A Faster, More Scalable Interconnect
NVLink allows for significantly accelerated speeds and performances in your applications. With the ability to create direct connections between multiple GPUs, as well as offering flexibility in connections between CPUs and GPUs, NVLink accelerates data exchange rates ranging from 5 to 12 times faster than before with PCIe.
The higher bandwidth between GPUs offers much better multi-GPU scalability. This increased flexibility, accessibility, and efficiency send your computational power into hyperdrive, driving the future in high performance computing.
Increase System Throughput
NVIDIA NVLink can bring up to 31% more performance to an otherwise identically configured server. Its dramatically higher bandwidth and reduced latency will enable even larger deep learning workloads to scale in performance as they continue to grow.
NVLink on Tesla V100 has increased the signaling rate from 20 to 25 GB/second in each direction. It can be used for GPU-to-CPU or GPU-to-GPU communication, as in the DGX-1V server with Tesla V100.
NVLink Capbable GPUs
|Product||Tesla P100: SXM2||Quadro GP100: PCIe||Quadro V100: SXM2||Quadro GP100: PCIe|
|2x Precision||5.3 TFLOPS||5.2 TFLOPS||7.5 TFLOPS||7.4 TFLOPS|
|1x Precision||10.6 TFLOPS||10.3 TFLOPS||15 TFLOPS||14.8 TFLOPS|
|Memory||16 GB||16 GB||16 GB||32 GB|
|Memory Bandwidth||720 GB/s||720 GB/s||900 GB/s||870 GB/s|