Unprecedented Acceleration at Every Scale
NVIDIA Ampere Architecture In-Depth
Groundbreaking Innovations
NVIDIA AMPERE ARCHITURE
Whether using MIG to partition an A100 GPU into smaller instances, or NVLink to connect multiple GPUs to speed large-scale workloads, A100 can readily handle different-sized acceleration needs, from the smallest job to the biggest multi-node workload. A100’s versatility means IT managers can maximize the utility of every GPU in their data center, around the clock.
THIRD-GENERATION TENSOR CORES
NVIDIA A100 delivers 312 teraFLOPS (TFLOPS) of deep learning performance. That’s 20X the Tensor floating-point operations per second (TOPS) for deep learning inference compared to NVIDIA Volta GPUs.
NEXT GENERATION NVLINK
NVIDIA NVLink in A100 delivers 2X higher throughput compared to the previous generation. When combined with NVIDIA NVSwitch™, up to 16 A100 GPUs can be interconnected at up to 600 gigabytes per second (GB/sec), unlashing the highest application performance possible on a single server. NVLink is available A100 SXM GPUs via HGX A100 server boards and in PCIe GPUs via an NVLink Bridge for up to 2 GPUs.
MULTI-INSTANCE GPU (MIG)
An A100 GPU can be partitioned into as many as seven GPU instances, fully isolated at the hardware level with their own high-bandwidth memory, cache, and compute cores. MIG gives developers access to breakthrough acceleration for all their applications, and IT administrators can offer right-sized GPU acceleration for every job, optimizing utilization and expanding access to very user and application.
HIGH-BANDWIDTH MEMORY (HBM2E)
With up to 80 gigabytes of HBM2e, A100 delivers the world’s fastest, GPU memory bandwidth of over 2TB/s, as well a dynamic random-access memory (DRAM) utilization efficiency of 95%. A100 delivers 1.7X higher memory bandwidth over the previous generation.
STRUCTURAL SPARSITY
AI networks have millions to billions of parameters. Not all of these parameters are needed for accurate predictions, and some can be converted to zeros, making the models “sparse” without compromising accuracy. Tensor Cores in A100 can provide up to 2X higher performance for sparse models. While the sparsity feature more readily benefits AI inference, it can also improve the performance of model training.