CAD

For programs like AutoCAD, Solidworks, Sketchup

Almost all CAD software, such as AutoCAD and Solidworks are predominantly single-threaded applications, meaning that the only utilise a handful of cores for the majority of tasks; with Rendering being the exception.

Therefore we recommend systems that are configured around a single CPU with a high core clock speed, with fast NVMe hard drives to optimise assemblies save/load time.

Graphics requirements are relatively light within CAD workflows, although Solidworks new graphics engine can take advantage of NVIDIA Quadro GPUs to massively improve viewport performance, especially in 3D model viewing.

For CAD workloads, we advise on using NVIDIA Quadro graphics cards. Although in most cases 'consumer-grade' cards perform comparatively, NVIDIA Quadro GPUs are tested, certified and supported by companies like Autodesk and Dassault.

BIM

For programs like AutoDesk Revit.

Much like traditional CAD work, the majority of tasks carried out in BIM packages such as Autodesk Revit, are all carried out on a single-core. Although support for multi-threaded processes is expanding, meaning that Revit can use up to 16 cores for a growing number of functions. (There is a full list of functions here)

We suggest systems that are configured around a single CPU with a high core clock speed, with fast NVMe hard drives to optimise assemblies save/load time, and a minimum of 16GB memory.

For BIM workloads, we advise on using NVIDIA Quadro graphics cards. Although in most cases 'consumer-grade' cards perform comparatively, NVIDIA Quadro GPUs are tested, certified and supported by companies like Autodesk and Dassault.

Photogrammetry

For programs like AutoDesk ReCap, Agisoft Metashape, Pix4D

Photogrammetry algorithms require computational processing power. Luckily popular software such as AutoDesk ReCap, AgiSoft Metashape and Pix4D are all multi-threaded applications, so we recommend a fast CPU with at least eight cores.

GPU acceleration is utilised heavily within Photogrammetry applications, with support for multiple GPUs within specific software. Although it has been shown that there is little performance increase in 'professional' GPUs, so we advise NVIDIA GeForce RTX GPUs over the more expensive Quadro cards.

Depending on the size of your dataset, we recommend a minimum of 32GB memory, with up to 128GB for large projects. Fast HDD read/write times are crucial to performance, so we recommend SSDs for OS and project drives.