Dense GPU Server Configurations

11/29/2017 | Hardware, Servers, Solution Overviews

Equus delivers market leading high density GPU configurations using the latest NVIDIA and Intel Xeon Phi™ technologies.

Equus continues to enable a highly scalable, energy efficient future for parallel computing with our latest supercomputing solution. Utilizing our expertise in maximizing compute density, performance and power efficiency, Equus has configured the following server for high GPU density and performance in a 1U server. This 1U server can hold 4 NVIDIA Tesla K80 GPUs, and uses 9 standard and 2 optional 4cm heavy duty fans to keep this powerful server cool. This system is currently configured with Dual Xeon® E5-2650v4 12-Core 2.2Ghz 30M Cache, 256GB DDR4 2400 RDIMM, Intel Dual Port 10GbE SFP+ with Optics and Single Port Mellanox ConnectX-3 VPU Pro for 10/40/56Gb connectivity, Micron 5100 Pro 240GB SATA SSD boot drive, Tool less Quick Rails, and DCMS Data Center Management Software with iKVM. This server will run in up to 35 degree ambient temperature allowing for warm datacenter operation to save on cooling costs.

Key Benefits:
• A CPU direct connect design optimizes signal integrity and airflow to eliminate the GPU pre-heating, PCI-E extension cables, and re-drivers.
• Reduced complexity, cost, power consumption and latency as compared to other vendor solutions
• Higher power efficiency by using fully redundant Titanium Level (96%+) power supplies
• Supports up to 4x 2.5” drive bays with fully redundant power supplies
• Uniquely supports NVIDIA GeForce GPU cards for customer applications

Dense GPU Server Configuration Details

Key Features:
• Dual socket R3 (LGA 2011) supports Intel Xeon processor E5-2600 v4/v3 family; QPI up to 9.6GT/s
• Up to 512GB ECC 3DS LRDIMM, up to DDR4-2400 MHz; 16x DIMM slots
• 4x PCI-E 3.0 x16 slots (4x GPU/Xeon Phi cards opt.), 2x PCI-E 3.0×8 (in x 16) LP slot
• Intel i350 Dual port GbE LAN
• 2x 2.5″ Hot-swap drive bays, 2x 2.5″ internal drive bays
• 9x 4cm heavy duty counter-rotating fans with air shroud & optimal fan speed control
• 2000W Redundant Power Supplies Platinum Level (94%+)

QuantityVendorDescription
1SupermicroSupermicro 1028GQ-TR 1U Xeon E5-2600v4 4xGPU DDR4 2x 2.5″HS 2000W RPSU
2IntelE5-2650V4 2.2GHZ 30M 12-Core Xeon CPU
8Crucial32GB DDR4 2400 ECC REG
1MellanoxMellanox ConnectX-3 10/40 GbE / 56Gb Infiniband Network Adapter
1Intel2port SFP+ 10G PCIe3.0x8 Direct Attach Twinax
2IntelINTEL DUAL RATE SFP+ MODULE XFP 1x10GBASE-SR
1MicronMicron 5100 PRO 240GB SATA 2.5
4NVIDIANVIDIA Tesla K80 2x12GB/DDR5 PCI-E3.0 300W
1SupermicroPower Cable for Teslaq K80 1028GQ-TR
2SupermicroPower Cable for Teslaq K80 1028GQ-TR
1SupermicroPower Cable for Teslaq K80 1028GQ-TR
2Supermicro40x40x56 mm 23.3K-20.3K RPM Counter-rotating Fan
1SupermicroDataCenter management package (per node license)
1EquusMid Server – 3 Year Onsite Warranty

Management Software

Manage and Maintain Servers with Included and Optional Management Software

  1. Remotely manage servers deployed worldwide.
  2. Manage server health, power consumption and firmware remotely using agent and agent-less mechanisms.
  3. Manage hardware with no impact on applications.
  4. Perform monitoring, configuration and update operations without affecting application performance or continuity.
  5. Using out-of-band (OOB) utilities.
  6. Server management functions accessed through utilities’ command line interfaces to support existing datacenter automated management frameworks.
  7. Additionally, our server management utilities provide seamless integration with Nagios and other industry standard plugins.
Standard ManagementOOB ManagementDCMS Management
Included with ServerIncludes Standard Management PlusIncludes Standard + OOB Management Plus
KVM/JAVAOut-of-BAND System Checks,Configuring, Mousemode, Fanmode,
KVM/HTML5 supportSystem Utilization, Asset InformationRadius, AD Through APIs
IPMI 2.0OOB/in-band BIOS/BMC ManagementRemote Syslog
DCMIGetting/Clearing Event LogRestful APIs
Web Based GUITrusted Platform Module ProvisioningScripted Virtual Media
In-band BIOS updatesMount/Unmounts ISO images fromUnified Hardware Management
BMC FW updatesSAMBA/HTTPRemote Power Management/Monitoring
LDAP/Active DirectoryRemote Screenshot CaptureDynamic DNS
Virtual MediaRemote Keyboard Operation3rd-party SW Integration through RedFish
SNMP AlertsRedfish APIs24/7 Health and Power Management
SMTP AlertsChanging system boot orderLevel II Support on utilities
SMASH and CLP SupportvCenter, SCOM and OpenStack Plugins
VLAN SupportFeature Updates and Support
Event LogWS-MAN API
Serial over LAN
Remote Power Control
Hardware Health Monitoring
HTTPS
DHCP
SSH CLI
Dedicated NIC
Local Users
Embedded Remote Support
Agentless Management
Role Based Authority
Multiple User Profiles
IPv6