| Sign In | Join Free | My himfr.com |
|
Shenzhen Zhizhenyuan Technology Co., Ltd.
Verified Supplier
1 Years
| Sign In | Join Free | My himfr.com |
|
1 Years
| Ask Lasest Price | |
| Brand Name : | Lenovo |
| Model Number : | SR650 |
| Price : | 8888USD/unit |
| Payment Terms : | T/T |
| Supply Ability : | 1000 |
| Delivery Time : | Depends on the order amount |
Products Highlight
From machine vision and recommendation systems to natural language processing and more complex large AI models, different AI application scenarios have varying requirements for AI servers. The Lenovo Wentian WA5480 G3 AI Training Server, a 4U rack-mountable server within the Lenovo Wentian brand, supports versatile computing power and a rich ecosystem. It delivers extremely reliable computing power for AI model training and inference, accelerates AI deployment across industries, and drives the intelligent transformation of various sectors and society as a whole.
Versatile Computing Power
Powered by two 4th or 5th Gen Intel® Xeon® Scalable processors and supporting the latest PCIe 5.0, the Lenovo Wentian WA5480 G3 AI Training Server can accommodate up to 10 multi-type, multi-brand AI accelerator cards via PCIe expansion. It is flexibly applicable to various scenarios including general AI model training, large model inference, AI generation, cloud gaming, and computational science, providing diverse computing power for multifaceted AI demands.
Flexible Topology
The hardware design of the Lenovo Wentian WA5480 G3 emphasizes flexibility in CPU-GPU interconnection. Based on different AI workloads, it offers customers multiple CPU-GPU connection modes, including Direct Connect, Balance, and Common. This avoids potential performance bottlenecks and reduced system efficiency caused by a single communication method mismatched to the workload. Combined with the choice of different types and quantities of accelerator cards, it achieves perfect alignment with diverse, complex AI scenarios.
Secure & Reliable Design
The Lenovo Wentian WA5480 G3 AI Training Server incorporates extensive design redundancy, providing a secure and stable operating environment for AI computing. Key features include:
N+N Redundant Power Supplies: Mutual backup within the node, connectable to dual utility power sources to ensure 24/7 online operation.
Critical Component Monitoring: Functional monitoring of CPU, GPU, memory, etc.
System Power Capping: Dynamic regulation to maintain operation within a reasonable power range, ensuring stable performance.
N+I Redundant Fans: Among other safety designs ensuring high-standard uninterrupted operation.
Products Datasheet
| Form Factor | 4U Rack |
| Processor | Up to 2x 4th/5th Gen Intel® Xeon® Scalable (Sapphire Rapids, Emerald Rapids), up to 64 cores, 350W TDP |
| Memory | 32x DDR5 DIMMs (8 channels per CPU), 1DPC: 5600 MHz max, 2DPC: 4800 MHz max |
| GPU Support | 8x dual-wide or 4x triple-wide GPUs (Max: 10x dual-wide) |
| Storage | Front: |
| - 24x 2.5" SAS/SATA/NVMe (12x NVMe max: 8 mobo + 4 PCIe switch) | |
| or 12x 3.5" SAS/SATA/NVMe (4x NVMe max from mobo) | |
| Internal: | |
| - 2x M.2 SATA/NVMe boot (RAID supported) | |
| - Direct NVMe, VROC/VMD & onboard SATA with SW RAID (0,1,5,10) | |
| I/O Expansion | SKU 1 (8GPU): |
| - 8x PCIe Gen5 x16 FHFL (switch) | |
| - 2x PCIe Gen5 x16 FHFL (switch) | |
| - 3x PCIe Gen5 x16 FHFL (CPU) | |
| - 2x PCIe Gen5 x8 FHHL (CPU, RAID) | |
| - 1x OCP 3.0 Gen5 x16 (CPU) | |
| SKU 2 (8GPU Direct): | |
| - 8x PCIe Gen5 x16 FHFL (CPU) | |
| - 2x PCIe Gen5 x16 FHFL (CPU) or 1x PCIe + OCP 3.0 | |
| SKU 3 (4GPU Direct): | |
| - 4x PCIe Gen5 x16 FHFL (CPU) | |
| - 2x PCIe Gen5 x16 FHFL (CPU) | |
| - 1x OCP 3.0 Gen5 x16 (CPU) | |
| Integrated Network | OCP 3.0 card supported |
| Ports | Front: |
| - 1x USB 3.0, 1x USB 2.0 (BMC), 1x VGA (optional) | |
| Rear: | |
| - 2x USB 3.0, 1x VGA, 1x 1GbE RJ45 (Mgmt), 1x Serial (optional) | |
| HBA/RAID Support | SATA with SW RAID (0,1,5,10); Standard/Internal PCIe RAID/HBA adapters |
| Power & Cooling | 4x Hot-swap redundant CRPS PSUs (Platinum/Titanium): 1600W, 2200W, 2700W, 3200W |
| Management | Lenovo BMC with management suite; Dual BIOS/BMC flash with mirroring |
| OS Support | Microsoft, SUSE, Red Hat, VMware, Ubuntu, etc. |
| Dimensions (HxWxD) | 174.5mm x 447mm x 842.8mm (excl. bezel) |
| Operating Temp | 10°C ~ 35°C |
CPU Comparison
| Processor Model | Cores / Threads | Base Clock | Max Turbo | L3 Cache | TDP | Memory Support | Max Memory | Memory Channels |
| Xeon Platinum 8592+ | 64 / 128 | 1.9 GHz | 3.9 GHz | 320 MB | 350W | DDR5-5600 | 8 TB | 8 |
| Xeon Platinum 8580 | 60 / 120 | 2.0 GHz | 4.0 GHz | 300 MB | 350W | DDR5-5600 | 8 TB | 8 |
| Xeon Gold 6548Y | 32 / 64 | 2.5 GHz | 4.1 GHz | 60 MB | 300W | DDR5-5600 | 4 TB | 8 |
| Xeon Gold 6526Y | 16 / 32 | 2.8 GHz | 4.1 GHz | 37.5 MB | 300W | DDR5-5600 | 4 TB | 8 |
| Xeon Silver 4514Y | 12月24日 | 2.0 GHz | 3.8 GHz | 30 MB | 165W | DDR5-4800 | 4 TB | 8 |
Complex Customization (Silk Screen)
| Phase | Stage | Key Activities & Deliverables |
|---|---|---|
| 1 | Inquiry | • Client submits detailed customization requirements |
| 2 | Requirements Confirmation | • Supplier provides feasibility previews • Client approves technical specifications |
| 3 | Feasibility Analysis | • Technical validation of design/production • Risk assessment report |
| 4 | Quotation | • Detailed cost breakdown • Production timeline with milestones |
| 5 | Place Order | • Contract execution with payment terms • Final requirement verification (double-check) |
| 6 | Mass Production | • Bulk manufacturing launch • Quality control during production |
| 7 | Pre-Delivery Audit | • Export documentation (HS code, invoice) • Declared value verification • Final QC inspection |
| 8 | Delivery & Support | • Logistics coordination • Real-time shipment tracking • Post-delivery support |
|
|
