Cryo-electron microscope image analysis workstation

Our customer inquired about a workstation for analysis and processing that reconstructs three-dimensional structures from a large number of images acquired with a cryo-electron microscope.
With an eye toward 3D reconstruction using large amounts of image data, we have received requests to operate multiple analysis software programs, such as RELION and CryoSPARC, in a Linux environment.
Furthermore, they place importance on GPU performance in order to improve research efficiency, and would like a quote with a budget of around 1,000 million yen.

Based on the consultation, we proposed the following structure:

CPU	Intel Xeon W7-3545 2.70GHz (up to 3.0GHz at TB4.8) 24C/48T
memory	Total 512GB DDR5 5600 REG ECC 64GB x 8
Storage 1	1TB SSD M.2 NVMe Gen4
Storage 2	4TB SSD M.2 NVMe Gen4
Video	NVIDIA RTX PRO 6000 Max-Q 96GB x 4
network	on board (1GbE x1 /10GbE x1)
Housing + power supply	GPGPU Full Tower 2800W Titanium
OS	Ubuntu 24.04

GPU configuration

For analysis processing, not only the number of GPUs is important, but also the VRAM capacity and stability over long periods of operation. Therefore, we proposed a configuration that would provide the best balance of performance for a workstation, equipping it with four NVIDIA RTX PRO 6000 Max-Q (96GB) cards.
This GPU is equipped with a large capacity VRAM, allowing it to process large-scale data analyses with ease. It also has a sufficient total memory bandwidth, allowing it to deliver stable processing performance in various analysis software.

CPU and memory

To handle large amounts of metadata and intermediate files in parallel with GPU processing, the Xeon W series was adopted and equipped with 512GB of memory.
By ensuring memory capacity that exceeds the total GPU VRAM, processing is less likely to stall even with large-scale tasks on RELION, CryoSPARC, etc., and stability can be maintained even during long-term analyses.

Storage and OS configuration

Because multiple analysis software will be used, Ubuntu 24.04 was selected as the standard OS due to its robustness against dependencies and version differences.
The storage is configured with two 1TB + 4TB NVMe SSDs, prioritizing high-speed access, ensuring a balance that can handle the load of reading and writing analysis data.

Chassis and power supply design

To ensure stable operation of a four-GPU configuration, a full-tower GPGPU chassis with excellent heat dissipation performance is used.

The power supply is a high-capacity 2800W product, designed for continuous operation and resistance to load fluctuations.
This configuration allows for safe operation even in research applications where high loads continue.

Keyword ・What is RELION? RELION (REgularized LIkelihood Optimization) is software widely used for 3D reconstruction using cryo-electron microscopy (cryo-EM) data. It features highly accurate particle classification, 3D reconstruction, and structure refinement based on Bayesian inference, and has become a standard analysis tool in structural biology and protein complex analysis. It supports large-scale GPU acceleration, enabling high-speed, high-resolution structural analysis. Reference: RELION 4.0 Installation Guide (relion.readthedocs.io) ・What is cryoSPARC? cryoSPARC is an analysis platform that integrates high-speed processing and high-resolution refinement of cryo-electron microscopy (cryo-EM) data. It offers a variety of workflows, including particle picking, 2D classification, ab initio reconstruction, and uniform and non-uniform refinement, and its strength lies in its high-speed processing through GPU optimization. It is widely used for large-scale structural analysis, complex dynamics research, and high-throughput analysis. Reference: cryoSPARC Hardware & System Requirements (guide.cryosparc.com)

 

■ Click here for details and inquiries about this PC case Cryo-electron microscope image analysis workstation * Please enter the name of the case or your desired conditions.