ZC4530
4U dual-socket AI accelerated computing server
ZC4530 Overview
Supports 1/2 3rd Gen Intel® Xeon® Scalable series processors, up to 205W TDP. Supports 10 high-speed PCIe expansion slots for rich scalability and 5 full-height dual-width high-performance GPU cards to meet high computing demands. Supports 16 DDR4 memory slots with frequencies of 2666/2933/3200MHz. High performance: up to 2 3rd Gen Intel® Xeon® Scalable series processors deliver powerful computing; supports Intel Turbo Boost Technology 2.0, optimized for deep learning with Intel® SSE4.2, Intel® AVX, Intel® AVX2 and Intel® AVX-512 instruction sets.
High Reliability
Advanced memory fault tolerance; optional SAS RAID card supporting RAID 0/1/5/6/50/60, with cache super-capacitor protection, RAID state migration and RAID configuration retention.
High Scalability
16 memory DIMM slots, up to 2TB 3DS ECC LRDIMM/RDIMM; supports 12 x 2.5"/3.5" SAS/SATA drives or SSDs; supports up to 4 full-length full-height dual-width professional graphics and compute accelerator cards.
High Manageability
Supports smart dynamic fan speed control that adjusts fan speed by load, effectively reducing system noise and power consumption; integrated IPMI enables remote server operation and maintenance with fault indication, effectively reducing O&M costs.
| System Model | ZC4530 |
|---|---|
| Processor | Supports 1 or 2 3rd Gen Intel® Xeon® Scalable ICE Lake series processors, CPU TDP up to 250W |
| Chipset | Intel® C621A series server-grade chipset |
| GPU Card | Supports 5 RTX 3090, 4 A800 or 4 A16 dual-width full-height full-length cards |
| Memory | 16 DDR4 memory slots, supports DDR4 2666/2933/3200MHz LRDIMM/RDIMM memory |
| Storage Controller | PCH supports RAID 0/1/5 (for Windows only) |
| Storage | Supports up to 12 x 2.5"/3.5" SAS/SATA hot-swap drives; optional rear 2 x 2.5" SAS/SATA or U.2 drives; Internal storage: 1 x PCIe 3.0 x4 M.2 slot, 3 x MiniSAS HD ports, 2 x SATA DOM ports and 1 x COM port |
| Network | Onboard 2 x gigabit RJ45 service ports |
| I/O Ports | Front: 2 x USB3.0; Rear: integrated 2 x gigabit service ports, dedicated 1 x RJ45 management port, 2 x USB3.0, 1 x VGA, 1 x COM |
| PCIe Expansion | Rear supports 10 full-height PCIe x16 slots (supports 6 x PCIe 3.0 x8 in PCIe 3.0 x16 slot, 1 x PCIe 3.0 x16 and 3 x PCIe 4.0 x16); optional 4 x PCIe 3.0 x8 (in PCIe x16 slot), 2 x PCIe 3.0 x16 and 3 x PCIe 4.0 x16 |
| GPU Quantity | 5 full-length half-height single-width GPU cards |
| Power Supply | 1600W/2000W/2200W hot-swap 1+1 redundant power supplies |
| Fans | Standard 8 x 8038 hot-swap N+1 redundant fans |
| Management | IPMI 2.0 |
| Security | TPM/TCM (optional) |
| Operating System | Microsoft Windows Server, Red Hat Enterprise Linux, SUSE Linux Enterprise Server, CentOS, Ubuntu |
| Server Dimensions | 695 x 433 x 176.5mm (D x W x H) |
| Operating Temperature | Operating temperature: 5°C - 35°C (no direct sunlight); Storage temperature: -40°C - 70°C |
Related Products
ZC2230
General-purpose server. A dual-socket general-purpose computing server based on the 3rd Generation Intel® Xeon® Scalable processors, delivering outstanding computing performance with versatile I/O expansion. It also provides ample storage capacity and is ideal for virtualization, data analytics, public and private clouds, as well as scenarios such as VDI, AI vision, big data processing and enterprise applications.
ZC7850
ZC7850 is an outstanding 7U dual-socket 8-GPU AI server with distinct advantages. Built on the 4th/5th Gen Intel® Xeon® Scalable processors and paired with high-performance GPUs, it delivers comprehensive upgrades across CPU, GPU and I/O specs, featuring high computing power, strong scalability, rich configuration and high reliability. It is ideal for AI inference, large...
ZC8850
Inference and training in one. With direct CPU-GPU interconnect, it supports 8 x 600W dual-width, triple-width, 3.5-slot and 4.5-slot GPUs. It features high performance and parallel computing, high scalability, high reliability, high energy efficiency, smart maintenance, modular design and an open ecosystem, meeting the needs of large-model training, AI inference, graphics processing and generative AI. Ideal for large models...
ZC6850
Inference and training in one. With direct CPU-GPU interconnect, it supports 600W dual-width, triple-width and 3.5-slot GPUs. It features high performance and parallel computing, high scalability, high reliability, high energy efficiency, smart maintenance, modular design and an open ecosystem, meeting the needs of various AI workloads. Ideal for cloud gaming, AI...