NVIDIA A30X
Converged Accelerators | Networking and Compute, Unified
In one unique, efficient architecture, NVIDIA converged accelerators combine the powerful performance of NVIDIA GPUs with the enhanced network and security of NVIDIA smart network interface cards (SmartNICs) and data processing units (DPUs). Deliver maximum performance and enhanced security for I/O intensive GPU accelerated workloads, from the data center to the edge.
The A30X combines the NVIDIA A30 Tensor Core GPU with the BlueField-2 DPU. The design of this card provides a good balance of compute and input/output (IO) performance use cases such as 5G vRAN and AI-based cybersecurity. Multiple services can run on the A30X GPU, with the low latency and predictable performance provided by the onboard PCIe switch.
Performance Highlights
Peak FP64
|
5.2 TFLOPS |
Peak FP64 Tensor Core
|
10.3 TFLOPS | Sparsity |
Peak FP32
|
10.3 TFLOPS |
TF32 Tensor Core
|
82.6 TFLOPS | Sparsity |
Peak FP16 Tensor Core
|
165 TFLOPS | Sparsity |
Peak INT8 Tensor Core
|
330 TOPS | Sparsity |
Multi-Instance GPU Support
|
Yes | Up to 4 |
GPU Memory
|
24 GB HBM2e |
Memory Bandwidth
|
1223 GB/s |
Media Engines
|
1x Optical Flow Accelerator (OFA)
1x JPEG Decoder (NVJPEG)
4x Video Decoders (NVDEC) |
Interconnect
|
PCIe Gen4 (x16 Physical, x8 Electrical | NVLink Bridge) |
Networking
|
2x 100 Gbps ports, Ethernet or InfiniBand |
Form Factor
|
Dual-Slot, Full-Height, Full-Length |
Thermal Solution
|
Passive |
Maximum Power Consumption
|
230 W |
High-Performance 5G
- NVIDIA converged accelerators like the A30X provide a high-performance platform for running 5G workloads. Because data doesn't need to go through the host PCIe system, processing latency is greatly reduced. The resulting higher throughput also allows for a greater subscriber density per server. NVIDIA A30X teams with the NVIDIA Aerial SDK, an application framework for building high-performance, software-defined, cloud-native 5G networks to address increasing user demand. It enables GPU-accelerated signal and data processing for 5G virtual radio access networks (vRANs). Aerial is 100% software defined and delivers a highly-programmable PHY layer and has the capability to support L2+ functions seamlessly. The A30X's GPU-accelerated processing lets complex computations run faster than existing L1 processing solutions, giving improved performance results. Together, NVIDIA A30X and the Aerial SDK offer commercial-off-the-shelf (COTS) hardware support, making it easier to deploy cloud-native platforms such as NVIDIA EGX. It's Kubernetes based and provides container orchestration for ease of deployment and management. Available as a .zip package or NVIDIA NGC container image the A30X and Aerial duo is a compelling platform on which to build and deploy GPU-accelerated 5G vRANs. Join the Aerial early access program today.
AI-Based Cybersecurity
- Converged accelerators like the A30X open up a new range of possibilities for AI-based cybersecurity and networking. The BlueField-2 DPU's Arm cores can be programmed using the NVIDIA Morpheus application framework to implement GPU accelerated advanced network functions such as threat detection, data leak prevention, and anomalous behavior profiling, GPU processing can be applied directly to network traffic at a high data rate, and data travels on a direct path between the GPU and GPU, providing better isolation. The NVIDIA A30X is a performant engine for NVIDIA Morpheus, an open applications framework that enables cybersecurity developers to create optimized AI pipelines for filtering, processing, and classifying large volumes of real-time data. Bringing a new level of information security to the data center, cloud, and edge, Morpheus uses AI to identify, capture, and act on threats and anomalies that were previously impossible to identify. NVIDIA Morpheus can be downloaded here.
Warranty
Free dedicated phone and email technical support
(1-800-230-0130)
Dedicated NVIDIA professional products Field Application Engineers
Resources
Contact gopny@pny.com for additional information.