Medicine (sometimes called healthcare) affects everyone, especially since our population is aging. The latest hardware brings a lot to the party in terms of performance to address the need for high responsiveness, significantly better accuracy and improved imaging capabilities. Most importantly, GPU computing has the power to accelerate and improve medical imaging to help medical professionals make better and faster decisions.
Thanks to NVIDIA's advancements in GPU technology, the industry is witnessing an evolution in how medical images are reconstructed. Taking the next step and actually implementing NVIDIA technology is where ADLINK shines. In fact, the GPU can convert the signal into a 2D or 3D image, which requires a complex series of calculations.
GPU solutions allow for Intelligent Video Analytics (IVA), which takes video data and turns it into actionable insights. For its part, ADLINK focuses on edge computing. Using GPU-accelerated edge computing improves image quality, accelerates image construction, and performs image analysis to help medical professionals treat patients. Ultimately, ADLINK is making medical imaging more widely used in clinical practice and surgery.
Clearly, no one company owns all the IP to build a successful platform. The partnership between ADLINK and NVIDIA enables scalability and future-proofing. GPUs — NVIDIA's specialty — clearly play an important role in medical imaging, especially with the company's GPU architecture and its various software stacks. ADLINK's embedded mobile PCI express modules (MXMs) and GPU computing platforms powered by NVIDIA GPUs provide computational acceleration for these enhanced imaging applications now and in the future.
The challenge of medical imaging
A time-tested way to customize a medical application platform is to use modules. These modules can accommodate any kind of processor to suit a specific application. For the applications mentioned here, ADLINK's embedded MXM GPU modules enable performance improvements not possible with traditional CPUs, including image processing and analytics, computational acceleration, and artificial intelligence.
The company's latest MXM GPU modules are designed with the NVIDIA Turing architecture, integrating CUDA cores, RT cores and Tensor cores in a single GPU. For example, the EGX-MXM-RTX3000 module features advanced NVIDIA Turing GPU technology in an MXM 3.1 Type B form factor that is one-quarter the size of a full-length PCI Express Graphics (PEG) card. Starting at 80 W of total graphics power, MXM GPU modules are suitable for mission-critical medical imaging applications where size, weight, and power are constrained, such as C-arms and other mobile devices.
Also keep in mind that, like industrial applications, the medical field is also looking for the longevity of its products. Two to five years is not uncommon, we often see a platform need 10 years of support. As a result, ADLINK is designing these systems using NVIDIA's embedded GPU solutions to give customers confidence that they will receive longer life support.
Finally, a feature implemented in the latest ADLINK systems is NVIDIA's GPUDirect Remote DMA (RDMA). In medical imaging applications, GPUDirect RDMA enables external data sources to directly access the GPU's memory. Without this feature, data would be copied into the CPU's memory before reaching the GPU, unnecessarily increasing data transfer latency and latency.
A typical use case for RDMA is ultrasound. In most cases, ultrasound utilizes a front-end device such as an FPGA to perform analog-to-digital conversion of the data before it reaches the GPU. This can be a lot of communication back and forth between the FPGA and GPU. With RDMA enabled, the increased bandwidth provides the necessary performance.
While the goal of better medical imaging is obvious, it is far from easy to achieve. The goal is simply to take the right action at the right time, in the right place, with the right data. The key to achieving this is combining domain knowledge with the platform and building blocks, and then tailoring these pieces to the individual needs of the end user.
Reviewing Editor: Guo Ting