Machine for training phase of deep learning

A customer planning deep learning for image systems asked us about a GPU machine for use in the training phase (the phase in which the Deep Neural Network automatically adjusts the weights and biases of the network).Your requirements are as follows.

Budget: 300 million yen or less
GPU: NVIDIA RTX A5000 x 4
Storage: about 2TB
OS: Ubuntu 18.x or 20.x
Power supply environment: Power supply environment for courses (100V)
Others: Hope for a quiet configuration

The maximum power consumption of the NVIDIA RTX A5000 is 1W per unit, which is relatively low for a high-end product, so even if four units are installed, it can be used with a household power supply.This configuration also assumes power supply with 230V, one system.

The chassis is also selected with an emphasis on quietness, but space for 2 slots is required for the convenience of mounting 4 cards that occupy 8 expansion slots.Since space is reserved for only 7 slots of the ATX form factor in a general middle tower housing, it is inevitable to select a large housing compatible with E-ATX etc.

【Main Specifications】