More Power to Small/Medium-sized GPU Providers

1
More Power to Small/Medium-Sized GPU Providers

• Run Pod

2
Customization

• RunPod currently supports background host jobs
  that you can run while your machine isn't being
  rented. You can mine, do your own machine
  learning, whatever you want to do. In addition to
  this, your background job can be loaded from a
  local docker image and it supports power limiting
  and overclocking for just your background jobs,
  so you don't affect your clients.

3
Metrics

• RunPod's supports some basic metrics out of the
  box. These are stored in our time series
  database, so it will be easy for us to generate
  more robust charts in the future like historical
  hardware performance and how much money your
  machines are earning over time. For now, it is
  easy to see how your machine hardware is
  operating at a glance.

4
Mobile Friendly

• All of RunPod's dashboard pages are mobile
  friendly. You can check in on your machines from
  your phone!

5
Easy

• The RunPod host daemon is a one-line command that
  will take care of all everything you need to get
  started. All you need to do is install Ubuntu LTS
  and set up your SSH access. That's it. You can
  learn more in our short installation guide.

6
Goals

• We have our sights set high. The RunPod platform
  is already capable of serving the needs of a
  large consumer base, but we are just getting
  started. We eventually aim to build SaaS services
  on top of the framework that we created. This
  means that hosts will get exposure to an entirely
  new pool of clients that are willing to pay more
  for convenience and features that other platforms
  don't have.

7
Agile

• One of the biggest strengths of the RunPod
  platform is that the team is devoted to making it
  successful. The entire platform as it stands was
  built in less than 3 months of development time
  and is a testament to how fast it can evolve.
  Join the hustle!

8
Client Friendly

• Last, but probably the most important, RunPod is
  client centric. After all, without clients, what
  is the use in having a bunch of dashboards that
  show your machines are idling? Here's a clip of
  the client rental experience for renting 12 GPUs
  on a basic Jupyter notebook with tensorflow
  pre-installed.

9

• https//runpod.io/