AI GPU workstations and servers
SR600 6U 8-GPU AI training server (, AMD EPYC)
型号 SR600
照片即将上线
概览
特点
- Packing up to 8 high-performance GPUs into a 6U footprint
- powered by the AMD® EPYC™ processor platform
用途
- deep learning, model training, inference, high-performance computing (HPC), and data analytics; Training Clusters, Data Centers, Supercomputing Centers (HPC)
规格
- Form Factor: 6U供应商自述
- CPU: Dual-Socket AMD EPYC™ 9005 / 9004 Series Processors (Socket SP5); Up to 192 cores per CPU (384 cores total in 2P configuration); Up to 500W TDP per CPU供应商自述
- Memory: 24× DDR5 DIMM Slots; Up to 6400 MT/s (with AMD EPYC™ 9005 Series)供应商自述
- Storage: Drive Types: SAS / SATA / NVMe (AnyBay): 16× 2.5-inch Hot-Swap SAS / SATA Drive Bays (up to 4× NVMe SSDs supported), 2× Onboard M.2 NVMe Slots; or 8× 2.5-inch Hot-Swap SAS / SATA Drive Bays (up to 4× NVMe SSDs supported), 2× Onboard M.2 NVMe Slots供应商自述
- RAID: Optional RAID levels: 0, 1, 10, 5, 50, 6, 60; Cache protection via Supercapacitor Backup Unit (SCP)供应商自述
- GPU Acceleration: Up to 10× Full-Height Full-Length (FHFL) Double-Width Air-Cooled GPUs; Up to 700W TDP per GPU供应商自述
- Networking: 1× Onboard 1GbE Dedicated HDM Management Port; 1× OCP 3.0 Mezzanine Slot (for NIC modules)供应商自述
- I/O: Front Panel I/O：2× USB 3.0 Type-A Ports;1× 1GbE Dedicated Management Port (HDM/BMC);1× Management Serial Port (RJ-45 or DB-9); Mounting Ear I/O (Bezel)：1× USB 2.0+1× USB 3.0+1× VGA+1× TypeC Debug供应商自述
- Power Supply & Thermal: Up to 10× Hot-Swappable Redundant Power Supply Modules (Optional Wattages: 2000W / 2700W / 3200W), Redundancy: N+N供应商自述
- Operating Environment: 5℃ - 35℃供应商自述
- 型号
- SR600
- 最低起订量
- 随报价确认
- 销售单位
- unit
这些信息是供应商自述，我们尚未测试。
运输与订购
付款时选择海运或空运。运费、进口关税和税费之后按代理实际成本收取，不计入商品价格。
付款方式
- 现在：支付货款。
- 之后：运费、进口关税和进口税费，按我们货代的实际成本收取。货物到达货代后我们开具账单，须在 7 天内付款，付款后我们安排派送。
- 运费和关税的区间只是预估，并非上限：您按实际成本支付，即使超出该区间。
- 进口记录人是我们的货代，或是通过货代办理进口的您，具体以您的运费报价为准。进口关税和税费按成本转收。
- 我们的价格不另加销售税。
还没有价格？无需登录，告诉我们您会买多少。
目前在 英国 提出需求的企业少于 5 家。达到 5 家后我们会显示总数。
我会买这个