Technical Requirements
Copyright © 2026 by Symetry, Inc. 14 Pine Street, Ste 6 Morristown, NJ 07960 All Rights Reserved
| Requirement | Description |
|---|
| Docker Engine | Docker Engine installed on the host |
| Operating System | Any Linux distribution that supports Docker |
| GPU (optional) | NVIDIA GPU with Compute Capability >= 3.5, NVIDIA drivers, and NVIDIA Container Toolkit |
| Spark (optional) | External Spark 4.1.0 cluster |
| Recommendation | Description |
|---|
| Spark Cluster worker memory | Minimum: 8 GB Recommended: 16 GB and more |
SymetryML memory requirements depend on the number of projects and the number of attributes in each project. The following table provides a rough guideline as to the compute and memory requirements for your projects. These numbers are for dense datasets. For sparse datasets, the number will vary.
| Number of Attributes | Type of Project | RAM Memory Requirement |
|---|
| 10 | CPU | 512m |
| 64 | CPU | 512m |
| 128 | CPU or GPU | 1g |
| 512 | GPU | 1g |
| 1024 | GPU | 1g |
| 4096 | GPU or multi-GPU | 4g |
| 8192 | GPU or multi-GPU | 16g |
| 12,500 | GPU or multi-GPU | 24g |
| 20,000-25,000 | multi-GPU | 64g |
| 25,000-100,000 | multi-GPU on Nvidia V100 | 64G to 256G |
| Data Source | Description |
|---|
| JDBC / ODBC | |
| S3 | |
| SFTP | |
| HTTP / HTTPS | |
| Local File | Local file where the application server resides |
| Kafka Streams | |
| RedShift | |
| Plugins | It's possible to add new 'plugin' data source to SymetryML |
| SymetryML Rest API | One can use the SymetryML Rest API to push data into a SymetryML project. Please consult the following section for details (section called Real Time Streaming of Data into SymetryML Projects). |