# NVIDIA A100 Tensor Core GPU

Unprecedented acceleration at every scale

## Accelerating the Most Important Work of Our Time

NVIDIA A100 Tensor Core GPU delivers unprecedented acceleration at every scale to power the world’s highest-performing elastic data centers for AI, data analytics, and HPC. Powered by the NVIDIA Ampere Architecture, A100 is the engine of the NVIDIA data center platform. A100 provides up to 20X higher performance over the prior generation and can be partitioned into seven GPU instances to dynamically adjust to shifting demands. The A100 80GB debuts the world’s fastest memory bandwidth at over 2 terabytes per second (TB/s) to run the largest models and datasets.

[Read NVIDIA A100 80GB PCIe Product Brief (PDF 380 KB)](https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/a100/pdf/PB-10577-001_v02.pdf)

[Read NVIDIA A100 40GB PCIe Product Brief (PDF 332 KB)](https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/a100/pdf/A100-PCIE-Prduct-Brief.pdf)

## Enterprise-Ready Software for AI

The NVIDIA EGX™ platform includes optimized software that delivers accelerated computing across the infrastructure. With NVIDIA AI Enterprise, businesses can access an end-to-end, cloud-native suite of  AI and data analytics software that’s  optimized, certified, and supported by NVIDIA to run on VMware vSphere  with  NVIDIA-Certified  Systems. NVIDIA AI Enterprise includes key enabling technologies  from NVIDIA for  rapid deployment, management, and scaling of AI workloads  in the modern hybrid cloud.

[Learn More](https://www.nvidia.com/en-us/data-center/products/ai-enterprise-suite.md)

## The Most Powerful End-to-End AI and HPC Data Center Platform

A100 is part of the complete NVIDIA data center solution that incorporates building blocks across hardware, networking, software, libraries, and optimized AI models and applications from [NGC™](https://www.nvidia.com/en-us/gpu-cloud.md). Representing the most powerful end-to-end AI and HPC platform for data centers, it allows researchers to rapidly deliver real-world results and deploy solutions into production at scale.

### Making of Ampere Video

[**WATCH VIDEO**](#)

### Deep Learning Training

#### Up to 3X Higher AI Training on Largest Models

DLRM Training

DLRM on HugeCTR framework, precision = FP16 | ​NVIDIA A100 80GB batch size = 48 | NVIDIA A100 40GB batch size = 32 | NVIDIA V100 32GB batch size = 32.

​AI models are exploding in complexity as they take on next-level challenges such as conversational AI. Training them requires massive compute power and scalability.

NVIDIA A100 [Tensor Cores](https://www.nvidia.com/en-us/data-center/tensorcore.md) with Tensor Float (TF32) provide up to 20X higher performance over the NVIDIA Volta with zero code changes and an additional 2X boost with automatic mixed precision and FP16. When combined with NVIDIA® [NVLink®](https://www.nvidia.com/en-us/data-center/nvlink.md), NVIDIA [NVSwitch™](https://www.nvidia.com/en-us/data-center/nvlink.md), PCI Gen4, NVIDIA® InfiniBand®, and the [NVIDIA Magnum IO™](https://www.nvidia.com/en-us/data-center/magnum-io.md) SDK, it’s possible to scale to thousands of A100 GPUs.

A training workload like BERT can be solved at scale in under a minute by 2,048 A100 GPUs, a world record for time to solution.

For the largest models with massive data tables like deep learning recommendation models (DLRM), A100 80GB reaches up to 1.3 TB of unified memory per node and delivers up to a 3X throughput increase over A100 40GB.

NVIDIA’s leadership in [MLPerf](https://www.nvidia.com/en-us/data-center/resources/mlperf-benchmarks.md), setting multiple performance records in the industry-wide benchmark for AI training.

[**Learn More About A100 for Training**](https://devblogs.nvidia.com/nvidia-ampere-architecture-in-depth/)

### Deep Learning Inference

A100 introduces groundbreaking features to optimize inference workloads. It accelerates a full range of precision, from FP32 to INT4. Multi-Instance GPU ([MIG](https://www.nvidia.com/en-us/technologies/multi-instance-gpu.md)) technology lets multiple networks operate simultaneously on a single A100 for optimal utilization of compute resources. And structural sparsity support delivers up to 2X more performance on top of A100’s other inference performance gains.

On state-of-the-art conversational AI models like BERT, A100 accelerates inference throughput up to 249X over CPUs.

On the most complex models that are batch-size constrained like RNN-T for automatic speech recognition, A100 80GB’s increased memory capacity doubles the size of each MIG and delivers up to 1.25X higher throughput over A100 40GB.

NVIDIA’s market-leading performance was demonstrated in [MLPerf Inference](https://www.nvidia.com/en-us/data-center/resources/mlperf-benchmarks.md). A100 brings 20X more performance to further extend that leadership.

[**Learn More About A100 for Inference**](https://devblogs.nvidia.com/nvidia-ampere-architecture-in-depth/)

#### Up to 249X Higher AI Inference Performance Over CPUs

BERT-LARGE Inference

BERT-Large Inference | CPU only: Xeon Gold 6240 @ 2.60 GHz, precision = FP32, batch size = 128 | V100: NVIDIA TensorRT™ (TRT) 7.2, precision = INT8, batch size = 256 | A100 40GB and 80GB, batch size = 256, precision = INT8 with sparsity.​

#### Up to 1.25X Higher AI Inference Performance Over A100 40GB

RNN-T Inference: Single Stream

MLPerf 0.7 RNN-T measured with (1/7) MIG slices. Framework: TensorRT 7.2, dataset = LibriSpeech, precision = FP16.

### High-Performance Computing

To unlock next-generation discoveries, scientists look to simulations to better understand the world around us.

NVIDIA A100 introduces double precision Tensor Cores  to deliver the biggest leap in HPC performance since the introduction of GPUs. Combined with 80GB of the fastest GPU memory, researchers can reduce a 10-hour, double-precision simulation to under four hours on A100. HPC applications can also leverage TF32 to achieve up to 11X higher throughput for single-precision, dense matrix-multiply operations.

For the HPC applications with the largest datasets, A100 80GB’s additional memory delivers up to a 2X throughput increase with Quantum Espresso, a materials simulation. This massive memory and unprecedented memory bandwidth makes the A100 80GB the ideal platform for next-generation workloads.

[**Review Latest GPU Performance on HPC Applications**](https://developer.nvidia.com/hpc-application-performance)

#### 11X More HPC Performance in Four Years

Top HPC Apps​

Geometric mean of application speedups vs. P100: Benchmark application: Amber [PME-Cellulose\_NVE], Chroma [szscl21\_24\_128], GROMACS  [ADH Dodec], MILC [Apex Medium], NAMD [stmv\_nve\_cuda], PyTorch (BERT-Large Fine Tuner], Quantum Espresso [AUSURF112-jR]; Random Forest FP32 [make\_blobs (160000 x 64 : 10)], TensorFlow [ResNet-50], VASP 6 [Si Huge] | GPU node with dual-socket CPUs with 4x NVIDIA P100, V100, or A100 GPUs.

#### Up to 1.8X Higher Performance for HPC Applications

Quantum Espresso​

Quantum Espresso measured using CNT10POR8 dataset, precision = FP64.

### High-Performance Data Analytics

#### 2X Faster than A100 40GB on Big Data Analytics Benchmark

Big data analytics benchmark |  30 analytical retail queries, ETL, ML, NLP on 10TB dataset | V100 32GB, RAPIDS/Dask | A100 40GB and A100 80GB, RAPIDS/Dask/BlazingSQL​

Data scientists need to be able to analyze, visualize, and turn massive datasets into insights. But scale-out solutions are often bogged down by datasets scattered across multiple servers.

Accelerated servers with A100 provide the needed compute power—along with massive memory, over 2 TB/sec of memory bandwidth, and scalability with NVIDIA® [NVLink® and NVSwitch™](https://www.nvidia.com/en-us/data-center/nvlink.md), —to tackle these workloads. Combined with InfiniBand, [NVIDIA Magnum IO™](https://www.nvidia.com/en-us/data-center/magnum-io.md) and the [RAPIDS™](https://www.nvidia.com/en-us/deep-learning-ai/software/rapids.md) suite of open-source libraries, including the RAPIDS Accelerator for Apache Spark for GPU-accelerated data analytics, the NVIDIA data center platform accelerates these huge workloads at unprecedented levels of performance and efficiency.

On a big data analytics benchmark, A100 80GB delivered insights with a 2X increase over A100 40GB, making it ideally suited for emerging workloads with exploding dataset sizes.

[**Learn More About Data Analytics**](https://www.nvidia.com/en-us/deep-learning-ai/solutions/data-science.md)

### Enterprise-Ready Utilization

#### 7X Higher Inference Throughput with Multi-Instance GPU (MIG)

BERT Large Inference

BERT Large Inference | NVIDIA TensorRT™ (TRT) 7.1 | NVIDIA T4 Tensor Core GPU: TRT 7.1, precision = INT8, batch size = 256 | V100: TRT 7.1, precision = FP16, batch size = 256 | A100 with 1 or 7 MIG instances of 1g.5gb: batch size = 94, precision = INT8 with sparsity.​

A100 with [MIG](https://www.nvidia.com/en-us/technologies/multi-instance-gpu.md) maximizes the utilization of GPU-accelerated infrastructure. With MIG, an A100 GPU can be partitioned into as many as seven independent instances, giving multiple users access to GPU acceleration. With A100 40GB, each MIG instance can be allocated up to 5GB, and with A100 80GB’s increased memory capacity, that size is doubled to 10GB.

MIG works with Kubernetes, containers, and [hypervisor-based server virtualization](https://www.nvidia.com/en-us/data-center/products/ai-enterprise-suite.md). MIG lets infrastructure managers offer a right-sized GPU with guaranteed quality of service (QoS) for every job, extending the reach of accelerated computing resources to every user.

[**Learn More About Mig**](https://www.nvidia.com/en-us/technologies/multi-instance-gpu.md)

### Get the Most From Your Systems

An NVIDIA-Certified System, comprising of A100 and NVIDIA Mellanox SmartnNICs and DPUs is validated for performance, functionality, scalability, and security allowing enterprises to easily deploy complete solutions for AI workloads from the NVIDIA NGC catalog.

[Learn More](https://www.nvidia.com/en-us/data-center/products/certified-systems.md)

## Data Center GPUs

### NVIDIA A100 for HGX

Ultimate performance for all workloads.

### NVIDIA A100 for PCIe

Highest versatility for all workloads.

## Specifications

|  | A100 80GB PCIe | A100 80GB SXM |
| --- | --- | --- |
| FP64 | 9.7 TFLOPS | | | |
| FP64 Tensor Core | 19.5 TFLOPS | | | |
| FP32 | 19.5 TFLOPS | | | |
| Tensor Float 32 (TF32) | 156 TFLOPS | 312 TFLOPS\* | | | |
| BFLOAT16 Tensor Core | 312 TFLOPS | 624 TFLOPS\* | | | |
| FP16 Tensor Core | 312 TFLOPS | 624 TFLOPS\* | | | |
| INT8 Tensor Core | 624 TOPS | 1248 TOPS\* | | | |
| GPU Memory | 80GB HBM2e | 80GB HBM2e |
| GPU Memory Bandwidth | 1,935 GB/s | 2,039 GB/s |
| Max Thermal Design Power (TDP) | 300W | 400W \*\*\* |
| Multi-Instance GPU | Up to 7 MIGs @ 10GB | Up to 7 MIGs @ 10GB |
| Form Factor | PCIe Dual-slot air-cooled or single-slot liquid-cooled | SXM |
| Interconnect | NVIDIA® NVLink® Bridge for 2 GPUs: 600 GB/s \*\* PCIe Gen4: 64 GB/s | NVLink: 600 GB/s  PCIe Gen4: 64 GB/s |
| Server Options | Partner and NVIDIA-Certified Systems™ with 1-8 GPUs | NVIDIA HGX™ A100-Partner and NVIDIA-Certified Systems with 4,8, or 16 GPUs NVIDIA DGX™ A100 with 8 GPUs |

\* With sparsity  
 \*\* SXM4 GPUs via HGX A100 server boards; PCIe GPUs via NVLink Bridge for up to two GPUs  
 \*\*\* 400W TDP for standard configuration. HGX A100-80GB custom thermal solution (CTS) SKU can support TDPs up to 500W

See the Latest MLPerf Benchmark Data

[View Results](https://www.nvidia.com/en-us/data-center/resources/mlperf-benchmarks.md)

### Inside the NVIDIA Ampere Architecture

Learn what’s new with the NVIDIA Ampere architecture and its implementation in the NVIDIA A100 GPU.

[Read Whitepaper](https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/nvidia-ampere-architecture-whitepaper.pdf)