GPU Leaderboard
Compare performance metrics for AI workloads across different NVIDIA GPUs
NVIDIA Tesla V100 PCIe 16 GB
NVIDIA Tesla V100 SXM2 16 GBArchitecture
NVIDIA VoltaArchitecture
NVIDIA VoltaRelease Date
2017.06
Release Date
2019.11
Memory
16 GB HBM2
Memory
16 GB HBM2
FP8 Performance
Not supported
FP8 Performance
Not supported
FP16 Performance
28.26 TFLOPS
FP16 Performance
9.8% faster
31.33 TFLOPS
FP32 Performance
14.13 TFLOPS
FP32 Performance
9.8% faster
15.67 TFLOPS
FP64 Performance
7.07 TFLOPS
FP64 Performance
9.7% faster
7.83 TFLOPS
CUDA Cores
5120
CUDA Cores
5120
Power Consumption
16.7% less
250 W
Power Consumption
300 W