The offshore oil and gas industry leverages AI to monitor complex topside and subsea operations and analyze sensor data to uncover suboptimal operations and impending problems, including equipment failures, before they occur. AI computing analytics for seismic processing and interpretation in the field is used to decide where to drill for oil. Reservoir modeling is used in constructing computer models of petroleum reservoirs in order to improve the estimation of the oil reserves and to aid in the development of the fields. Ocean based oil rigs collect vast amounts of IoT data that needs to be processed at the edge in a harsh remote environment for real time action in response to emergency conditions. Oil refineries use IoT data and real time AI to shut down industrial equipment autonomously to prevent catastrophic failures. 

AI is being deployed on fleets of mining and construction equipment to support hybrid autonomous operation in harsh and potentially dangerous environments. Data is captured from a wide range of sensors to aid in the optimal functioning of mining operations using AI algorithms to provide predictive intelligence and guide exploration decisions. Large scale construction projects utilize AI applications to monitor fleets of heavy equipment and provide predictive maintenance guidance as well as deployment of autonomous or semi-autonomous construction equipment in the field. Agriculture is increasingly automated and connected. The introduction of private 5G radio services will be leveraged to bring autonomous farm vehicles, smart irrigation, smart harvesting, and herd monitoring to the modern farm. These systems rely on thousands of sensors on the edge and will require local “transportable” equipment for efficiency of use.  

Healthcare demands new computing paradigms to meet the need for personalized medicine, enhanced quality of care, and breakthroughs in biomedical research to treat disease. Artificial Intelligence is becoming a critical component of many of the medical systems deployed to deliver on these needs. AI-powered tools can be an extra set of "eyes," helping clinicians to quickly read images, monitor changes, and identify urgent findings to optimize workflows and enhance patient care. Medical imaging and diagnostic equipment hosting AI applications process and analyze scans for faster, more accurate and more granular diagnoses. High end medical equipment used for things like X-ray, ultrasound and eye surgery require significant processing power and are normally contained within the “transportable” equipment to allow them to be moved from one room to the next and are dependent on hardware which fits in the limited space and power of the medical environment.   

The appliance supports up to 4 NVIDIA A100 PCIe GPUs which deliver 2.5x FP64 performance compared to the NVIDIA V100, with two PCIe Gen 4 x16 HBA/NIC slots for up to 128GB/s of sustained data throughput. Alternatively, the EB4400 can be configured to provide 8 single-width PCIe Gen 4 x8 slots for FPGA data ingest or the latest storage add-in cards. Additional features including dynamic fan speed control, IPMI based system monitoring, replaceable fan filters, and optional SmartNIC host configuration elevate the EB4400 to the ideal expansion platform for the entire AI workflow. The EB4400 combines a small, rugged form factor with the latest PCIe Gen 4 add-in cards and features to meet the needs of any airborne, naval, or ground based transportable AI application.

The 4U Pro provides optimized PCIe Gen 4 configurable expansion for edge HPC/AI applications at twice the performance of the previous generation PCIe Gen 3. The appliance supports up to 8 NVIDIA A100 PCIe GPUs which deliver 2.5x FP64 performance compared to the NVIDIA V100, with four PCIe Gen 4 x16 HBA/NIC slots for up to 256GB/s of sustained data throughput. Alternatively, the 4U Pro can be configured to provide 16 single-width PCIe Gen 4 x8 slots for FPGA data ingest or the latest storage addin cards.

The Gen 4 PCIe SDS-3U-4a contains options for Dual or Single-Socket AMD EPYC™ 7003 Series Processors and provides a feature packed server platform for GPU or FPGA accelerated computing, U.2/U.3 NVMe storage and the most native PCIe Gen 4 x16 slots available on the market. This functionality is delivered in a size efficient edge ready rugged chassis with 3U height and 20” depth. This allows the highly integrated server to stand alone as a hyperconverged PCIe Gen 4 server or form the core CPU and memory resources for a scale-out, rack level, expandable or composable solution in the shallowest available racks. The SDS-3U-4a features up to 7 PCIe Gen 4 x16 slots, and 16 SATA/SAS/NVMe drives. The server supports up to 2TB of memory per CPU socket and a resource expanded BIOS for scale-out device enumeration and large memory mapped I/O used for GPUs and accelerators.

The EOS-2U-4a contains an AMD EPYC™ 7002 Series Processor and the EOS-2U-4i contains dual Intel Xeon Ice Lake Processors. These servers can stand alone or form the core CPU and memory resources for a scale-out, rack level, expandable solution. The EOS servers feature multiple PCIe slots with options for several add-in cards or U.2 or U.3 drives. The servers support 2-4TB of memory and a resource expanded BIOS for scale-out device enumeration and large memory mapped I/O used for GP-GPUs and accelerators.