├── NOTICE ├── README.md ├── LICENSE └── machinelearning.template /NOTICE: -------------------------------------------------------------------------------- 1 | Copyright 2016 Amazon.com, Inc. or its affiliates. All Rights Reserved. 2 | 3 | Licensed under the Apache License, Version 2.0 (the "License"). You may not use this file except in compliance with the License. A copy of the License is located at 4 | 5 | http://aws.amazon.com/apache2.0/ 6 | 7 | or in the "license" file accompanying this file. This file is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. 8 | 9 | ecs-machine-learning 10 | Copyright 2016 Amazon.com, Inc. or its affiliates. All Rights Reserved. 11 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | 2 | # Orchestrating GPU-Accelerated Workloads on Amazon ECS 3 | 4 | ## Background 5 | AWS Solutions Architects are seeing an emerging type of application for ECS: GPU-accelerated workloads, or, more specifically, workloads that need to leverage large amounts of GPUs across many nodes. For example, at Amazon.com, the Amazon Personalization Team runs significant Machine Learning workloads that leverage many GPUs on Amazon ECS. Let’s take a look at how ECS enables GPU workloads. 6 | 7 | ## Solution 8 | 9 | In order to run GPU-enabled work on an ECS cluster, a Docker image configured with [Nvidia CUDA drivers][1], which allow the container to communicate with the GPU hardware, is built and stored in EC2 Container Registry. An [ECS Task Definition][2] is used to point to the container image in ECR and specify configuration for the container at runtime, like how much CPU and memory each container should use, the command to run inside the container, if a data volume should be mounted in the container, where the source dataset lives in S3, and so on. 10 | 11 | Once the ECS Tasks are run, the ECS [scheduler][3] finds a suitable place to run the containers by identifying an instance in the cluster with available resources. As shown in the below architecture diagram, ECS can place containers into the cluster of GPU instances (“GPU slaves” in the diagram) 12 | 13 | 14 | 15 | ## Deploying the architecture 16 | 17 | In this template, we spin up an ECS Cluster with a single GPU instance in an autoscaling group. You can, however, adjust the ASG desired capacity to run a larger cluster if you’d like. The instance is configured with all of the necessary software, like Nvidia drivers, that DSSTNE requires for interaction with the underlying GPU hardware. We also install some development tools, like Make and GCC, so that we can compile the DSSTNE library at boot time. We then build a Docker container with the DSSTNE library packaged up and upload it to EC2 Container Registry. We grab the URL of the resulting container image in ECR and build an ECS Task Definition that points to the container. 18 | 19 | Once the CloudFormation template completes, take a look at the “Outputs” tab to get an idea of where to look for your new resources. 20 | 21 | ### Prerequisistes 22 | 23 | #### Network configuration 24 | The instances launched will need to have access to the internet hence either be in a public subnet and provided a public IP or in a private subnet with acess to a NAT gateway. 25 | 26 | #### Accepting terms 27 | 28 | 0. Accept AWS Marketplace terms for Amazon Linux AMI with NVIDIA GRID GPU Driver by going to the [MarketPlace page][9] 29 | 30 | 1. Click Continue on the right. 31 | 32 | 2. Click on the Manual Launch tab and click on the Accept Software Terms button. 33 | 34 | 3. Wait for an email confirmation that your marketplace subscription is active. 35 | 36 | ### Launch the stack 37 | 38 | 2. Choose **Launch Stack** to launch the template in the us-east-1 region in your account: 39 | [![Launch ECS Machine Learning into North Virginia with CloudFormation](http://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/images/cloudformation-launch-stack-button.png)](https://console.aws.amazon.com/cloudformation/home?region=us-east-1#/stacks/new?stackName=ecs-machine-learning&templateURL=https://s3.amazonaws.com/ecs-machine-learning/machinelearning.template) 40 | 41 | (The template will build a DSSTNE container on the ECS cluster instance. Note this can take up to 25 minutes and the CloudFormation stack will not report completion until the entire build process is done.) 42 | 43 | 2. Give a Stack Name and select your preferred key name. If you do not have a key available, see [Amazon EC2 Key Pairs][4]. 44 | 45 | ### Run the model 46 | 47 | 3. Find the name of the DSSTNE ECS Task Definition in CloudFormation stack outputs. It will start with "arn:aws[...]" and contain the CloudFormation template name right after "task-defintion/". 48 | 49 | 4. Go to the [ECS Console][10], click on Task Definitions (left column) and find the one you spotted in the step above. 50 | 51 | 5. Tick the one revision you see. Click on the Actions drop-down menu and hit Run task. Make sure to select the ECS Cluster that was brought up by the CloudFormation template. By running this task, you are essentially running the DSSTNE sample modeling as described on the [amazon-dsstne GitHub page][5]. 52 | 53 | 6. You can easily check that the GPU is being used by logging to the EC2 instance and running `watch -n1 nvidia-smi` 54 | 55 | ### Collect predictions 56 | 57 | 5. You should be able to find the name of the relevant [CloudWatch][7] Logs Group in CloudFormation stack outputs 58 | 59 | 6. Look at the task logs for details and output from the task run and location of results file in S3 60 | 61 | 8. Navigate to this S3 bucket via the [S3 Console][11]. This is where you will be able to access the results file and confirm that this GPU-enabled Machine Learning run was successful. 62 | 63 | ### Bonus activities 64 | - Bonus activity #1: Repeat step 2 in the *Run the model* section, but change the config URL and training command by overriding the task definition environment variables to perform a [benchmark][6]. 65 | - Bonus activity #2: Modify the CloudFormation template (or launch a new stack) and use a g2.8xlarge instead of g2.2xlarge (you could also try a [P2 instance][12]...) Repeat step 2 in the *Run the model* section, but override the training command in the task definition environment variables to use MPI to take advantage of all 4 GPUs: add `mpirun -np ` in front of the `train` command. For this you will have to create a new revision of the task definition. 66 | 67 | In both bonus steps, look at the CloudWatch Logs to view the task logs (different training commands, taking advantage of multiple GPUs, etc.) 68 | 69 | ## Conclusion 70 | 71 | You should now have a good grasp on how to leverage ECS and GPU-optimized EC2 instances for your Machine Learning needs. Head on over to the [AWS Big Data blog][8] to learn more about how DSSTNE interacts with Apache Spark, trains models, generates predictions, and other fun Machine Learning concepts. 72 | 73 | 74 | [1]: http://www.nvidia.com/object/cuda_home_new.html 75 | [2]: http://docs.aws.amazon.com/AmazonECS/latest/developerguide/task_defintions.html 76 | [3]: http://docs.aws.amazon.com/AmazonECS/latest/developerguide/scheduling_tasks.html 77 | [4]: http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-key-pairs.html 78 | [5]: https://github.com/amznlabs/amazon-dsstne/blob/master/docs/getting_started/examples.md 79 | [6]: https://github.com/amznlabs/amazon-dsstne/blob/master/benchmarks/Benchmark.md 80 | [7]: https://console.aws.amazon.com/cloudwatch/home 81 | [8]: https://blogs.aws.amazon.com/bigdata/post/TxGEL8IJ0CAXTK/Generating-Recommendations-at-Amazon-Scale-with-Apache-Spark-and-Amazon-DSSTNE 82 | [9]: https://aws.amazon.com/marketplace/pp/B00FYCDDTE 83 | [10]: https://console.aws.amazon.com/ecs/ 84 | [11]: https://console.aws.amazon.com/s3/home 85 | [12]: https://aws.amazon.com/ec2/instance-types/p2/ 86 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | 2 | Apache License 3 | Version 2.0, January 2004 4 | http://www.apache.org/licenses/ 5 | 6 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 7 | 8 | 1. Definitions. 9 | 10 | "License" shall mean the terms and conditions for use, reproduction, 11 | and distribution as defined by Sections 1 through 9 of this document. 12 | 13 | "Licensor" shall mean the copyright owner or entity authorized by 14 | the copyright owner that is granting the License. 15 | 16 | "Legal Entity" shall mean the union of the acting entity and all 17 | other entities that control, are controlled by, or are under common 18 | control with that entity. For the purposes of this definition, 19 | "control" means (i) the power, direct or indirect, to cause the 20 | direction or management of such entity, whether by contract or 21 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 22 | outstanding shares, or (iii) beneficial ownership of such entity. 23 | 24 | "You" (or "Your") shall mean an individual or Legal Entity 25 | exercising permissions granted by this License. 26 | 27 | "Source" form shall mean the preferred form for making modifications, 28 | including but not limited to software source code, documentation 29 | source, and configuration files. 30 | 31 | "Object" form shall mean any form resulting from mechanical 32 | transformation or translation of a Source form, including but 33 | not limited to compiled object code, generated documentation, 34 | and conversions to other media types. 35 | 36 | "Work" shall mean the work of authorship, whether in Source or 37 | Object form, made available under the License, as indicated by a 38 | copyright notice that is included in or attached to the work 39 | (an example is provided in the Appendix below). 40 | 41 | "Derivative Works" shall mean any work, whether in Source or Object 42 | form, that is based on (or derived from) the Work and for which the 43 | editorial revisions, annotations, elaborations, or other modifications 44 | represent, as a whole, an original work of authorship. For the purposes 45 | of this License, Derivative Works shall not include works that remain 46 | separable from, or merely link (or bind by name) to the interfaces of, 47 | the Work and Derivative Works thereof. 48 | 49 | "Contribution" shall mean any work of authorship, including 50 | the original version of the Work and any modifications or additions 51 | to that Work or Derivative Works thereof, that is intentionally 52 | submitted to Licensor for inclusion in the Work by the copyright owner 53 | or by an individual or Legal Entity authorized to submit on behalf of 54 | the copyright owner. For the purposes of this definition, "submitted" 55 | means any form of electronic, verbal, or written communication sent 56 | to the Licensor or its representatives, including but not limited to 57 | communication on electronic mailing lists, source code control systems, 58 | and issue tracking systems that are managed by, or on behalf of, the 59 | Licensor for the purpose of discussing and improving the Work, but 60 | excluding communication that is conspicuously marked or otherwise 61 | designated in writing by the copyright owner as "Not a Contribution." 62 | 63 | "Contributor" shall mean Licensor and any individual or Legal Entity 64 | on behalf of whom a Contribution has been received by Licensor and 65 | subsequently incorporated within the Work. 66 | 67 | 2. Grant of Copyright License. Subject to the terms and conditions of 68 | this License, each Contributor hereby grants to You a perpetual, 69 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 70 | copyright license to reproduce, prepare Derivative Works of, 71 | publicly display, publicly perform, sublicense, and distribute the 72 | Work and such Derivative Works in Source or Object form. 73 | 74 | 3. Grant of Patent License. Subject to the terms and conditions of 75 | this License, each Contributor hereby grants to You a perpetual, 76 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 77 | (except as stated in this section) patent license to make, have made, 78 | use, offer to sell, sell, import, and otherwise transfer the Work, 79 | where such license applies only to those patent claims licensable 80 | by such Contributor that are necessarily infringed by their 81 | Contribution(s) alone or by combination of their Contribution(s) 82 | with the Work to which such Contribution(s) was submitted. If You 83 | institute patent litigation against any entity (including a 84 | cross-claim or counterclaim in a lawsuit) alleging that the Work 85 | or a Contribution incorporated within the Work constitutes direct 86 | or contributory patent infringement, then any patent licenses 87 | granted to You under this License for that Work shall terminate 88 | as of the date such litigation is filed. 89 | 90 | 4. Redistribution. You may reproduce and distribute copies of the 91 | Work or Derivative Works thereof in any medium, with or without 92 | modifications, and in Source or Object form, provided that You 93 | meet the following conditions: 94 | 95 | (a) You must give any other recipients of the Work or 96 | Derivative Works a copy of this License; and 97 | 98 | (b) You must cause any modified files to carry prominent notices 99 | stating that You changed the files; and 100 | 101 | (c) You must retain, in the Source form of any Derivative Works 102 | that You distribute, all copyright, patent, trademark, and 103 | attribution notices from the Source form of the Work, 104 | excluding those notices that do not pertain to any part of 105 | the Derivative Works; and 106 | 107 | (d) If the Work includes a "NOTICE" text file as part of its 108 | distribution, then any Derivative Works that You distribute must 109 | include a readable copy of the attribution notices contained 110 | within such NOTICE file, excluding those notices that do not 111 | pertain to any part of the Derivative Works, in at least one 112 | of the following places: within a NOTICE text file distributed 113 | as part of the Derivative Works; within the Source form or 114 | documentation, if provided along with the Derivative Works; or, 115 | within a display generated by the Derivative Works, if and 116 | wherever such third-party notices normally appear. The contents 117 | of the NOTICE file are for informational purposes only and 118 | do not modify the License. You may add Your own attribution 119 | notices within Derivative Works that You distribute, alongside 120 | or as an addendum to the NOTICE text from the Work, provided 121 | that such additional attribution notices cannot be construed 122 | as modifying the License. 123 | 124 | You may add Your own copyright statement to Your modifications and 125 | may provide additional or different license terms and conditions 126 | for use, reproduction, or distribution of Your modifications, or 127 | for any such Derivative Works as a whole, provided Your use, 128 | reproduction, and distribution of the Work otherwise complies with 129 | the conditions stated in this License. 130 | 131 | 5. Submission of Contributions. Unless You explicitly state otherwise, 132 | any Contribution intentionally submitted for inclusion in the Work 133 | by You to the Licensor shall be under the terms and conditions of 134 | this License, without any additional terms or conditions. 135 | Notwithstanding the above, nothing herein shall supersede or modify 136 | the terms of any separate license agreement you may have executed 137 | with Licensor regarding such Contributions. 138 | 139 | 6. Trademarks. This License does not grant permission to use the trade 140 | names, trademarks, service marks, or product names of the Licensor, 141 | except as required for reasonable and customary use in describing the 142 | origin of the Work and reproducing the content of the NOTICE file. 143 | 144 | 7. Disclaimer of Warranty. Unless required by applicable law or 145 | agreed to in writing, Licensor provides the Work (and each 146 | Contributor provides its Contributions) on an "AS IS" BASIS, 147 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 148 | implied, including, without limitation, any warranties or conditions 149 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 150 | PARTICULAR PURPOSE. You are solely responsible for determining the 151 | appropriateness of using or redistributing the Work and assume any 152 | risks associated with Your exercise of permissions under this License. 153 | 154 | 8. Limitation of Liability. In no event and under no legal theory, 155 | whether in tort (including negligence), contract, or otherwise, 156 | unless required by applicable law (such as deliberate and grossly 157 | negligent acts) or agreed to in writing, shall any Contributor be 158 | liable to You for damages, including any direct, indirect, special, 159 | incidental, or consequential damages of any character arising as a 160 | result of this License or out of the use or inability to use the 161 | Work (including but not limited to damages for loss of goodwill, 162 | work stoppage, computer failure or malfunction, or any and all 163 | other commercial damages or losses), even if such Contributor 164 | has been advised of the possibility of such damages. 165 | 166 | 9. Accepting Warranty or Additional Liability. While redistributing 167 | the Work or Derivative Works thereof, You may choose to offer, 168 | and charge a fee for, acceptance of support, warranty, indemnity, 169 | or other liability obligations and/or rights consistent with this 170 | License. However, in accepting such obligations, You may act only 171 | on Your own behalf and on Your sole responsibility, not on behalf 172 | of any other Contributor, and only if You agree to indemnify, 173 | defend, and hold each Contributor harmless for any liability 174 | incurred by, or claims asserted against, such Contributor by reason 175 | of your accepting any such warranty or additional liability. 176 | 177 | END OF TERMS AND CONDITIONS 178 | 179 | APPENDIX: How to apply the Apache License to your work. 180 | 181 | To apply the Apache License to your work, attach the following 182 | boilerplate notice, with the fields enclosed by brackets "[]" 183 | replaced with your own identifying information. (Don't include 184 | the brackets!) The text should be enclosed in the appropriate 185 | comment syntax for the file format. We also recommend that a 186 | file or class name and description of purpose be included on the 187 | same "printed page" as the copyright notice for easier 188 | identification within third-party archives. 189 | 190 | Copyright 2016 Amazon Web Services 191 | 192 | Licensed under the Apache License, Version 2.0 (the "License"); 193 | you may not use this file except in compliance with the License. 194 | You may obtain a copy of the License at 195 | 196 | http://www.apache.org/licenses/LICENSE-2.0 197 | 198 | Unless required by applicable law or agreed to in writing, software 199 | distributed under the License is distributed on an "AS IS" BASIS, 200 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 201 | See the License for the specific language governing permissions and 202 | limitations under the License. 203 | -------------------------------------------------------------------------------- /machinelearning.template: -------------------------------------------------------------------------------- 1 | AWSTemplateFormatVersion: '2010-09-09' 2 | Description: Creates environment for running Amazon DSSTNE on EC2 Container Service 3 | Mappings: 4 | AmazonLinuxAMI: 5 | ap-northeast-1: {AMI: ami-78ba6619} 6 | ap-southeast-1: {AMI: ami-56e84c35} 7 | ap-southeast-2: {AMI: ami-2589b946} 8 | eu-central-1: {AMI: ami-d9d62ab6} 9 | eu-west-1: {AMI: ami-b9bd25ca} 10 | us-east-1: {AMI: ami-50b4f047} 11 | us-west-1: {AMI: ami-699ad409} 12 | us-west-2: {AMI: ami-002bf460} 13 | Metadata: 14 | AWS::CloudFormation::Interface: 15 | ParameterGroups: 16 | - Label: {default: SSH Access} 17 | Parameters: [KeyName, SSHLocation] 18 | - Label: {default: 'VPC, Subnets, and Instance'} 19 | Parameters: [VpcId, Subnets, ECSInstanceType] 20 | - Label: {default: Data and Config} 21 | Parameters: [DataUrl, TrainingConfigUrl] 22 | ParameterLabels: 23 | DataUrl: {default: Data URL} 24 | ECSInstanceType: {default: ECS instance type} 25 | KeyName: {default: Key name} 26 | SSHLocation: {default: SSH CIDR range} 27 | Subnets: {default: Subnets} 28 | TrainingConfigUrl: {default: Training config URL} 29 | VpcId: {default: VPC} 30 | Maintainer: {Description: Chad Schmutzer } 31 | Outputs: 32 | CloudWatchLogsGroup: 33 | Description: Name of the CloudWatch Logs Group 34 | Value: {Ref: CloudWatchLogsGroup} 35 | DsstneTaskDefinition: 36 | Description: The task definition for Amazon DSSTNE 37 | Value: {Ref: ECSTaskDefinitionDsstne} 38 | ECRRepository: 39 | Description: The ECR Repository for the Amazon DSSTNE container 40 | Value: {Ref: ECRRepository} 41 | ECSClusterName: 42 | Description: Name of the ECS cluster 43 | Value: {Ref: ECSCluster} 44 | S3BucketResults: 45 | Description: The S3 bucket for storing results 46 | Value: {Ref: S3BucketResults} 47 | Parameters: 48 | DataUrl: {ConstraintDescription: Must be valid URL, Default: 'https://s3-us-west-2.amazonaws.com/amazon-dsstne-samples/data/ml20m-all', 49 | Description: Neural Network Modeling data URL, Type: String} 50 | ECSInstanceType: 51 | AllowedValues: [g2.2xlarge, g2.8xlarge, p2.xlarge, p2.8xlarge, p2.16xlarge] 52 | ConstraintDescription: Must be g2.2xlarge, g2.8xlarge, p2.xlarge, p2.8xlarge, 53 | p2.16xlarge 54 | Default: g2.2xlarge 55 | Description: ECS instance type 56 | Type: String 57 | KeyName: {Description: Name of an existing EC2 KeyPair, Type: 'AWS::EC2::KeyPair::KeyName'} 58 | SSHLocation: {AllowedPattern: '(\d{1,3})\.(\d{1,3})\.(\d{1,3})\.(\d{1,3})/(\d{1,2})', 59 | ConstraintDescription: Must be a valid CIDR range of the form x.x.x.x/x, Default: 0.0.0.0/0, 60 | Description: Restrict SSH access to CIDR range (default can be accessed from anywhere), 61 | MaxLength: '18', MinLength: '9', Type: String} 62 | Subnets: {ConstraintDescription: Must be a list of existing subnets in the selected 63 | Virtual Private Cloud, Description: Please make sure you select 1 or more subnets 64 | in the Virtual Private Cloud (VPC) ID you chose above, Type: 'List'} 65 | TrainingConfigUrl: {ConstraintDescription: Must be a valid URL, Default: 'https://s3-us-west-2.amazonaws.com/amazon-dsstne-samples/configs/config.json', 66 | Description: Training config URL, Type: String} 67 | VpcId: {ConstraintDescription: Must be the VPC ID of an existing Virtual Private 68 | Cloud, Description: VPC ID of your existing Virtual Private Cloud (VPC), Type: 'AWS::EC2::VPC::Id'} 69 | Resources: 70 | CloudWatchLogsGroup: 71 | Properties: {RetentionInDays: 7} 72 | Type: AWS::Logs::LogGroup 73 | ECRRepository: {Type: 'AWS::ECR::Repository'} 74 | ECSCluster: {Type: 'AWS::ECS::Cluster'} 75 | ECSInstanceAutoScalingGroup: 76 | CreationPolicy: 77 | ResourceSignal: {Count: '1', Timeout: PT120M} 78 | Properties: 79 | DesiredCapacity: '1' 80 | LaunchConfigurationName: {Ref: ECSInstanceLaunchConfiguration} 81 | MaxSize: '1' 82 | MinSize: '1' 83 | Tags: 84 | - Key: Name 85 | PropagateAtLaunch: 'true' 86 | Value: {Ref: 'AWS::StackName'} 87 | VPCZoneIdentifier: {Ref: Subnets} 88 | Type: AWS::AutoScaling::AutoScalingGroup 89 | ECSInstanceLaunchConfiguration: 90 | Metadata: 91 | AWS::CloudFormation::Init: 92 | config: 93 | files: 94 | /etc/ecs/ecs.config: 95 | content: 96 | Fn::Join: 97 | - '' 98 | - - ECS_CLUSTER= 99 | - {Ref: ECSCluster} 100 | - ' 101 | 102 | ' 103 | - 'ECS_AVAILABLE_LOGGING_DRIVERS=["json-file","awslogs"] 104 | 105 | ' 106 | group: root 107 | mode: '000644' 108 | owner: root 109 | /home/ec2-user/.aws/config: 110 | content: 111 | Fn::Join: 112 | - '' 113 | - - '[default] 114 | 115 | ' 116 | - 'region = ' 117 | - {Ref: 'AWS::Region'} 118 | - ' 119 | 120 | ' 121 | group: root 122 | mode: '000644' 123 | owner: ec2-user 124 | /home/ec2-user/Dockerfile: 125 | content: 126 | Fn::Join: 127 | - '' 128 | - ['FROM nvidia/cuda:7.5-cudnn5-devel 129 | 130 | ', ' 131 | 132 | ', 'RUN apt-get update && \ 133 | 134 | ', ' apt-get -y dist-upgrade && \ 135 | 136 | ', ' apt-get -y install \ 137 | 138 | ', ' kmod \ 139 | 140 | ', ' make \ 141 | 142 | ', ' build-essential \ 143 | 144 | ', ' cmake \ 145 | 146 | ', ' cpp \ 147 | 148 | ', ' g++ \ 149 | 150 | ', ' gcc \ 151 | 152 | ', ' libatlas-base-dev \ 153 | 154 | ', ' curl \ 155 | 156 | ', ' python-pip \ 157 | 158 | ', ' openmpi-bin \ 159 | 160 | ', ' libopenmpi-dev \ 161 | 162 | ', ' libjsoncpp-dev \ 163 | 164 | ', ' libhdf5-dev \ 165 | 166 | ', ' openssh-client \ 167 | 168 | ', ' zlib1g-dev 169 | 170 | ', ' 171 | 172 | ', 'RUN pip install awscli 173 | 174 | ', ' 175 | 176 | ', 'ENV GPU_DRIVER_VERSION=352.99 177 | 178 | ', ' 179 | 180 | ', 'RUN cd /tmp && \ 181 | 182 | ', ' curl -LO http://us.download.nvidia.com/XFree86/Linux-x86_64/$GPU_DRIVER_VERSION/NVIDIA-Linux-x86_64-$GPU_DRIVER_VERSION.run 183 | && \ 184 | 185 | ', ' chmod +x ./NVIDIA-Linux-x86_64-$GPU_DRIVER_VERSION.run && 186 | \ 187 | 188 | ', ' ./NVIDIA-Linux-x86_64-$GPU_DRIVER_VERSION.run -s --no-kernel-module 189 | && \ 190 | 191 | ', ' rm -rf /tmp/* 192 | 193 | ', ' 194 | 195 | ', 'RUN ln -s /usr/include/hdf5/serial/* /usr/include/ 196 | 197 | ', ' 198 | 199 | ', 'RUN ln -s /usr/lib/x86_64-linux-gnu/hdf5/serial/libhdf5* /usr/lib/x86_64-linux-gnu/ 200 | 201 | ', ' 202 | 203 | ', 'RUN cd /tmp && \ 204 | 205 | ', ' curl -LO ftp://ftp.unidata.ucar.edu/pub/netcdf/netcdf-4.1.3.tar.gz 206 | && \ 207 | 208 | ', ' tar xvf netcdf-4.1.3.tar.gz && \ 209 | 210 | ', ' cd netcdf-4.1.3 && \ 211 | 212 | ', ' ./configure --prefix=/usr/local && \ 213 | 214 | ', ' make -j16 && \ 215 | 216 | ', ' make install && rm -rf /tmp/* 217 | 218 | ', ' 219 | 220 | ', 'RUN cd /tmp && \ 221 | 222 | ', ' curl -LO http://www.unidata.ucar.edu/downloads/netcdf/ftp/netcdf-cxx4-4.2.tar.gz 223 | && \ 224 | 225 | ', ' tar xvf netcdf-cxx4-4.2.tar.gz && \ 226 | 227 | ', ' cd netcdf-cxx4-4.2 && \ 228 | 229 | ', ' ./configure --prefix=/usr/local && \ 230 | 231 | ', ' make -j16 && \ 232 | 233 | ', ' make install && rm -rf /tmp/* 234 | 235 | ', ' 236 | 237 | ', 'RUN cd /tmp && \ 238 | 239 | ', ' curl -LO https://github.com/NVlabs/cub/archive/1.5.2.zip 240 | && \ 241 | 242 | ', ' apt-get install -y unzip && \ 243 | 244 | ', ' unzip 1.5.2.zip && \ 245 | 246 | ', ' cp -rf cub-1.5.2/cub/ /usr/local/include/ && \ 247 | 248 | ', ' rm -rf /tmp/* 249 | 250 | ', ' 251 | 252 | ', 'ENV PATH=/usr/local/openmpi/bin/:/usr/local/cuda/bin/:${PATH} 253 | \ 254 | 255 | ', ' LD_LIBRARY_PATH=/usr/local/lib/:${LD_LIBRARY_PATH} 256 | 257 | ', ' 258 | 259 | ', 'RUN ln -sf /usr/lib/openmpi /usr/local/openmpi 260 | 261 | ', ' 262 | 263 | ', 'COPY src /opt/amazon/dsstne/src 264 | 265 | ', ' 266 | 267 | ', 'RUN sed -i "s/-gencode arch=compute_60,code=sm_60//" /opt/amazon/dsstne/src/amazon/dsstne/Makefile.inc 268 | 269 | ', ' 270 | 271 | ', 'RUN cd /opt/amazon/dsstne/src/amazon/dsstne && \ 272 | 273 | ', ' make 274 | 275 | ', ' 276 | 277 | ', 'ENV PATH /opt/amazon/dsstne/src/amazon/dsstne/bin:${PATH} 278 | 279 | '] 280 | group: root 281 | mode: '000644' 282 | owner: root 283 | /root/.aws/config: 284 | content: 285 | Fn::Join: 286 | - '' 287 | - - '[default] 288 | 289 | ' 290 | - 'region = ' 291 | - {Ref: 'AWS::Region'} 292 | - ' 293 | 294 | ' 295 | group: root 296 | mode: '000644' 297 | owner: root 298 | packages: 299 | yum: 300 | docker: [] 301 | ecs-init: [] 302 | git: [] 303 | services: 304 | sysvinit: 305 | docker: {enabled: 'true', ensureRunning: 'true'} 306 | Properties: 307 | BlockDeviceMappings: 308 | - DeviceName: /dev/xvda 309 | Ebs: {VolumeSize: '30'} 310 | IamInstanceProfile: {Ref: ECSInstanceProfile} 311 | ImageId: 312 | Fn::FindInMap: 313 | - AmazonLinuxAMI 314 | - {Ref: 'AWS::Region'} 315 | - AMI 316 | InstanceType: {Ref: ECSInstanceType} 317 | KeyName: {Ref: KeyName} 318 | SecurityGroups: 319 | - {Ref: ECSInstanceSecurityGroup} 320 | UserData: 321 | Fn::Base64: 322 | Fn::Join: 323 | - '' 324 | - - '#!/bin/bash -xe 325 | 326 | ' 327 | - 'yum -y update --exclude=kernel* --exclude=nvidia* 328 | 329 | ' 330 | - '# Install the files and packages from the metadata 331 | 332 | ' 333 | - '/opt/aws/bin/cfn-init -v ' 334 | - ' --stack ' 335 | - {Ref: 'AWS::StackName'} 336 | - ' --resource ECSInstanceLaunchConfiguration ' 337 | - ' --region ' 338 | - {Ref: 'AWS::Region'} 339 | - ' 340 | 341 | ' 342 | - '/sbin/start ecs 343 | 344 | ' 345 | - 'cd /root/ && git clone https://github.com/amznlabs/amazon-dsstne.git 346 | 347 | ' 348 | - 'cd /root/amazon-dsstne/ 349 | 350 | ' 351 | - 'cp /home/ec2-user/Dockerfile . 352 | 353 | ' 354 | - 'docker build -t ' 355 | - {Ref: ECRRepository} 356 | - ' . 357 | 358 | ' 359 | - 'aws ecr get-login | sh 360 | 361 | ' 362 | - 'docker tag ' 363 | - {Ref: ECRRepository} 364 | - ':latest ' 365 | - {Ref: 'AWS::AccountId'} 366 | - .dkr.ecr. 367 | - {Ref: 'AWS::Region'} 368 | - .amazonaws.com/ 369 | - {Ref: ECRRepository} 370 | - ':latest 371 | 372 | ' 373 | - 'docker push ' 374 | - {Ref: 'AWS::AccountId'} 375 | - .dkr.ecr. 376 | - {Ref: 'AWS::Region'} 377 | - .amazonaws.com/ 378 | - {Ref: ECRRepository} 379 | - ':latest 380 | 381 | ' 382 | - '# Signal the status from cfn-init 383 | 384 | ' 385 | - '/opt/aws/bin/cfn-signal -e $? ' 386 | - ' --stack ' 387 | - {Ref: 'AWS::StackName'} 388 | - ' --resource ECSInstanceAutoScalingGroup ' 389 | - ' --region ' 390 | - {Ref: 'AWS::Region'} 391 | - ' 392 | 393 | ' 394 | Type: AWS::AutoScaling::LaunchConfiguration 395 | ECSInstanceProfile: 396 | DependsOn: ECSInstanceRole 397 | Properties: 398 | Path: / 399 | Roles: 400 | - {Ref: ECSInstanceRole} 401 | Type: AWS::IAM::InstanceProfile 402 | ECSInstanceRole: 403 | Properties: 404 | AssumeRolePolicyDocument: 405 | Statement: 406 | - Action: ['sts:AssumeRole'] 407 | Effect: Allow 408 | Principal: 409 | Service: [ec2.amazonaws.com] 410 | Version: '2012-10-17' 411 | ManagedPolicyArns: ['arn:aws:iam::aws:policy/service-role/AmazonEC2ContainerServiceforEC2Role'] 412 | Path: / 413 | Policies: 414 | - PolicyDocument: 415 | Statement: 416 | - Action: s3:ListBucket 417 | Effect: Allow 418 | Resource: 419 | Fn::Join: 420 | - '' 421 | - - 'arn:aws:s3:::' 422 | - {Ref: S3BucketResults} 423 | - Action: ['s3:PutObject', 's3:GetObject', 's3:DeleteObject'] 424 | Effect: Allow 425 | Resource: 426 | Fn::Join: 427 | - '' 428 | - - 'arn:aws:s3:::' 429 | - {Ref: S3BucketResults} 430 | - /* 431 | - Action: ['ecr:DescribeRepositories', 'ecr:ListImages', 'ecr:InitiateLayerUpload', 432 | 'ecr:UploadLayerPart', 'ecr:CompleteLayerUpload', 'ecr:PutImage'] 433 | Effect: Allow 434 | Resource: 435 | Fn::Join: 436 | - '' 437 | - - 'arn:aws:ecr:' 438 | - {Ref: 'AWS::Region'} 439 | - ':' 440 | - {Ref: 'AWS::AccountId'} 441 | - :repository/ 442 | - {Ref: ECRRepository} 443 | Version: '2012-10-17' 444 | PolicyName: Amazon-DSSTNE 445 | Type: AWS::IAM::Role 446 | ECSInstanceSecurityGroup: 447 | Properties: 448 | GroupDescription: Security Group for ECSInstance 449 | SecurityGroupIngress: 450 | - CidrIp: {Ref: SSHLocation} 451 | FromPort: '22' 452 | IpProtocol: tcp 453 | ToPort: '22' 454 | Tags: 455 | - {Key: Name, Value: ECSInstanceSecurityGroup} 456 | VpcId: {Ref: VpcId} 457 | Type: AWS::EC2::SecurityGroup 458 | ECSServiceRole: 459 | Properties: 460 | AssumeRolePolicyDocument: 461 | Statement: 462 | - Action: ['sts:AssumeRole'] 463 | Effect: Allow 464 | Principal: 465 | Service: [ecs.amazonaws.com] 466 | Version: '2012-10-17' 467 | ManagedPolicyArns: ['arn:aws:iam::aws:policy/service-role/AmazonEC2ContainerServiceRole'] 468 | Path: / 469 | Type: AWS::IAM::Role 470 | ECSTaskDefinitionDsstne: 471 | Properties: 472 | ContainerDefinitions: 473 | - Command: ['curl -Lo data $DATAURL && echo "running $GENERATEINPUTCMD" && $GENERATEINPUTCMD 474 | && echo "running $GENERATEOUTPUTCMD" && $GENERATEOUTPUTCMD && curl -Lo 475 | config $TRAININGCONFIGURL && echo "running $TRAINCMD" && $TRAINCMD && 476 | echo "running $PREDICTCMD" && $PREDICTCMD && DATE=`date -Iseconds` && 477 | echo "results being written to s3://$S3BUCKETRESULTS/results.$HOSTNAME.$DATE" 478 | && aws s3 cp results s3://$S3BUCKETRESULTS/results.$HOSTNAME.$DATE && 479 | echo "Task complete!"'] 480 | EntryPoint: [/bin/bash, -c] 481 | Environment: 482 | - Name: DATAURL 483 | Value: {Ref: DataUrl} 484 | - Name: TRAININGCONFIGURL 485 | Value: {Ref: TrainingConfigUrl} 486 | - Name: S3BUCKETRESULTS 487 | Value: {Ref: S3BucketResults} 488 | - {Name: GENERATEINPUTCMD, Value: generateNetCDF -d gl_input -i data -o gl_input.nc 489 | -f features_input -s samples_input -c} 490 | - {Name: GENERATEOUTPUTCMD, Value: generateNetCDF -d gl_output -i data -o 491 | gl_output.nc -f features_output -s samples_input -c} 492 | - {Name: TRAINCMD, Value: train -c config -i gl_input.nc -o gl_output.nc -n 493 | gl.nc -b 256 -e 10} 494 | - {Name: PREDICTCMD, Value: predict -b 1024 -d gl -i features_input -o features_output 495 | -k 10 -n gl.nc -f data -s results -r data} 496 | Image: 497 | Fn::Join: 498 | - '' 499 | - - {Ref: 'AWS::AccountId'} 500 | - .dkr.ecr. 501 | - {Ref: 'AWS::Region'} 502 | - .amazonaws.com/ 503 | - {Ref: ECRRepository} 504 | - :latest 505 | LogConfiguration: 506 | LogDriver: awslogs 507 | Options: 508 | awslogs-group: {Ref: CloudWatchLogsGroup} 509 | awslogs-region: {Ref: 'AWS::Region'} 510 | Memory: '15000' 511 | Name: Amazon-DSSTNE 512 | Privileged: 'true' 513 | Type: AWS::ECS::TaskDefinition 514 | S3BucketResults: 515 | Properties: {AccessControl: BucketOwnerFullControl} 516 | Type: AWS::S3::Bucket 517 | --------------------------------------------------------------------------------