When purchasing a new system or cluster, preparing the power requirements is integral. Power draw from the systems is key in addition to ensuring there is enough power available from the power source/circuit to fulfill the wattage needed.
...
For example, we will examine the power draw for the:
...
Quantum TXR430-
...
1500R (Deep Learning
...
Solution)
http://exxactcorp.com/index.php/solution/solu_detail/225198
Components that requires power listed below:
| Parts Listing ( |
| Quantum TXR430-1500R - Deep Learning Certified System) | ||
| Description | Quantity | Power Draw |
| @ Max | ||
| Quantum TXR430-1500R - 4U Intel Dual CPU Chassis, 24x 2.5IN Hot-Swap HDD Bays, Supports Up To 2x E5-2600 v3 CPU, Dedicated IPMI Port, 8x PCIE 3.0 x16 Slots, Up To 512GB DDR3 1866MHz ECC/REG Memory, 1600W Redundant Power Supplies | 1 | ~115W |
| Intel® Xeon® processor E5-2620 v4, 8C, 2.1 GHz 20M, 85W | 2 | 85W Per |
| 16GB DDR4 2400MHz ECC Registered 1.2V Memory Module (128GB system memory) | 8 | ~5-10W Per |
| 1TB 2.5 SATA III Internal Solid State Drive (SSD) (OS drive) | 1 | <1W |
NVIDIA Tesla M40 24GB GDDR5 PCIe 3.0 - Passive Cooling | 8 | 250W per |
| Total Power Requirement |
| ~1366W |
Cooling
To calculate the appropriate amount of cooling required for a single system or cluster, use the following:
...
P(BTU/hr) = 3.412141633 × 1325W 1366W
P(BTU/hr) = 45214660.0876 9855 (BTU/hr)