We were consulted about introducing a high-performance workstation for robot machine learning.
The desired specifications are Ubuntu 22.04 OS, Intel Core Ultra 9 CPU, 128GB memory, 4TB SSD + 2TB HDD storage, and four RTX6000 Ada GPUs.
| CPU | Intel Xeon W5-3535X 2.90GHz (up to 3.0GHz at TB4.8) 20C/40T |
| memory | Total 128GB DDR5 5600 REG ECC 16GB x 8 |
| Storage 1 | 4TB SSD S-ATA |
| Storage 2 | 2TB HDD S-ATA |
| Video | NVIDIA RTX6000 Ada 48GB (DisplayPort x4) x 4 |
| network | on board (1GbE x1 /10GbE x1) |
| Housing + power supply | Tower type chassis + 2800W TITANIUM |
| OS | Ubuntu 22.04 |
| Others | 15A 200V Power Cable C19 – C14 |
About CPU
The Core Ultra series you requested is designed for consumers and does not support more than two GPUs due to the PCI Express lane count specifications.
Alternatively, we can use a multi-plate configuration.Xeon-basedplatform was adopted.
The increased number of PCIe lanes ensures sufficient bandwidth even with a four-GPU configuration.Stable performanceYou can maintain.
GPU and power supply environment
In this configuration with four RTX6000 Ada cards,The power consumption is very high
(Including CPU and other devices,It may exceed 2000W.)
To ensure a stable power supply, a power supply unit with a maximum output of 2800W is used, and a 200V power cable is also included.
*The installed power supply is a 200V dedicated module.
Cannot be used with a household 100V outlet.
Please check in advance whether the equipment you plan to install is compatible with 200V.
Memory and Storage
It is equipped with 128GB of memory, a 4TB SSD, and a 2TB HDD according to customer requests.
Please note that the main memory capacity isIt is recommended to have a sufficient margin beyond the total VRAM capacity of the installed GPU.Will be
If the memory capacity is less than the VRAM capacity, a bottleneck may occur when loading training data onto the GPU, or processing errors may occur.
In this configuration, RTX 6000 Ada × 4 (total VRAM 192GB),256GB or more of memory recommendedCapacity.
After explaining these points, the initial configuration will be 128GB of memory.Configuration assuming future expansionWe are providing the following information.
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 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.
|
■ Click here for details and inquiries about this PC case
* Please enter the name of the case or your desired conditions. |
