product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-0

server

Home >  Products >  Others >  server

High-performance AI GPU Server Intel Xeon Scalable Processors 10 GPU Support PCIe 5.0 for AI Training & Cloud Computing

Product Description
Product Detail
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-2
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-3
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-4
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-5
1. Product Overview
The SuperCloud R8428 G13 is a high-performance GPU server designed based on Intel's fourth-generation/fifth-generation XeonScalable series processors. This product adopts a heterogeneous acceleration architecture, and its excellent product design brings it ultimate scalability. Its superior cooling performance enables it to support the cooling needs of up to 10 high-performance GPUs in a single machine. It features high performance, high scalability, and easy management and maintenance, making it highly suitable for various AI business applications such as AI cloud computing, AI training and inference, rendering,and video acceleration.
2. Product Features
2.1 Leading technical specifications
* Compatible with Hopper architecture GPU accelerator cards, supports MIG multi-instance, sparse matrix, NVlink Bridge connection and other leading technologies.
* Supports PCIe 5.0 connectivity, with an interconnect bandwidth of up to 64GB/s, and is backward compatible with PCIe 4.0.
2.2 Strong performance
* Equipped with two Intel® Xeon® Sapphire Rapids/Emerald Rapids series processors, with a maximum clock speed of up to 4.2 GHz, a single machine can support up to 128 computing cores, featuring 4 UPI interconnects, and a TDP (Thermal Design Power) of up to 350W.
* Supports 32 DDR5 ECC memory slots, 16 memory channels, and supports RDIMM and LRDIMM up to 5600MT/s,.
* It supports up to 24 hot-swappable 2.5-inch SAS/SATA hard drives or 12 3.5-inch SAS/SATA hard drives, and optionally supports 16 hot-swappable NVMe SSDs and 1 M.2 SATA/NVMe SSD. It achieves all-flash connectivity within the entire machine, greatly enhancing data exchange speed.
2.3 Intelligent management system
* By embedding an intelligent management chip in the server, various management functions such as complete IPMI2.0 remote system monitoring, remote KVM, virtual media, and more can be achieved. When paired with the SuperCloud cloud platform management suite,it greatly simplifies user device deployment, enables one-click virtualization applications, facilitates the deployment of virtual machines, and achieves dynamic adjustment of physical resources in a strategic manner. The visual management module enables remote fault diagnosis and alarming of servers, significantly enhancing the efficiency of data center operation and maintenance and reducing the workload of management personnel.
2.4 Excellent and flexible scalability
* The front supports 3 full-height half-length PCI-E 5.0 interfaces and 1 OCP 3.0 expansion card, while the rear supports 13 PCI-E 5.0 slots and 1 OCP 3.0 expansion card, with a maximum support of 10 dual-wide PCIE GPU cards.
* Supporting the internal fan wall of the machine greatly enhances heat dissipation efficiency.
* Support 400G Remote Direct Memory Access (RDMA) high-speed network, optimizing cluster bandwidth.
3. Product technical specifications

3.1 Host performance
Processor
Supports 2 Intel® Xeon® Sapphire Rapids/Emerald Rapid series processors, with TDP (Thermal Design Power) ≤ 350W
Chipset
Intel® C741 series chipset
Memory type
Supports DDR5 ECC memory, with a memory voltage of 1.1V
(The operating frequency varies depending on the CPU and memory configuration)
Slot and capacity
32 memory slots, with a maximum expandable memory capacity of 12TB
Supports DDR5 5600/4800 ECC LRDIMM; RDIMM;
Supports advanced features such as advanced memory error correction, memory mirroring, and memory hot standby
Hard disk
It supports up to 24 standard 2.5-inch/12 standard 3.5-inch hot-swappable hard drives, and supports SAS/SATA/NVME interface types
Supports up to 16 U.2 NVMe hard Drives
Supports one M.2 NVMe hard drive
Memory Controller
Optional SAS controller (12G), supports RAID 0, 1, 10
The optional RAID controller (12G) supports a maximum of 2G of cache and supports RAID 0/1/10/5/50/6/60
It can also be extended with a cache data protection suite and a cache protection acceleration function
Network
1 IPMI remote management network interface (RJ45 interface)
1 OCP network card expansion
CD-ROM
Ultra-thin DVD-RW drive with optional external USB interface
3.2 Scalability
I/O expansion
The rear supports up to 13 standard PCIe 5.0 slots
The front supports up to 3 standard PCIe 5.0 slots
GPU
Supports up to 10 dual-wide GPU cards
Peripheral Interface
1×VGA interface
2×USB3.0 interfaces
1xRJ45 management interface
1× Serial interface
3.3 Chassis power supply
Chassis
SuperCloud 4U GPU server chassis
175mm(H) x 447mm (W) x 900mm(D)
Fan
12 intelligent control fans, supporting N+1 redundancy
Power supply
Standard configuration includes a 2700W titanium (2+2/3+1) redundant power supply with a conversion rate of 96%
 3.4 Accessories included
Manual
Hyper Cloud Server User Manual
Rail
A set of standard rack-mount rails
Power line
National standard AC power cord
Recommend Products
Company Profile
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-6
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-7
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-8
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-9
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-10
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-11
FAQ
Q: Are you a manufacturer or a trading company? A: We have been professional antenna manufacturer for 18 years. Q: Have gps gsm combo antenna 100% assembles well in stock? A: No, all gps gsm combo antenna will be newly assembled according to your orders including the samples. Q: Could I use my own LOGO or design on goods? A: Yes, customized logo and design on mass production are available. Q: What's the payment term? A: T/T, AliPay,Paypal,Afterpay,Klarna etc. Q:What certifications do you have? A:Certified by FCC, CE, RoHS,ISO available upon request. Q: Where is your factory located? Can I visit your factory? A: Factory address: 2nd Floor, Building D4-1, Hongfeng Science and Technology Park, Nanjing Economic and Technological Development Zone, Jiangsu Province. Just call us, we will pick you up with no delay.
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-12
product high performance ai gpu server  intel xeon scalable processors  10 gpu support  pcie 50  for ai training  cloud computing203-13

Get a Free Quote

Our representative will contact you soon.
Email
Name
Company Name
Message
0/1000