Third-Generation Tensor Cores

Performance and Versatility for HPC and AI

First introduced in the NVIDIA Volta^™ architecture, NVIDIA Tensor Core technology has brought dramatic speedups to AI training and inference operations, bringing down training times from weeks to hours and providing massive acceleration to inference.

The NVIDIA Ampere architecture builds upon these innovations by providing up to 20X higher FLOPS for AI. It does so by improving the performance of existing precisions and bringing new precisions—TF32, INT8, and FP64—that accelerate and simplify AI adoption and extend the power of NVIDIA Tensor Cores to HPC.

Multi-Instance GPU (MIG)

Securely, Isolated Multi-Tenancy

Each 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.

Third-Generation NVIDIA NVLink

Scale to Multiple GPUs With 80GB of Memory

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.

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 40 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.

Data Science and Data Analytics

Accelerate end-to-end data science and analytics workflows with powerful performance to extract meaningful insights from large-scale datasets quickly. By combining the high-performance computing capabilities of the A800 40GB Active with NVIDIA AI Enterprise, data practitioners can leverage a large collection of libraries, tools, and technologies to accelerate data science workflows- from data prep and training to inference.

Training and Development

With 40GB of HBM2 memory and powerful third-generation Tensor Cores that deliver up to 2X the performance of the previous generation, the A800 40GB Active GPU delivers incredible performance to conquer demanding AI development and training workflows on workstation platforms, including data preparation and processing, model optimization and tuning, and early-stage training.

The NVIDIA AI Enterprise software platform accelerates and simplifies deploying AI at scale, allowing organizations to develop once and deploy anywhere. Coupling this powerful software platform with the A800 40GB Active GPU enables AI developers to build, iterate, and refine AI models on workstations using included frameworks, simplifying the scaling process and reserving costly data center computing resources for more expensive, large-scale computations.

Inference

Inference is where AI delivers results, providing actionable insights by operationalizing trained models. With 432 third-generation Tensor Cores and 6,912 CUDA^® Cores, A800 40GB Active delivers 2X the inference operation performance versus the previous generation with support for structural sparsity and a broad range of precisions, including TF32, INT8, and FP64. AI developers can use NVIDIA inference software including NVIDIA TensorRT™, NVIDIA Triton™ Inference Server, and NVIDIA Triton™ Management Service that are part of NVIDIA AI Enterprise to simplify and optimize the deployment of AI models at scale.

Generative AI

Using neural networks to identify patterns and structures within existing data, generative AI applications enable users to generate new and original content from a wide variety of inputs and outputs, including images, sounds, animation, and 3D models. Leverage the NVIDIA generative AI solution, NeMo™ Framework, included in NVIDIA AI Enterprise along with NVIDIA A800 40GB Active GPU for easy, fast, and customizable generative AI model development.

Manufacturing, Product Development, and Engineering

The A800 40GB Active GPU delivers incredible performance for GPU-accelerated computer-aided engineering (CAE) applications. Engineering and product development professionals can run large-scale simulations for finite element analysis (FEA), computational fluid dynamics (CFD), construction engineering management (CEM), and other engineering analysis codes in full FP64 precision with incredible speed, shortening development timelines and accelerating time to value. With the addition of RTX-accelerated GPUs providing display capabilities, scientists and engineers can visualize large-scale simulations and models in full design fidelity.

Energy and Geosciences

With 9.7 TFLOPS of FP64 compute performance, the A800 40GB Active GPU enables geoscience professionals to power the latest AI-augmented exploration and production software workflows and accelerate simulation processes to gain faster insight into subsurface data. For large-scale datasets, two A800 40GB Active GPUs can be connected with NVLink to provide 80GB of memory and twice the processing power.

Life Sciences

With the A800 40GB Active professionals across life science disciplines can accelerate complex data processing tasks, enable faster discovery, and improve decision-making. AI-accelerated life science applications like genomics sequencing, medical imaging, and personalized medicine can benefit from faster training and inference performance to accelerate the analysis of large datasets. For complex simulations and data processing tasks requiring high accuracy, FP64 capabilities allow for scientific applications like molecular dynamics, drug discovery, and genomic analysis to run with higher accuracy and precision, yielding more reliable results.