PNY A800 NVIDIA RTX A800 40 GB High Bandwidth Memory 2 (HBM2)

• NVIDIA RTX A800 40 GB High Bandwidth Memory 2 (HBM2)
• 5120 bit PCI Express x16 4.0
• 1x 16-pin

High-Performance Data Science and AI Platform
Rapid growth in workload complexity, data size, and the proliferation of emerging workloads like generative AI are ushering in a new era of computing, accelerating scientific discovery, improving productivity, and revolutionizing content creation. As models continue to explode in size and complexity to take on next-level challenges, an increasing number of workloads will need to run on local devices. Next-generation workstation platforms will need to deliver high-performance computing capabilities to support these complex workloads.

The NVIDIA A800 40GB Active GPU accelerates data science, AI, and HPC workflows with 432 third-generation Tensor Cores to maximize AI performance and ultra-fast and efficient inference capabilities. With third-generation NVIDIA NVLink technology, A800 40GB Active offers scalable performance for heavy AI workloads, doubling the effective memory footprint and enabling GPU-to-GPU data transfers up to 400 gigabytes per second (GB/s) of bidirectional bandwidth. This board is an AI-ready development platform with NVIDIA AI Enterprise, and delivers workstations ideally suited to the needs of skilled AI developers and data scientists.

PERFORMANCE AND USEABILITY FEATURES
NVIDIA Ampere Architecture
NVIDIA A800 40GB Active is one of the world’s most powerful data center GPUs for AI, data analytics, and high-performance computing (HPC) applications. Building upon the major SM enhancements from the Turing GPU, the NVIDIA Ampere architecture enhances tensor matrix operations and concurrent executions of FP32 and INT32 operations.

More Efficient CUDA Cores
The NVIDIA Ampere architecture’s CUDA® cores bring up to 2.5x the single-precision floating point (FP32) throughput compared to the previous generation, providing significant performance improvements for any class or algorithm, or application that can benefit from embarrassingly parallel acceleration techniques.

Third-Generation Tensor Cores
Purpose-built for deep learning matrix arithmetic at the heart of neural network training and inferencing functions, the NVIDIA A800 40GB Active includes enhanced Tensor Cores that accelerate more datatypes (TF32 and BF16) and includes a new Fine-Grained Structured Sparsity feature that delivers up to 2x throughput for tensor matrix operations compared to the previous generation.

PCIe Gen 4
The NVIDIA A800 40GB Active supports PCI Express Gen 4, which provides double the bandwidth of PCIe Gen 3, improving data-transfer speeds from CPU memory for data-intensive tasks like AI and data science.

Multi-Instance GPU (MIG): Securely, Isolated Multi-Tenancy
Every AI and HPC application can benefit from acceleration, but not every application needs the performance of a full A800 40GB Active GPU. Multi-Instance GPU (MIG) maximizes the utilization of GPU-accelerated infrastructure by allowing an A800 40GB Active GPU to be partitioned into as many as seven independent instances, fully isolated at the hardware level. This provides multiple users access to GPU acceleration with their own high-bandwidth memory, cache, and compute cores. Now, developers can access breakthrough acceleration for all their applications, big and small, and get guaranteed quality of service. And IT administrators can offer right-sized GPU acceleration for optimal utilization and expand access to every user and application.

Ultra-Fast HBM2 Memory
To feed its massive computational throughput, the NVIDIA A800 40GB Active GPU has 40GB of high-speed HBM2 memory with a class-leading 1,555GB/s of memory bandwidth—a 79 percent increase compared to NVIDIA Quadro GV100. In addition to 40GB of HBM2 memory, A800 40GB Active has significantly more on-chip memory, including a 48 megabyte (MB) level 2 cache, which is nearly 7x larger than the previous generation. This provides the right combination of extreme bandwidth on-chip cache and large on-package high-bandwidth memory to accelerate the most compute-intensive AI models.

Compute Preemption
Preemption at the instruction-level provides finer grain control over compute and tasks to prevent longer-running applications from either monopolizing system resources or timing out.

MULTI-GPU TECHNOLOGY SUPPORT
Third-Generation NVLink
Connect a pair of NVIDIA A800 40GB Active cards with NVLink to increase the effective memory footprint and scale application performance by enabling GPU-to-GPU data transfers at rates up to 100GB/s (bidirectional) for a total bandwidth of 200GB/s. Scaling applications across multiple GPUs requires extremely fast movement of data. The third generation of NVLink in A800 40GB Active provides 400GB/s of GPU-to-GPU direct bandwidth.

SOFTWARE SUPPORT
Software Optimized for AI
Deep learning frameworks such as Caffe2, MXNet, CNTK, TensorFlow, and others deliver dramatically faster training times and higher multi-node training performance. GPU-accelerated libraries such as cuDNN, cuBLAS, and TensorRT deliver higher performance for both deep learning inference and High-Performance Computing (HPC) applications.

NVIDIA CUDA Parallel Computing Platform
Natively execute standard programming languages like C/C++ and Fortran, and APIs such as OpenCL, OpenACC, and Direct Compute to accelerate techniques such as ray tracing, video and image processing, and computation fluid dynamics.

Unified Memory
A single, seamless 49-bit virtual address space allows for the transparent migration of data between the full allocation of CPU and GPU memory.

NVIDIA AI Enterprise
Enterprise adoption of AI is now mainstream and leading to an increased demand for skilled AI developers and data scientists. Organizations require a flexible, high-performance platform consisting of optimized hardware and software to maximize productivity and accelerate AI development. NVIDIA A800 40GB Active and NVIDIA AI Enterprise provide an ideal foundation for these vital initiatives.