The customer needed a scalable, user-friendly HPC setup for Abaqus and AutoCAD simulations on AWS
Simulation applications were running on standalone Windows/Linux VMs on AWS, without a cohesive HPC framework for orchestration or scaling.
The existing environment did not provide an elastic cluster capable of launching multiple compute nodes dynamically that is based on simulation demand.
Researchers needed a way to run Abaqus jobs, perform pre/post-processing, and access shared storage without relying on manual infrastructure intervention.
End-to-end AWS ParallelCluster–based HPC environment tailored for Abaqus engineering workloads.
A fully managed HPC environment was deployed using AWS ParallelCluster with a dedicated HeadNode and multiple SLURM queues. The design enabled structured, scalable execution of Abaqus workloads for different engineering profiles.
Compute nodes were configured to launch automatically when jobs are submitted & terminate upon completion. This ensured zero idle infrastructure & delivered true on-demand HPC elasticity.
Amazon FSx for NetApp ONTAP was implemented as the central shared file system for applications, input decks, & simulation output. It was seamlessly integrated with Amazon S3 for configuration, scripts, & persistent artifact storage.
The HPC environment was deployed in a private VPC with high-throughput networking & encrypted data pathways. Active Directory integration via LDAPS ensured secure, role-based access for all engineering users.
Cluster-wide monitoring was enabled using CloudWatch dashboards & logs for CPU, memory, I/O, & job health. Infrastructure-as-code templates & automation scripts ensured consistent provisioning, software setup, & operational repeatability.
The HPC POC was structured to deliver tangible improvements in performance, usability, and governance.
Submit Abaqus jobs to appropriately sized SLURM queues, automatically launching multiple compute nodes depending on job requirements.
Use FSx-backed shared storage to access input decks, store job outputs, and support pre- and post-processing on remote workstations.
Track node health, CPU/memory utilisation, network throughput, and job status through CloudWatch metrics and logs.
Leverage AD-integrated authentication, IAM roles, and KMS-backed encryption to enforce secure access and data protection across the HPC stack.
SLURM Scheduler
AWS KMS
Amazon S3
CloudWatch
AWS IAM
FSx Lustre
AWS Parallel Cluster
AWS LDAPS/AD Integration
Amazon Aurora Serverless MySQL
Learn how Apollo Tyres leveraged Tachyon to scale engineering simulations & drive innovation
