API Support
CUDA, OpenCL, DirectCompute
Brand
HPE
Compatibility
HPE Servers
Computational Capability
Volta Architecture
Form Factor
SXM2 Module for Servers
GPU
NVIDIA Tesla V100
Interconnect Support
NVLink
Interface
PCI Express Gen3 x16
Manufacturer's Warranty
Limited Warranty
Memory
16GB HBM2
Memory Bandwidth
Up to 900 GB/s
Module Type
SXM2
NVIDIA CUDA Cores
5120
Operating Systems Supported
Linux and Windows Server
Other Features
Volta-optimized Tensor Cores, NVLink Interconnection, High-Bandwidth Memory
Performance
14 TFLOPS (single-precision), 7 TFLOPS (double-precision)
Product Name
NVIDIA Tesla V100 SXM2 16GB Module
Product Number
Q2N66A
TDP
300W
Thermal Solution
Passive Heatsink
Use Case
High-Performance Computing, Deep Learning, Machine Learning, AI Research