In the first part of this two-part post about high-performance computing, I already addressed three commonly-held myths associated with HPC. Now I’ll address three myths that are related to particular concerns about HPC adoption.
Myth #4: “Without internal IT support, HPC cluster adoption is undoable” Continue reading
You’ve heard all the talk about simulation-based design. You’ve listened to colleagues— maybe even some of your competitors — wax on about how doing robust simulation studies early on in the design cycle leads to more and better product ideas while also optimizing use of materials. In fact, you’re sold on the need to embrace advanced analysis, but you just don’t see how it’s feasible given the perceived complexity and cost of the simulation software — not to mention, the high-powered workstation gear. Continue reading
Looking back at my notes from conversations with many engineers during our recent ANSYS Convergence Conferences, I must admit that I still came across some myths and misconceptions about high-performance computing (HPC) for engineering simulation. Let me share six really striking ones with you:
- HPC is available on supercomputers only
- HPC is only useful for CFD simulations
- I don’t need HPC – my job is running fast enough
- Without internal IT support, HPC cluster adoption is undoable
- Parallel scalability is all about the same, right?
- HPC software and hardware are relative expensive
You may have heard about the grounding of an Alaskan oil rig in January, 2013. The 28,000-tonne rig was pushed toward the shore by waves up to 35 feet and winds up to 62 mph, dragging its main towing vessel and a tug behind it. There have been several such oil rig incidents over the past few decades. The below image shows the failure of an another oil rig platform due to extreme wave forces. A huge wave hitting the offshore platform leads to high wave impact loads that can eventually result in significant platform damage and collapse. These incidents can cause fatalities and damages that can cost hundreds of millions of dollars. Continue reading
The What-If studies that our software performs go a long way toward influencing future product development. With tons of features and their resulting design complexity, short development cycles, and consumer/regulatory demand for safety, innovative products call for computing tools — and most people think of CAD and CAE. The silent partner in this mix is the hardware and its ability to quickly and accurately perform the calculations.
Over the past few decades, engineering simulation providers have satisfied demanding design requirements by expanding and deepening their multiphysics capabilities. Now we can imagine what if … and we can conduct studies to test what if … But the truth is that any given simulation tool paired with the best hardware platform is no longer a recipe to fast, reliable solutions. The simulation tool itself must be designed to reap the hardware’s full potential, such as highly scalable and distributed HPC environments. Continue reading
ANSYS 15.0 contains a number of amazing achievements in the area of high performance computing (HPC) for the Mechanical APDL product. The performance is up to 5 times faster than previous releases, especially at higher core counts, by means of improved domain decomposition algorithms.
In addition, new parallel functionality was added in this release. One of the most important new features was the subspace eigensolver for vibration analyses, which supports distributed memory parallel and can be several times faster than the widely used block Lanczos eigensolver. Continue reading
As each week begins, I realize what a privilege it is to work with leading HPC technology providers like Intel, NVIDIA, Dell, HP, IBM and many others. Apart from the pleasant inter-social aspects of our weekly meetings, these collaborations enable us to provide simulation solutions optimized on the latest computing platforms. I strongly believe this is necessary because the computing landscape changes so quickly. Our customers want to take advantage of the latest HPC technologies and expand the scope of what they can accomplish with simulation.
One example of a strong partnership is NVIDIA. As a result of this partnership, ANSYS and NVIDIA have developed GPU-accelerated solvers and algorithms across our full range of multiphysics solutions. We were one of the first commercial engineering simulation providers to introduce structural mechanics support of GPU computing, and we released the first major commercial GPU-accelerated fluid dynamics solver of its kind with ANSYS 15.0. Continue reading
Are you happy with the way you access HPC resources? Some time back, I spoke with one of our customers about his daily experiences with his company’s HPC environment. He told me that the procedure to access the resources and dealing with the sessions is cumbersome and difficult to understand, especially for new employees.
I replied that ANSYS 15.0 comes with a solution for launching and managing batch jobs and remote visualization sessions based on a Web front end. That resonated well with my counterpart, such that the conversation evolved into a more detailed discussion on the individual features of the solution.
After the conversation, I was sharing my experience with some colleagues. We agreed that being able to quickly demonstrate the above mentioned features is essential for being able to communicate the advantages. However, as you can imagine, deploying such a tool is not done with the installation of the software, but requires some configuration, such as hooking up the submission tool with the compute resources. Even if the configuration step is designed to be comfortably done through the Web GUI itself, it will take some time until customers will be able use the tool and collect their own experiences. Continue reading