Onsite Resources | Center for Precision Medicine and Data Sciences

Computational Resources

Orbital GPU Pro Workstation

• Quantity: 2
• CPU: AMD Ryzen Threadripper 7970X, 32 cores / 64 threads, up to 5.3 GHz turbo
• Memory: 256 GB DDR5 ECC RDIMM
• GPU: 2 × NVIDIA RTX 5090 GPUs, 32 GB GDDR7 each, 21,760 CUDA cores each
• Storage: 4 TB PCIe 5.0 NVMe SSD, up to 14,100 MB/s; 8 TB SATA SSD
• Operating System / Software: Ubuntu 24.04 LTS and molecular modeling/simulation software stack, including AMBER, Rosetta, AlphaFold, Boltz-2
• Primary Use / Purpose: GPU-accelerated molecular dynamics simulations, molecular modeling, docking workflows, machine learning, AI model development, and high-throughput computational analysis

BIZON ZX5500 Deep Learning / AI Workstation

• Quantity: 1
• CPU: AMD Ryzen Threadripper PRO 5995WX, 64 cores / 128 threads, 2.7 GHz base, up to 4.5 GHz boost; 256 MB L3 cache
• Memory: 1024 GB DDR4 ECC/REG (8 × 128 GB) 3200 MHz
• GPU: Liquid-cooled NVIDIA H100 94 GB NVL
• Storage: 15.36 TB PCIe 4.0 NVMe SSD; 22 TB HDD
• Operating System / Software: Ubuntu 22.04 with BIZON OS and deep learning software stack, including TensorFlow, PyTorch, CUDA, cuDNN, SLURM, etc.
• Primary Use / Purpose: Large-scale AI/deep learning, data science, precision medicine modeling, GPU-accelerated scientific computing, and computationally intensive workflows requiring large memory capacity

BIZON ZX5500 Deep Learning / AI Workstation

• Quantity: 1
• CPU: AMD Ryzen Threadripper PRO 5995WX, 64 cores / 128 threads, 2.7 GHz base, up to 4.5 GHz boost; 256 MB L3 cache
• Memory: 1 TB DDR4-3200 ECC RDIMM, 8 × 128 GB Samsung modules
• GPU: 4 × NVIDIA RTX A6000 GPUs, 48 GB GDDR6 ECC VRAM each; 192 GB aggregate VRAM; 10,752 CUDA cores per GPU
• Storage: 7.68 TB Micron 7450 PCIe 4.0 NVMe SSD
• Operating System / Software: Ubuntu 22.04.3 LTS; Linux kernel 6.2; NVIDIA driver 575.51.03; CUDA 12.9. Suitable for PyTorch, TensorFlow, CUDA/cuDNN, SLURM, AlphaFold, Boltz-2, Rosetta, molecular dynamics, docking, and ML workflows
• Primary Use / Purpose: Large-scale GPU-accelerated molecular modeling, docking, protein design, deep learning, AI-model development, and high-throughput computational analysis

Warpcore (Mercury GPU)

• Quantity: 1
• CPU: AMD 2xEPCY 9684X, 96 cores 2.55 GHz
• Memory: 1536GB DDR5 4800MHz (24x64GB)
• GPU: 4xNVIDIA H100
• Storage: 2x960 GB NVMe; 30.72 TB NVMe
• Operating System / Software: Ubuntu server 26.04 deep learning software stack, including TensorFlow, PyTorch, CUDA, cuDNN, SLURM, etc.
• Primary Use / Purpose: Large-scale AI/deep learning, data science, precision medicine modeling, GPU-accelerated scientific computing, and computationally intensive workflows requiring large memory capacity

Silicon Mechanics Rackform R485.v6

• Quantity: 1
• CPU: Intel Xeon 8x Platinum 8160, 24-Core 2.1GHz
• Memory: 1536GB DDR4-2666 ECC (48x32GB)
• GPU: 8xNVIDIA Tesla P100, 16GB
• Storage: 4 × 2 TB Intel DC P4600 NVMe SSDs
• Operating System / Software: Ubuntu server 26.04 molecular modeling/simulation software stack
• Primary Use / Purpose: GPU-accelerated molecular dynamics simulations, molecular modeling, docking workflows.

Summary of Available Capacity

Approximate Total Across Listed Systems

• Total Systems: 4 workstations and 2 servers
• Total CPU Cores: 576 physical CPU cores
• Total System Memory: 5632 GB RAM
• Total GPUs: 21 GPUs
• GPU Memory: 128 GB from 4 × RTX 5090 GPUs, 470 GB from 5 × NVIDIA H100 NVL,192 GB from 4 × RTX A6000 GPUs, plus 128GB from NVIDIA P100
• High-speed NVMe Storage: 71.68 TB total
• Additional SSD/HDD Storage: 16 TB SATA SSD plus 22 TB HDD

Computing Environment

Many CPMDS computing resources, including the Orbital GPU workstations and WarpCore server, are housed in a dedicated server room within the UC Davis Health Administrative Support Building. The facility provides a secure, professionally managed environment with appropriate cooling, power, and network infrastructure to support reliable, high-performance computing operations.