Supermicro 4028GR-TR2 GPU Server, 4× NVIDIA Tesla V100 32GB PCIe Complete Solution
  • Product categories:Servers
  • Part number:X9DRG-O / X10DRG-O
  • Availability:In Stock
  • Condition:Used
  • Product features:Customizable
  • Min order:1 unit
  • List Price was:$5,955.00
  • Your Price: $5,588.00 You save $367.00
  • Chat Now Send Email

Breathe easy. Returns accepted.

Shipping: International shipment of items may be subject to customs processing and additional charges. See details

Delivery: Please allow additional time if international delivery is subject to customs processing. See details

Returns: 14 days returns.Seller pays for return shipping. See details

FREE Shipping. We are accepting NET 30 Days Purchase Orders. Get a decision in seconds, without affecting your credit.

If you need a large quantity of X9DRG-O / X10DRG-O product - call us through our toll-free number at Whatsapp: (+86) 151-0113-5020 or request a quote at live chat and our sales manager will contact you shortly.

Supermicro 4028GR-TR2 GPU Server, 4× NVIDIA Tesla V100 32GB PCIe Complete Solution

Keywords

Supermicro 4028GR-TR2, 4028GR-TR2, Tesla V100 32GB PCIe, NVIDIA Tesla V100, GPU server, 4U rack server, X10DRG-O+-CPU, E5-2600 v3 v4, AI training server, deep learning workstation, HPC server, PCIe 3.0 GPU node, 2000W titanium PSU, 24×2.5" bay, single root complex, GPUDirect RDMA, C612 chipset, LGA2011 R3, DDR4 ECC, supermicro GPU chassis, used gpu server, inference server, scientific computing, virtualization, high density gpu server, mining server, data center server, enterprise server, supermicro superserver, Tesla V100 PCIe, 4 gpu server, 2000w redundant power

Description

Complete Supermicro 4028GR-TR2 4U rack GPU server pre-configured with 4 pieces NVIDIA Tesla V100 32GB PCIe cards. Dual LGA2011 R3 motherboard X10DRG-O+-CPU, support Intel Xeon E5-2600 v3/v4 CPUs, up to 24 DDR4 ECC memory slots, 24× 2.5-inch SAS/SATA hot-swap drive bays, 2000W titanium redundant power supply, single-root PCIe complex design, ideal for AI training, deep learning, HPC, scientific simulation and GPU acceleration workloads.

Key Features

Built on Supermicro SYS-4028GR-TR2 4U rack chassis, equipped with X10DRG-O+-CPU mainboard and dedicated PCIe expansion backplane, stably runs 4× Tesla V100 32GB PCIe full-height double-width GPUs under single root complex, supports GPUDirect RDMA for low latency GPU communication.

Dual socket R3 LGA2011 platform compatible with Intel Xeon E5-2600 v3 / v4 series processors, 24 DIMM slots supporting large-capacity DDR4 ECC RDIMM/LRDIMM, up to 3TB total memory for large model loading and parallel computing tasks.

Comes with 24 × 2.5-inch hot-swap drive bays via SAS3 12Gbps backplane, 2000W titanium redundant power supply, high airflow fan array, IPMI remote management, suitable for data center deployment, AI inference, computing cluster construction and secondary rental business.

Configuration

  • Model: Supermicro SYS-4028GR-TR2 (4U Rackmount)
  • Motherboard: X10DRG-O+-CPU + X10DRG-O-PCIE PCIe daughterboard
  • CPU: Dual Intel Xeon E5-2600 v3 / v4 series (LGA2011 R3, max 160W TDP)
  • GPU: 4 × NVIDIA Tesla V100 32GB PCIe (double width)
  • Memory: 24 × DDR4 ECC RDIMM / LRDIMM, max up to 3TB
  • Storage: 24 × 2.5" SAS / SATA hot-swap bays, SAS3 12Gbps backplane BPN-SAS3-216A
  • Expansion: Multiple PCIe 3.0 x16 full-height full-length slots, single root complex
  • Power: 2000W Titanium redundant power supply
  • Network: Onboard Gigabit Ethernet, IPMI for remote management
  • Chipset: Intel C612 PCH

Compatibility

Supports Intel Xeon E5-2600 v3 / v4 processors; DDR4 ECC RDIMM, LRDIMM, 3DS LRDIMM; PCIe 3.0 double-width GPU cards including Tesla V100, P100, P40, P102 etc.; OS: Ubuntu, CentOS, RHEL, Windows Server, compatible with CUDA, cuDNN, TensorFlow, PyTorch, MXNet, MPI and HPC stacks.

Usage Scenarios

Deep learning model training, AI inference service, high-performance computing, molecular simulation, computational fluid dynamics, video rendering, GPU computing cluster, university research lab, data center acceleration node, private AI computing platform.

FAQ

Q: Is this system compatible with NVLink V100?
A: This listing is for PCIe version Tesla V100, this 4028GR-TR2 model does not support NVLink interconnect; choose 4028GR-TVRT for SXM2 NVLink V100 build.

Q: What is the maximum GPU quantity this chassis supports?
A: The 4028GR-TR2 backplane can support up to 8 double-width PCIe GPUs, this set is configured with 4× V100 32GB PCIe as complete ready-to-run solution.

Q: Does it include CPU, RAM and hard drives?
A: The base listing highlights the complete GPU solution with 4× V100; CPU, memory and storage can be customized according to customer requirements, we support full configuration matching before shipment.

Q: Can it be used for stable 24/7 running?
A: Original Supermicro enterprise-grade chassis, titanium high-efficiency redundant PSU and dedicated GPU air shroud, designed for continuous operation in data center environment with proper cooling maintenance.

PRODUCTS RELATED TO THIS ITEM
Lenovo ThinkSystem SR670V2 3U Rack Server for AI Training & High Performance Computing Recommend
ThinkSystem SR665V3 2U Rack Server 24 Bay 2.5 Inch Enterprise Server / 2*AMD EPYC 9474F/8* 128G  DDR5 / 960G M.2 *2/ RAID B540i-2i / XClarity Pro Standard 3Y Recommend
XFusion  G5500V7 /  8468*2 / 32G  5600 *4 /1.92T *3 / NVDIA  RTX4090D-48G*8 / 9560-4G / 3000W*4 Recommend
Dell PowerEdge R770 Mission Critical Rack Server with Dual Intel Xeon 6530P, 1.5TB DDR5 Memory, Enterprise RAID, 25GbE Networking & iDRAC10 Recommend
Dell PowerEdge R770 Mission Critical Rack Server with Dual Intel Xeon 6530P, 768GB DDR5, Enterprise RAID & 25GbE Networking Recommend
xFusion G5500 V6 AI GPU Server with Dual Intel Xeon Gold 6330, 8× NVIDIA A800 80GB SXM4, 200Gb InfiniBand & NVMe Storage Recommend
HPE ProLiant DL380 Gen11 2U Rack Server with Dual Intel Xeon Gold 6430, 256GB DDR5, Dual 25GbE, 32Gb FC, MR408i-o RAID & 3-Year Warranty Recommend
Dell  PowerEdge   R760 / 5418Y*2/ 64  DDR5  *12 /20TSATA*12/ H755/ 1400*2 Recommend