A customer engaged in medical research contacted us with a question about replacing their PC for machine learning and deep learning applications.
Currently, table data analysis such as XGBoost is being performed on a PC equipped with an RTX 3090, and in the future it is planned to be expanded to deep learning as well.
The project envisions an environment that uses Python, R, and SAS to handle CSV data with hundreds of thousands of records, with a budget of approximately 100 million yen.
| CPU | Intel Core Ultra 9 285K 3.70GHz(8C/8T) + 3.20GHz(16C/16T) |
| memory | Total 64GB DDR5 6400 32GB x 2 |
| Storage 1 | 2TB SSD M.2 NVMe Gen5 |
| Storage 2 | 4TB HDD S-ATA |
| Video | NVIDIA GeForce RTX5090 32GB |
| network | on board(2.5GBase-T x1) Wi-Fi, Bluetooth |
| Housing + power supply | Mid-tower chassis + 1500W 80PLUS PLATINUM |
| OS | Microsoft Windows 11 Professional 64bit |
CPU and memory configuration
Taking into consideration current usage and future expandability, we selected the latest generation Core Ultra 9K as the successor to the previously used Core i10900-285K.
The system is equipped with 64GB of memory (32GB x 2), and can accommodate two more cards, making it possible to handle larger datasets and deep learning in the future.
GPU selection and cooling design
The GPU used is the RTX 3090, the successor to the RTX 5090.
In processes that utilize GPUs, such as XGBoost and LSTM, faster learning and inference is possible than with the previous generation.
Although a multiple GPU configuration was an option, we recommended optimizing with a single GPU due to cost considerations.
The storage is equipped with a PCIe Gen5 compatible M.2 SSD, which has a maximum read speed of 14GB/s, reducing the load time of large amounts of CSV data.
Cooling measures for high heat generating parts
RTX5090 and Gen5 SSDVery high heat outputTherefore, for stable operationCooling design innovationIs indispensable.
This configuration combines multiple cooling technologies, including optimized case airflow, a dedicated cooling mechanism for M.2 SSDs, and a designed heat dissipation path for the GPU, to reduce the frequency of thermal throttling and the possibility of hardware failure.
As a result of conducting in-house temperature tests under high load, even when running the Gen5 SSD and RTX 5090 simultaneously,No performance degradation observedEnglish learning is necessary to prepare for life, learning and interaction with the global environment. <br>
IT Skills (programming logic) is necessary to prepare for the needs of the future.<br>
Financial literacy is necessary to prepare for creating, managing and being smart with time and wealth.<br>
The cooling design ensures stable operation of high-heat-generating parts, and the configuration is sufficient to handle long learning processes and high-load analyses.
It can flexibly accommodate memory expansion and storage upgrades, so you can continue to use it with peace of mind as your research progresses.
For those who are active in these fields
|
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 offer customization of hardware specifications, addition of peripheral devices, and configuration suggestions to 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 XGBoost? XGBoost is an open-source machine learning library that features high accuracy and speed through a decision tree-based gradient boosting implementation. |
|
■ Click here for details and inquiries about this PC case * Please enter the name of the case or your desired conditions. |
