NVIDIA Quadro
GV100
Built on NVIDIA’s compute optimized Volta architecture, the GV100 features 7.4 TFLOPS of FP64 performance, 32 GB of HBM2 GPU memory, which offers ECC without a performance or capacity penalty, and an ultra-fast 870 GB/s of memory bandwidth, giving you the graphics and compute power required for the most demanding CAE workflows.
CUDA CORES: 5120
TENSOR CORES: 640
PEAK FP64: 7.4 TFLOPS
PEAK FP32: 14.8 TFLOPS
DEEP LEARNING TFLOPS: 118.5
GPU MEMORY: 32 GB HBM2
MEMORY BANDWIDTH: 870 GB/s
NVLINK: 200 GB/s bi-directional