Curtiss-Wright’s Defense Solutions division introduced its first 3U OpenVPX GPGPU processor module designed in compliance with the U.S. Army CCDC C5ISR Center’s C4ISR/EW Modular Open Suite of Standards (CMOSS) and aligned with standards currently being defined by The Open Group Sensor Open Systems Architecture (SOSA) Consortium. Developed to support compute-intensive ISR and EW systems, this fully rugged SOSA-aligned variant of the VPX3-4935, an NVIDIA Quadro Turing-based GPGPU processor card module, is the latest addition to Curtiss-Wright’s growing roster of CMOSS-compliant / SOSA aligned hardware. The SWaP-optimized VPX3-4935 is ideal for accelerating tensor/matrix computation used for deep learning neural network inference used in deployed artificial intelligence (AI) and machine learning (ML) applications requiring TFLOPS of accelerated processing. These applications include high-performance radar, SIGINT, EO/IR, data fusion ingest, processing and display, and autonomous vehicles
The VPX3-4935 SOSA-aligned module features an NVIDIA Quadro Turing TU104 (RTX5000E) GPU that delivers 11.2 TFLOPS/TIPS performance. It provides 3072 CUDA cores, 384 tensor cores, and 48 ray-tracing (RT) cores. Designed to work in conjunction with NVIDIA TensorRT and CUDA cores, the module’s Turing tensor cores add INT8 and INT4 matrix operation while continuing support for high-precision workloads. These state-of-the-art GPGPU modules further extend Curtiss-Wright’s proven leadership as a supplier of the most advanced computing solutions for embedded ISR applications. To meet demanding rugged military and aerospace specifications, these GPGPU boards feature a chip-down design.
The VPX3-4935 modules are fully interoperable with Curtiss-Wright’s broad family of 3U OpenVPX solutions. For example, SWaP-constrained systems can pair the single VPX3-4935 modules with the Intel Xeon D processor-based CHAMP-XD1 DSP engine, which together can augment the powerful sensor processing capabilities of our Xilinx FPGA-based transceiver modules.
Curtiss-Wright HPEC modules and systems are supported by the OpenHPEC Accelerator Suite of best-in-class software development tools. These powerful tools enable system developers to develop their software faster. For example, the suite includes the powerful Bright Cluster Manager from Bright Computing, an NVIDIA partner. Bright Cluster Manager provisions and monitors both the CPU and GPU boards and includes a fully configurable module environment. The OpenHPEC tool suite includes Bright’s deep learning libraries and tools from both Intel and NVIDIA, including Caffe and TensorFlow. It also provides the Arm Forge suite, which enables true system-level debugging and profiling for both CPU and GPUs, and supports MPI, OpenMP, and ACC.
For high-speed, low-latency, peer-to-peer communications, the OpenHPEC Accelerator Suite also includes the Dolphin eXpressWare PCIe communication library, which hides the complexities of directly programming the system’s PCIe devices. In addition to supporting GPU sharing between the CPUs, the Dolphin library also supports both CPU direct and remote direct memory access (RDMA). The use of the OpenHPEC Accelerator Suite simplifies, speeds, and lowers the cost of ISR application development. These tools deliver the benefits of open standard High-Performance Computing (HPC) software to the COTS market to effectively remove risk when developing large-scale embedded computer clusters.
Resulting from its Reseller Agreement with WOLF Advanced Technology, the VPX3-4935 modules have been pre-validated. They complement Curtiss-Wright’s previously announced family of NVIDIA Turing GPGPU modules by speeding and easing the integration of HPEC solutions into SOSA-aligned deployed systems.