Making HPC Accessible for SMEs
Numerical computation, often also named simulation, plays nowadays in many enterprises an important role in the development process. Especially in large companies in the automotive or aerospace industry it is actually impossible to develop a new product without simulation technology. In recent decades the necessary know-how and personnel has been built up, but there have also been and will be significant investments into the infra structure. Investments of a size, that can easily be handled by large companies, can be a real challenge for an SME. Especially a Return on Investment (RoI) calculation is much more critical for smaller companies. At least as long as it is not clear that the RoI will come in an acceptably short time frame, given that in comparison to the company’s size the investment can be quite large. A powerful computer that has to be depreciated over several years must be loaded strongly in order to be worthy. In addition in smaller companies the usage of the infrastructure depends stronger on the sequence of projects. For them it is much more difficult to keep the load well balanced, which in turn makes the technology more expensive. This effect is visible even in larger companies. A third issue comes up especially at smaller enterprises in the scope of a regular usage, but for some users due to the kind of the problems they have to solve already at the beginning: what can be done if a numerical simulation is too big for the available hardware or would block it for days or weeks and make it thus unavailable for other users. This can for example happen, when a very fine mesh is needed; or for parameter studies where a certain computation has to be carried out very often wirh with slightly varying parameters.
https://www.dynamore.de/de/download/papers/2015-ls-dyna-europ/documents/sessions-d-5-8-1/making-hpc-accessible-for-smes/view
https://www.dynamore.de/@@site-logo/DYNAmore_Logo_Ansys.svg
Making HPC Accessible for SMEs
Numerical computation, often also named simulation, plays nowadays in many enterprises an important role in the development process. Especially in large companies in the automotive or aerospace industry it is actually impossible to develop a new product without simulation technology. In recent decades the necessary know-how and personnel has been built up, but there have also been and will be significant investments into the infra structure. Investments of a size, that can easily be handled by large companies, can be a real challenge for an SME. Especially a Return on Investment (RoI) calculation is much more critical for smaller companies. At least as long as it is not clear that the RoI will come in an acceptably short time frame, given that in comparison to the company’s size the investment can be quite large. A powerful computer that has to be depreciated over several years must be loaded strongly in order to be worthy. In addition in smaller companies the usage of the infrastructure depends stronger on the sequence of projects. For them it is much more difficult to keep the load well balanced, which in turn makes the technology more expensive. This effect is visible even in larger companies. A third issue comes up especially at smaller enterprises in the scope of a regular usage, but for some users due to the kind of the problems they have to solve already at the beginning: what can be done if a numerical simulation is too big for the available hardware or would block it for days or weeks and make it thus unavailable for other users. This can for example happen, when a very fine mesh is needed; or for parameter studies where a certain computation has to be carried out very often wirh with slightly varying parameters.