Machine specifications (human-readable)

Available classes (summary)

Class	Description	vCPUs / GPUs	Memory	Storage	Network
gpu-small	Single-GPU instance for development, small training and inference.	8 CPU, 1 GPU	32 GB	200 GB	1 Gbps
gpu-standard	Multi-GPU instance for model training and larger inference.	32 CPU, 4 GPU	256 GB	1 TB	10 Gbps
cpu-standard	CPU-optimised instances for batch processing, ETL and general compute.	16 CPU	64 GB	500 GB	1-10 Gbps
cpu-large	High-core instances for large-scale batch analytics and rendering.	64 CPU	256 GB	2 TB	10 Gbps

Example GPU models

Examples listed in the JSON; exact models depend on availability. Examples include: NVIDIA T4, A10, A40, and nodes used in multi-GPU A100 deployments.

Developer usage examples

Download the machine JSON from a script or CI pipeline:

curl -sS https://solarcompute.net.au/machines-specs.json -o machines-specs.json
jq . machines-specs.json

Use the JSON to populate instance selection UIs, validate required resources, or pass to quoting tools. For exact inventory and tailored quotes, contact hello@solarcompute.net.au.

Workload guidance: GPU vs CPU

A concise guide to help choose the right class for common workloads. For a developer walkthrough, see the Developer Quickstart.

When to choose GPU

• Deep learning training, fine-tuning and large neural network workloads.
• Model inference requiring low latency or high throughput.
• Video processing, encoding, and highly parallelisable ML tasks.

When to choose CPU

• Batch jobs, ETL, data preprocessing and pipeline orchestration.
• Web services, CI builds and general-purpose compute tasks.
• Workloads that scale horizontally across many smaller cores.

Still unsure? Request a tailored quote: hello@solarcompute.net.au.