Steps to reproduce Windows AMI 635198895848/codex-preprocess-1.0.0, ami-xxx
Following this guide is only required to re-create a CODEX analysis server from a base Windows 2019 server image. The Amazon Machine image described in the README is preconfigured for CODEX processing.
The base Amazon Machine Image used is the latest official Microsoft Windows Server 2019 with NVIDIA Tesla Driver image, described here.
For security, we rely on incoming connection restriction at the AWS Security Group level. Under Server Management, disable for all regions:
- Windows Defender Firewall
- IE Enhanced Security Configuration
- Windows Defender Antivirus
- install pipeline dependencies
- (NVIDIA Tesla drivers compatible with G5 instances are pre-installed)
- CUDA (latest 11.x)
- Matlab 2021b
- additional toolboxes
- Simulink
- Bioinformatics
- Computer Vision
- Databse
- Deep learning
- GPU Coder
- Image Aquisition
- Image Processing
- Matlab Coder
- Matlab Compiler
- Optimization
- Parallel Computing
- Spreadsheet link
- Statistics and Machine learning
- additional toolboxes
- ImageJ (pinned, custom version), extracted to
\C:\Program Files.
- install dependencies for parallelization and mounting remote file storage
- git bash (latest), used to run Unix pipeline tools on Windows
- GNU Parallel via git bash
- rclone (latest), used to mount remote storage to the instance
Y:\drive to avoid copying files over the network in bulk - WinFSP (latest), an
rclonedependency
You will be required to activate Matlab using your own license and configure access via rclone to input data.
- mysys2, used for Unix shell emulation
- AWS comand line interface
- Google Chrome
- VS Code
Finally, download this repository and extract to \C:\Program Files\codex-preprocess. Then move the two *.jar files in the source code top level directory to the Matlab jar folder as described in the README.