their Horizon cloud platform for CONVERGE joining the other ISV’s with a direct cloud offering. Horizon may be a good solution if you need pay as you go CONVERGE licensing. Some advantages TotalCAE has when considering a cloud strategy for your CAE department:

  1. We host in your company’s cloud subscription, so you can pay the lowest possible costs for cloud infrastructure directly to AWS or Azure, starting at $2.88 an hour for 96 cores.
  2. Run large scale jobs with multiple nodes connected with InfiniBand or AWS EFA. See InfiniBand benefits for CONVERGE here.
  3. Hosting in your cloud subscription enables transparent access to your existing paid CAE/CFD licensing.
  4. Hosting in your cloud environment enables IT control and visibility into security and possible higher bandwidth downloads if you have Azure Express Route or AWS Direct Connect.
  5. Support 100’s of CAE and CFD applications with one solution, and one place to go for IT CAE support.
  6. Have an identical interface and workflow for TotalCAE on-prem (for lowest constant workload cost) and cloud, with one vendor supporting all your CAE applications for all your workloads on-prem in your data center or in your cloud subscription.
  7. Support for all the major HPC cloud infrastructure providers.
  8. Support for integrated cloud desktops on-demand starting at $1.34 per hour for pre/post processing to avoid the cloud download penalty.

Below is a comparison of various CONVERGE jobs and their relative cost on different types of AMD compute hardware.

ModelHardwareJob Cost ( Hardware Only On-Demand)
NASA_low-speed_fan_MRF_steady_RANSAWS hpc6a.48xlarge AMD
– 96 cores
$0.58
NASA_low-speed_fan_MRF_steady_RANSCONVERGE Hosted OCI BM.Standard.E4.128 AMD  – 128 cores$0.85
NASA_low-speed_fan_MRF_steady_RANSAzure  Standard HB120rs v3  AMD
– 120 cores
$0.54

ModelHardwareJob Cost ( Hardware Only On-Demand)
Pouch_battery_liquid_cooled_CHT_steady_laminarAWS hpc6a.48xlarge AMD
– 96 cores
$7.10
Pouch_battery_liquid_cooled_CHT_steady_laminarAzure  Standard HB120rs v3  AMD
– 120 cores
$8.97

ModelHardwareJob Cost ( Hardware Only On-Demand)
SI8_engine_premix_G-Equation_transient_RANSAWS hpc6a.48xlarge AMD
– 96 cores
$17.33
SI8_engine_premix_G-Equation_transient_RANSCONVERGE Hosted OCI BM.Standard.E4.128 AMD 
– 128 cores
$22.55
SI8_engine_premix_G-Equation_transient_RANSAzure  Standard HB120rs v3  AMD
– 120 cores
$20.14

For the raw data head over to https://caebench.org