Putting GPUs to work with Kubernetes
medium.com
Putting GPUs to work with Kubernetes
1–10 of 54 posts
Re: Putting GPUs to work with Kubernetes
#2Great read, on a smaller scale, I have found nvidia-docker with nvidia-docker-compose to be a great solution for deploying docker containers on AWS P2 machines with 8 GPUs.
Re: Putting GPUs to work with Kubernetes
#3I keep seeing Kubernetes appear on HackerNews. Is there a quick thing I can read to explain why everyone's so excited about? I know it's container orchestration, but I'm not sure what people are using it for or what pain point it is revealing.
Re: Putting GPUs to work with Kubernetes
#4Is the author of this working on official support or just testing? I know there's a gpu roadmap for k8s, but I can't tell from this blog if this was part of it.
Re: Putting GPUs to work with Kubernetes
#5Unprofitable for ETH mining maybe, but it seems like a natural fit to rent time on it to deep learning people with slow training models. Although that could still be unprofitable after the cost of electricity, I guess it's a question of market size/demand. A lot of deep learning is already at big infrastructure players anyway who wouldn't need the service, leaving academics / smaller companies. But maybe some people would find a reliable, scalable GPU cluster valuable.
Re: Putting GPUs to work with Kubernetes
#6I keep seeing Kubernetes appear on HackerNews. Is there a quick thing I can read to explain why everyone's so excited about? I know it's container orchestration, but I'm not sure what people are using it for or what pain point it is revealing.
Re: Putting GPUs to work with Kubernetes
#7I keep seeing Kubernetes appear on HackerNews. Is there a quick thing I can read to explain why everyone's so excited about? I know it's container orchestration, but I'm not sure what people are using it for or what pain point it is revealing.
https://vishh.github.io/docs/concepts/overview/what-is-kuber...
If you'd like to engage with me further, how does a company know it needs Kubernetes? If I'm Soylent and I'm processing a few orders a minute, I'm probably safe with a few redundant monoliths. Do I have to be Uber? What's the middle-ground between Soylent and Uber that would still need this?
Is the answer the same as the question "who needs a microservice architecture"?
Re: Putting GPUs to work with Kubernetes
#8Is the author of this working on official support or just testing? I know there's a gpu roadmap for k8s, but I can't tell from this blog if this was part of it.
Canonical will officially support GPUs when they lands GA upstream. The flag is beta as of now in the Canonical Distribution of Kubernetes.
Paying customers either for the managed or supported solutions get a best effort for GPU, and this feature is enabled by default.
Re: Putting GPUs to work with Kubernetes
#9Earlier quoted context omitted.
https://vishh.github.io/docs/concepts/overview/what-is-kuber...
If you'd like to engage with me further, how does a company know it needs Kubernetes? If I'm Soylent and I'm processing a few orders a minute, I'm probably safe with a few redundant monoliths. Do I have to be Uber? What's the middle-ground between Soylent and Uber that would still need this? Is the answer the same as the question "who needs a microservice architecture"?
There are a few killer features that you would benefit at any size and that I really love
* self healing: when you create a deployment/replica set. it will be maintained at all cost, so if the app has a memory leak or anything goes wrong, it will be contained and kept up and running
* Rolling update: even when you run 5 frontends, it is a pain to use capistrano or other tools to just update at git repo. it is literally a one liner in Kubernetes. If you use CI/CD the setup is just a few lines in any Jenkins/Gitlab/Travis...
* Service discovery: the combination of ENV and predictable DNS endpoints is just awesome
* Ecosystem: PaaS, Serverless... Many of the new world infra is built on K8s, so it is a door to the next gen, whether you know you will use it or not.
As for Micro Service Architecture, just starting with the web frontend and a couple of lightly dockerize middleware makes it sooooo simple that you instantly want to get more out of it.
As the overhead of running K8s vs. set of servers is relatively low especially at small scale, it is definitely worth looking at. Happy to do a run through with you and show you how the deployment of a tiered app works as a demo, ping me on @SaMnCo_23 if interested.
Re: Putting GPUs to work with Kubernetes
#10Unprofitable for ETH mining maybe, but it seems like a natural fit to rent time on it to deep learning people with slow training models. Although that could still be unprofitable after the cost of electricity, I guess it's a question of market size/demand. A lot of deep learning is already at big infrastructure players anyway who wouldn't need the service, leaving academics / smaller companies. But maybe some people…
Ahah, good point. Really the ETH stuff was "because I can". But in the same charts repository you will find a Tensorflow chart. My previous series of blogs [0] was about exactly that. A nice addition as well for compute intensive workloads is the use of LXD [1]
Another use case is in media for transcoding. It is not a trivial job to orchestrate transcoding at scale, and Kubernetes with or without GPUs is an excellent solution for that as it is trivial to setup a completely automated job queue.
Also another interesting field will eventually be HPC but there are some constraints about compute that K8s does not tick scheduling wise at this point in time. There is a pluggable scheduler in the works I think, and this will eventually help. Also the LXD example is a nice optimization but it would not replace the scheduler in any way.
[0]: https://medium.com/intuitionmachine/gpus-kubernetes-for-deep...
[1]: https://hackernoon.com/job-concurrency-in-kubernetes-lxd-and...