AlphaFold3 Workstation

We received an inquiry regarding a GPU machine that supports AlphaFold3 and other programs for research purposes, such as molecular structure prediction and analysis.
The software we plan to use is AlphaFold3, Amber, and R.
The hardware is equipped with a Xeon W7-3545, RTX PRO 5000, DDR5-5600 ECC R-DIMM 32GB x 8 memory configuration, and storage is equipped with two 480GB SSDs and one 15.36TB U.S.B. NVMe Gen5 SSD, with emphasis on stable operation in the Ubuntu 24.04 environment.
The customer would like a quote that includes a three-year warranty and rack mounting, and is looking for a configuration that can be expanded in the future to accommodate RAID configurations and additional SSDs.

CPU	Intel Xeon W7-3545 2.70GHz (up to 3.0GHz at TB4.8) 24C/48T
memory	Total 256GB DDR5 5600 REG ECC 32GB x 8
Storage 1	480GB SSD S-ATA x2
Storage 2	15.36TB SSD U.2 NVMe Gen5
Video	NVIDIA RTX PRO5000 48GB
network	on board (1GbE x1 /10GbE x1)
Housing + power supply	4U Rackmount Chassis for GPGPU + 1500W 80PLUS PLATINUM
OS	Ubuntu 24.04
Others	3-year send-back warranty CUDA Toolkit installation Rack mount slide rails SAS/SATA RAID card On-site installation (unpacking, racking, cabling, startup check)

CPU and memory configuration

As specified by the customer, it is equipped with a 24-core, 48-thread Xeon W7-3545.
The system has a total of 256GB of memory (32GB x 8), providing ample capacity for large-scale molecular dynamics calculations and machine learning processing.

GPU selection

This configuration is equipped with the desired RTX PRO 5000 48GB and is optimized for Alphafold3 inference processing and CUDA-based analysis.
We selected a motherboard with six PCIe x16 slots for future expansion, allowing for the addition of additional GPUs.
If you plan to add a GPU in the future, you will need to adjust the power supply and RAID card, so we recommend that you consult us at the planning stage.

Storage and RAID Support

In addition to a 480GB SATA SSD, it is equipped with a 15.36TB U.2 NVMe Gen5 SSD.
It supports analytical applications that require high-speed data access while also ensuring flexibility in RAID configuration.
It is equipped with a SAS/SATA compatible RAID card, allowing for future changes to the RAID configuration and the addition of SSDs.

Case, power supply and installation support

The chassis is compatible with both tower and 4U rack mounts, ensuring cooling and installation of GPUs and expansion cards.
Equipped with a 1500W 80PLUS PLATINUM power supply, it provides stable power supply.
We can also handle on-site installation (unpacking, rack installation, and startup check), and we also offer a 3-year send-back warranty.

Software environment

The CUDA Toolkit is pre-installed to ensure smooth setup of the AlphaFold 3 GPU environment.

In addition, upon request, we can provide molecular analysis and statistical analysis software such as Amber and R pre-installed.

For those who are active in these fields Molecular dynamics Statistical analysis Bioinformatics Medical science Biology

Tegara's custom-made PC production service not only caters to initial use, but also supports system expansion in anticipation of future expansion of research scale.
We not only propose configurations that meet various software requirements, but also accept consultations regarding the construction of an entire research environment.
Please feel free to contact us and we will provide the best solution to suit your needs.

Keyword ・What is AlphaFold? AlphaFold is an innovative protein structure prediction AI developed by DeepMind. It is capable of predicting three-dimensional structures with high accuracy from amino acid sequences, achieving overwhelming speed and cost-effectiveness compared to conventional experimental methods (X-ray crystallography and NMR). It is used in a wide range of fields, including structural biology, drug discovery, enzyme design, and disease research, and millions of structural information files are publicly available through the AlphaFold Protein Structure Database. Reference: AlphaFold – DeepMind ・What is Amber? Amber is a computational chemistry software that can perform molecular dynamics (MD) simulations with high accuracy and speed. It supports GPU acceleration, allowing efficient execution of long-term simulations of large-scale biomolecular systems. It is widely used for structural analysis and free energy calculations of proteins, nucleic acids, lipids, etc., and achieves dramatic speedups compared to conventional CPU-based calculations through parallel computing using CUDA-compatible GPUs. Reference: Amber GPU Performance Benchmarks – ambermd.org

■ Click here for details and inquiries about this PC case AlphaFold3 Workstation * Please enter the name of the case or your desired conditions.