Novatech CAD workstations

What makes a CAD workstation?

• CPU CPUs in workstations tend to be optimised towards larger core counts and overall slower clock speeds than those utilised in high-end desktops. This is due to the increasing need to parallelise workflows in simulation and render activities. These workloads increasingly prioritise multithreaded performance rather than a single high-speed thread.
• Memory CAD workstations have several requirements when it comes to RAM size and type. The most important of these is the size in comparison to the model or object that is being worked with. If the size of the model that is being manipulated is too large to be held entirely in RAM, this forces the program to page data either to a paging file or to a scratch drive. While this is often a high-speed storage drive such as an SSD, this is an order of magnitude slower than accessing the model in RAM. To allow greater flexibility, workstations often have a minimum of 16GB of RAM although they often scale to 64 or even 128GB to enable efficient working.
  
  For CAD professionals we highly recommend the use of Error Correcting Code RAM, a server and workstation variant of standard RAM that has inbuilt logic to check and correct most forms of in memory corruption. This has the dual benefit of preventing mistakes in simulated models such as Solidworks and also increasing the stability and reliability of any system using it, reducing the likelihood of crashes or unforced errors.
• GPU Key to performance in CAD applications, GPUs can be judged on the combination of two factors, their Frame Buffer and CUDA core count. In this case, much like with RAM, high amounts of memory are required to efficiently and effectively deal with more complex and potentially larger models, while the CUDA cores and clock speed determine the rate at which changes to these models are processed, ie manipulation or transformation of the subject.
  
  While those that are budget conscious may see the value of using a high-end gaming card, as these have similar CUDA counts and buffers, professional grade GPUs provide more reliability. Via a combination of driver packages customised for professional workflows, internal software validation (ISV) certification , and components and cards engineered for 24/7 loads over long periods of time. Other highly advanced feature sets are only available on Quadro cards, particularly in the rendering space.