8x Tesla V100 (16 GB) by Lambda Labs


Bandwidth testing are conducted with local and block storages attached to the instance. We use the well-known open-source tool FIO. To express bandwidth the following parameters are used:

  • 1M block
  • Sequential access
  • No filesystem (except for write access with root volume)
  • Avoidance of cache and buffer
Specifications:
CPU
92
RAM
448.0 GB
GPU
8 Tesla V100-PCIE-16GB
Root
6041 GB NVMe

It seems that we don't have any bandwidth results.


Sign up as Freemium to get more insight.