Visualise via the cloud
By The Record on January 17, 2017
Filed under Discrete Manufacturing
To meet a range of simulation and visualisation challenges, many manufacturers are turning to high performance computing and graphics accelerated application technologies. When used alongside cloud technologies, new levels of flexibility are attainable
Today’s manufacturers and engineers are facing a series of challenges around design complexity, compliance and market pressure – the biggest of which is around increasing speed and the pace of innovation.
The market is currently incredibly dynamic, and there is pressure to provide exceptional value at a higher rate. Enterprises need a solid roadmap and must look to protect their futures, as changes can be costly and time consuming. This increases the demand on original equipment manufacturers and original design manufacturers to create innovative platforms that can grow going forward.
For product-driven companies, simulating and visualising product design is critically important for achieving certification, compliance and other standards. It is also key to innovation and optimisation. For example, if an engineer is working on reducing the size of an electronics enclosure, simulating and visualising heat transfer and air flow can provide real time feedback to determine the amount of size reduction that is possible.
For certain applications, simulations can be performed during the design process easily, and the computing requirement is relatively low. However, more complex simulations must run within a supercomputing infrastructure to return results in a timely manner.
Today, most engineers must wait in a queue for their simulation to run within their company’s
supercomputing infrastructure, due to capacity and availability constraints. Simulation becomes a validation tool as opposed to a design tool. This is leading to a greater demand for high performance computing (HPC) and graphics accelerated applications (GAA).
Microsoft Azure provides a selection of virtual machines that are tailored for this. For HPC, virtual machines such as A8, A9, H16r, H16mr and NC24r may be clustered using RDMA to build any size of supercomputer, from 32 cores to 32,000 or higher. For GAA, virtual machines like NV6, NV12 and NV24 provide increasing compute, memory and storage capacity with the advantage of NVIDIA GRID capability. This allows applications such as computer aided design, rendering and animation, and other graphics-intensive professional applications, to run at optimal performance and be accessed via a remote desktop connection.
Simulation and visualisation tools have become essential to how designers and engineers work, and critical for ensuring compliance with governing standards. The real difference exists in how these capabilities are used. It’s common for most businesses to have on-premise infrastructure for HPC and GAA, but an increasing number are moving to the cloud to obtain the flexibility they need to be responsive.
In the majority of cases, performance is superior in the cloud when compared with on-premise implementations. Microsoft Azure offers the latest in NVIDIA GRID capabilities and higher levels of CPU and memory configuration. Designers and engineers can work remotely and easily, because their device is no longer a limiting factor.