The Cisco UCS C480 ML, a purpose-built Server for Deep Learning, is a four-rack-unit (4RU) server supporting Intel® Xeon® processor scalable family CPUs with 8 NVIDIA Tesla V100-32GB Tensor Core GPUs with NVLink Interconnect. It supports up to 3 terabytes (TB) of double-data-rate 4 (DDR4) memory1 in 24 slots, up to 24 small form factor (SFF) hot-swappable2 SAS/SATA SSD/HDD, up to 6 PCIe NVMe disk drives and upto two internal M.2 drives.4 PCI Express (PCIe) expansion slots support Cisco UCS C-Series and partner network adapters, with additional I/O provided by two 10Gbase-T LOM ports and one 1GbE dedicated out-of-band (OOB) management port. A separate PCIe slot is reserved inside the chassis for a RAID controller card