Our Process

Comprehensive HPC & AI System Testing

With decades of experience in all phases of HPC, Microway’s production and integration teams ensure your workstations, servers, clusters, storage, networking and their software arrive fully-tested and ready-to-run. All systems are thoroughly tested at our factory before shipment to your site.

Our testing process considers each component in the system to verify operation and reliability. Primary areas of focus include the processors (CPUs), memory, GPUs, storage and network fabric. Our test suite includes applications that have been shown to cause faults in the field, as well as detect infant mortality. We look at much more than a simple on/off test.

Then the operating system, compilers, SDKs, and application software you specify are installed and configured before final testing. Clusters are made to run complete jobs that span all nodes and accelerators. Thus, the full hardware+software stack is validated. System cooling and power supplies are also stress-tested.

Depending upon the complexity and size of your order, systems will spend between one and four weeks in our test area. All components must withstand sustained 100% load for multiple days before they are approved for final QC and shipment to customers.

Typically, systems are boxed and shipped on pallets separately from the rackmount cabinet(s). If desired, Microway can provide turn-key delivery of fully-integrated and racked systems. Factory pre-wiring of the cabinets and/or onsite installation services are available for an additional fee.