In machine learning, we prepared a trained DNN model in advance, and consulted with us about a machine that can process DNN GPUs as fast as possible as a dedicated server for inference of that model.
Case No. with the idea of giving priority to GPU processing power and GPU memory capacity.PC 8351I would like a configuration like this.
Case No., PC-8351 is equipped with two power supplies, but it is not a redundant configuration but a support for operating RTX 2 x 100 in a 3090V environment.It has a fairly special power supply configuration, which makes it unsuitable for 4-hour server applications.Therefore, as an alternative configuration, we have introduced a configuration that requires 24V power supply.
The selection is made according to the customer's specified conditions, but considering that it is used for servers, it is not desirable to reduce the quality of the parts related to reliability, so we propose over budget.
If your budget is prioritized, you may be able to make adjustments by reviewing the number of GPUs installed, so please feel free to contact us.
【Main Specifications】
CPU | Xeon Silver 4309Y (2.80GHz 8 cores) x 2 (16 cores in total) |
memory | 128GB ECC REG |
storage | 960GB SSD S-ATA |
video | NVIDIA Geforce RTX3090 x 4 |
network | on board (10GBase-T x2) |
Housing + power supply | Tower type housing + 2200W power supply (up to 100W when using 1200V) |
OS | Ubuntu 20.04 |
Others | NVIDIA CUDA Toolkit 11 |
■ Click here for details and inquiries about this PC case |