NVIDIA A100 Tensor Core GPU PCIe 80GB HBM2e
NVIDIA A100 80GB PCIe is a legacy high-performance data center Tensor Core GPU with 80GB HBM2e ECC memory. Optimized for LLM training & inference, HPC and data analytics. It supports MIG to split one GPU into up to seven isolated 10GB instances for elastic multi-user scheduling.
$22000.00
Quantity
1 items selected,Total $22000.00
+8618819251189
Delivery to
Please add the address first.
Specifications
Detail
The NVIDIA A100 80GB PCIe Tensor Core GPU is a widely deployed data center accelerator built for AI training, large-model inference, high-performance computing and big data analytics. It carries 80GB high-capacity HBM2e ECC memory with up to 1935 GB/s memory bandwidth, enabling it to fit huge models and massive datasets directly on GPU memory.
It supports multi-precision Tensor Core computing including TF32, BFLOAT16, FP16 and INT8, delivering excellent throughput for foundation model workloads. The MIG (Multi-Instance GPU) feature partitions a single A100 into up to seven independent GPU instances, allowing flexible resource allocation for multiple teams or small jobs on shared servers.
This is PCIe form factor (PCIe 4.0 x16), passive cooling, dual-slot FHFL design, easier to integrate into standard off-the-shelf rack servers compared with SXM variants. It is broadly adopted by research labs, cloud service providers and system integrators for stable AI cluster deployment.
Note: A100 PCIe version does not have native NVLink by default; NVLink bridge is optional for dual-card setups.
Key Specifications
GPU: NVIDIA A100 Tensor Core GPU
Memory: 80GB HBM2e, ECC enabled
Memory Bandwidth: 1935 GB/s
Interface: PCIe 4.0 x16
TDP: 300W
Cooling: Passive
Form Factor: Dual-slot, FHFL
Feature: MIG support (up to 7 × 10GB instances)
Supported Workloads: LLM training & inference, HPC, simulation, data analytics
WhatsApp：+86-18819251189