Create New Research Namespace

1Project & Scientific Metadata

sci-

Full Kubernetes namespace identifier: sci-my-project (RFC 1123 format)

2Resource Quota & Accelerator Tiers

Development & Prototyping

For testing workflows, container debugging, and small interactive Jupyter notebooks.

4 Cores16 GB RAM50 GB Storage

Standard Scientific Compute

Ideal for data processing pipelines, multi-core simulation packages, and API servers.

16 Cores64 GB RAM250 GB Storage

GPU Accelerated Research

2x GPU

Optimized for PyTorch, TensorFlow, AlphaFold, Cryo-EM, and deep neural network training.

32 Cores128 GB RAM1000 GB Storage

Custom Heavy Allocation

4x GPU

Tailor custom CPU, memory, GPU, and high-speed NVMe storage limits for large grants.

64 Cores256 GB RAM2000 GB Storage
Fine-Tune Resource Limits
CPU Quota (Cores)32 Cores
RAM Memory Limit128 GB
NVIDIA GPU Allocation2 GPUs
Persistent Storage Volume1000 GB

3Grant / Project Reference & Terms

ℹī¸ Science IT Policy: Compute namespaces are automatically audited for idle resources. Pods inactive for > 30 days may have replica counts scaled down to 0 to maximize GPU efficiency across Berkeley Lab.