DE · Topics · Resources · Sponsored · Sponsored Content


CAM Embraces GPU Acceleration

New CAM software releases leverage NVIDIA RTX GPU power to provide faster results.

New CAM software releases leverage NVIDIA RTX GPU power to provide faster results.

GPU acceleration can enable up to 10x faster material removal simulations. Image courtesy of NVIDIA.


Engineers have benefited from the use of GPU acceleration across a number of workflows, including CAD, rendering/visualization, and engineering simulation/analysis. At the IMTS 2024 conference in Chicago in September, the benefits of GPU were on display with a new group of software vendors – those in the CAM sector.

Both ModuleWorks and MachineWorks have announced GPU acceleration within their computer aided manufacturing solutions this year, and both were on hand at the IMTS show to discuss how GPUs (specifically, NVIDIA RTX GPUs) can save time and boost productivity.

ModuleWorks announced GPU acceleration features in August, which is expected to be made available for testing with the ModuleWorks 2024.12 software release. The new feature will be offered as part of the ModuleWorks Cutting Simulation software and can be integrated directly in existing CAM systems, enabling solution providers to upgrade to GPU-accelerated material removal simulation while retaining the full feature-set of the ModuleWorks simulation products.

In benchmark tests conducted at the ModuleWorks facilities in Aachen, Germany, the ModuleWorks Cutting Simulation with GPU processing was up to 10x faster than CPU-based simulations on mid range NVIDIA graphics cards, whereas more than 20 times accelleration has been observed with high-end NVIDIA graphic cards.

According to ModuleWorks Founder & Managing Director, Dr. Yavuz Murtezaoglu, the company is responding to the need for faster simulations that were constrained by the capabilities of the CPU. Rather than porting the CPU version of the software to support GPU compute, the company took the most time-consuming part of the simulation and developed new software code to natively calculate that on the GPU using OpenGL commands. 

ModuleWorks Founder & Managing Director, Dr. Yavuz Murtezaoglu, at the IMTS conference in Chicago.

 “We realized we could benefit from heavily optimized OpenGL [graphics standard] drivers and hardware from companies like NVIDIA, to convert our problems in a way that they can work on the GPU,” Murtezaoglu said. “We get the same accuracy, quality, and all of the features we have developed over 15 years, so we do not sacrifice any functionality, but we can run the simulations faster.”

Customers can enable the new features without any complex retrofitting; they just need to pay an additional license fee and make sure they have the right GPU.

ModuleWorks conducted initial testing with an NVIDIA GeForce GPU on a desktop system, and has plans to perform benchmarking on the professional-grade NVIDIA RTX GPUs. Both ModuleWorks and NVIDIA have facilities in Aachen, Germany, which has facilitated close collaboration.

“NVIDIA supports OpenGL and works out of the box with our software,” Murtezaoglu said. “The NVIDIA cards scale extremely well, and we expect additional boosts in performance as we move up to more powerful GPUs.”

NVIDIA continues to work with its OEM partners to help enable this type of performance on engineering workstations. “Combining professional NVIDIA RTX graphics with purpose-built Precision workstations from Dell demonstrates real-world customer performance benefits as ISVs like ModuleWorks and MachineWorks create new functionality leveraging GPUs to accelerate workflows.” says Matt Allard, Director of Strategic Alliances at Dell Technologies.

MachineWorks announced MachineWorks GPU in early September, a new software release that takes advantage of the high numbers of cores in modern GPUs to accelerate overall simulation performance. It is targeted for release in 2025 as part of MachineWorks 9.0.

Examples of MachineWorks GPU simulation results. Image courtesy of MachineWorks.

“Having developed early prototypes more than a decade ago, we have been watching the development of GPU hardware closely,” said Dr. Fenqiang Lin, managing director of MachineWorks, in a press release. “GPU has become more mature and, with advances in AI LLMs [artificial intelligence large language models], it is clear that massively parallel hardware will become a commodity item.” 

According to Richard Baxter, head of sales and marketing for MachineWorks, the company has adopted GPU-enabled CNC simulation based on customer demand and because there is a clear benefit in speed. The software uses simulation to produce a stock model that customers can then utilize for the toolpath generation process, and Baxter says that the company has seen between a 10X and 25X speedup (depending on which GPU is being used) of those processes compared to using the CPU alone. 

“We can see a reduction from several minutes down to just a few seconds, subject to the constraints of accuracy,” Baxter said.

The company has transitioned one of its geometry engines to GPU compute and has benchmarked performance on the NVIDIA RTX™ 2000 Ada Generation GPU as well as NVIDIA GeForce RTX products.

At IMTS, there was also significant activity around the use of artificial intelligence (AI) in the CAM and CNC programming space, and Murtezaoglu at ModuleWorks says he hopes that the work NVIDIA is doing in both AI and engineering will benefit the CAM software market. “We hope that it will drive technology even faster and further because of the big investments being made in AI, and it will raise the bar even higher,” he said.

More NVIDIA Coverage

Share This Article

Subscribe to our FREE magazine, FREE email newsletters or both!

Join over 90,000 engineering professionals who get fresh engineering news as soon as it is published.


#29420