Docker folding@home conatiner

Debian stretch imagine with folding at home set to run as part of the containers team. You can override all settings by adding the flags to the end of the docker run command:

"--user=YOUR_NAME", "--team=TEAM_NUMBER", "--gpu=false", "--smp=true"

Currently COVID19 is taking prioroty so its a great cause that we all need right now.

Running manually

To run simply type:

docker run -d -p7396:7396 richarvey/foldingathome:latest

You can get to the webUI on http://localhost:7396

Deploying on ECS/Fargate Using CDK

We've included CDK fiels to help you deploy on AWS in fargate, check out https://docs.aws.amazon.com/cdk/latest/guide/getting_started.html to get setup.

The cdk.json file tells the CDK Toolkit how to execute your app.

Once installed run:

cd cdk
npm install
npm run build
cdk deploy

Useful commands

• npm install -g aws-cdk install the aws cdk
• npm install install all required modules
• npm run build compile typescript to js
• cdk diff compare deployed stack with current state
• cdk synth emits the synthesized CloudFormation template
• cdk deploy deploy this stack to your default AWS account/region

Manually switching to SPOT (save some money)

Get your account id:

aws sts get-caller-identity

Using the aws CLI and changing the ARN / region:

aws ecs put-cluster-capacity-providers --cluster arn:aws:ecs:eu-west-1:YOUR_ACCOUNT_ID:cluster/CdkfoldingStack-FoldingClusterYOUR_CLUSTER_UID --capacity-providers FARGATE FARGATE_SPOT --default-capacity-provider-strategy capacityProvider=FARGATE_SPOT,weight=100 --region eu-west-1

Running on Kubernetes

To run on kubernetes in AWS, GCP, Azure and other simply run the following:

cd k8s
kubectl create namespace dev
kubectl apply -f foldingathome-deployment.yaml

GPU Support on EKS

You'll need eksctl installed for the following instructions

First of all build your EKS cluster with p2 instances

eksctl create cluster --node-type=p2.xlarge

Now enable the NVIDA drivers

kubectl apply -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/master/nvidia-device-plugin.yml

Now launch the deployment:

cd k8s
kubectl create namespace dev
kubectl apply -f foldingathome-GPU-deployment.yaml