Product	NVIDIA® Tesla™ V100 GPU Computing Accelerator - 16GB HBM2 - SXM2 NVLink	NVIDIA® Tesla™ V100 GPU Computing Accelerator - 32GB HBM2 - SXM2 NVLink
Action	Select	Select
Main Specifications
Product Series	Tesla P100	Tesla V100
Core Type	NVIDIA CUDA	NVIDIA CUDA
Host Interface	PCI Express 3.0 x16	PCI Express 3.0 x16
GPU Architecture	Volta	Volta
Detailed Specifications
Streaming Processor Cores	5120 CUDA Cores	5120 CUDA Cores
NVIDIA Tensor Cores	640 Tensor Cores	640 Tensor Cores
CoWoS HBM2 Stacked Memory Capacity	16 GB	32 GB
CoWoS HBM2 Stacked Memory Bandwidth	900 GB/s	900 GB/s
Max Memory Bandwidth	900 GB/s	900 GB/s
Peak Half-Precision Performance	21.2 TeraFLOPS	21.2 TeraFLOPS
Peak Single Precision floating point performance (GFLOP)	15.7 TeraFLOPS	15.7 TeraFLOPS
Peak Double Precision floating point performance (GFLOP)	7.8 TeraFLOPS	7.8 TeraFLOPS
Total NVLink Bandwidth	300 GB/s	300 GB/s
Tensor Performance	125 TeraFLOPS	125 TeraFLOPS
Cooling	Passive	Passive
Max Graphics Card Power (W)	300W	300W
Action	Select	Select