- Supports 5G network infrastructure
- Up to 4 x NVIDIA Tesla® PCIe GPGPU cards
- NVIDIA NGC Ready server
- 2nd Gen. Intel® Xeon® Scalable and Intel® Xeon® Scalable Processors
- 6-Channel RDIMM/LRDIMM DDR4, 24 x DIMMs
- Supports Intel® Optane™ DC Persistent Memory
- Dual 1Gb/s LAN ports (Intel® I350-AM2)
- 2 x 2.5″ hot-swappable + 2 x 2.5″ internal fixed HDD/SSD bays
- 2 x low profile PCIe Gen3 expansion slots
- 2 x 80 PLUS Platinum 2000W redundant PSU
- Optimized performance with Mellanox Infiniband and Ethernet 100G products
2nd Generation Intel® Xeon® Scalable Family Ready
GIGABYTE’s servers are fully ready to support the second generation of Intel® Xeon® Scalable Family processors, codenamed “Cascade Lake”, which bring the following major enhancements:
- Intel® Optane™ DC Persistent Memory: Built-in support for this revolutionary new product based on Intel’s 3D Xpoint technology
- Overall Performance: Higher CPU frequencies, improved turbo profiles vs. prior-gen Intel Xeon Scalable processors
- Increased DDR4 Memory Speed & Capacity: Up to 2933MHz (1 DIMM per channel on some SKUs), 16Gb based DIMM supported
- Intel Deep Learning Boost: Significantly accelerates inference performance for deep learning workloads optimized to use VNNI (Vector Neural Network Instructions)
- Security: Hardware mitigations for Meltdown / Spectre security vulnerabilities
Intel® Xeon® Scalable Family Processors
GIGABYTE’s Intel® Xeon® Scalable Processor family servers are available in dual socket configurations, and are compatible with the full family* of different SKUs (Bronze, Silver, Gold and Platinum) that are workload optimized to support different applications, from enterprise IT database, cloud and storage to the most high-demand HPC workloads.
* SKU compatibility will depend on the maximum CPU TDP supported for each system, please refer to each product’s specification page for further information
Advanced features built into the silicon include:
- Intel® QAT: Speed up data compression and cryptography with integrated Intel QuickAssist Technology (QAT).
- Intel® AVX-512: Intel Advanced Vector Extension 512 instruction set architecture enables workload-optimized performance and throughput increases for advanced analytics, high performance computing (HPC) applications, and data compression.
- Intel® VROC: Intel Xeon Scalable Processors come with built-in support for Intel VROC (Virtual RAID on CPU). And GIGABYTE Intel Xeon Scalable NVMe SKU servers all ship with an Intel VROC module as standard.
- Intel® Omni-Path: Selected GIGABYTE servers support CPU SKUs with integrated Intel Omni-Path Fabric connection
Intel® Optane™ DC Persistent Memory Ready
GIGABYTE’s second generation Intel® Xeon® Scalable family servers come ready to support Intel Optane DC Persistent Memory, a revolutionary new product that re-defines traditional memory & storage architectures by enabling a large persistent memory tier between DRAM and SSDs, that’s higher capacity than DRAM and faster than SSDs, enabling the user to bring more data closer to the CPU for faster time for insight.
Optane DC Persistent Memory Modules (“DCPMM”) integrate seamlessly into existing DDR4 DIMM slots, and use Intel’s 3D Xpoint memory technology, which unlike DRAM retains data after power is removed. The modules will be available in capacities up to 512GB – a massive increase compared to the current maximum 128GB of industry-leading DDR4 memory sticks – and can be addressed by the OS or application either as memory (Memory Mode) or storage (Application Direct Mode).
Note: the maximum number of DCPMM that can be installed per system and the DIMM slot positions that can be used for DCPMM will vary according to each server model due to thermal considerations. Please refer to each product’s specification page for further instructions.
G191-H44 System Block Diagram
G191-H44 Product Overview
Industry Leading GPU Density
Supports 4 x Computing Cards
The G191-H44 provides powerful parallel computing capabilities with a dense capacity of 4 x dual slot active or passive-cooled GPU cards. Each side of the system tray at the front and rear has a large bay for a GPU card, which is installed via a PCIe x16 (3.0 x16) expansion slot from a riser card connected to the main motherboard. The G191-H44 is ideal for workloads in artificial intelligence, machine learning, deep learning and for technical computing applications in financial services, life sciences and oil and gas exploration.
NVIDIA EGX Server: Accelerating AI at the Edge
This server has been officially recognized as an NVIDIA EGX Platform system: GPU accelerated platforms designed specifically for edge computing. The EGX Platform integrates an optimized hardware stack including GIGABYTE server hardware, NVIDIA Tesla® T4 accelerators and other components together with NVIDIA Edge Stack, a hardened software stack built for AI inferencing workloads using containers, that includes NVIDIA drivers, a CUDA Kubernetes plug-in, a CUDA Docker container runtime, CUDA-X libraries and containerized AI frameworks and applications, including TensorRT, TensorRT Inference Server and DeepStream.
As an EGX Platform system, this server can be used as an “on-premises AI-cloud in a box” that can be deployed as an edge server or as part of an edge micro data center, and provide powerful inferencing capabilities in locations where real-time answers are needed – in stores, hospitals, manufacturing plants and cities.
NVIDIA NGC Ready Server
NGC (NVIDIA GPU Cloud) is an online, free to use platform which features GPU accelerated containers, a container registry and software for deploying and running these containers. New NGC containers are published monthly and offer the latest features and superior performance on NVIDIA GPUs, can be used with multi-GPU & multi-node systems for scale-up & scale-out environments, and feature support for Docker & Singularity runtimes as well as VMware vSphere.
Supercharged with NVIDIA Tesla V100 or T4 accelerator cards, this GIGABYTE server has been fully validated and is ready to be used with NGC. NGC containers can be quickly and easily deployed onto this server to run GPU-accelerated workloads for deep learning, machine learning, and HPC. NGC takes care of all the plumbing so data scientists, developers, and researchers can focus on building solutions, gathering insights, and delivering business value.
Being NGC Ready gives data center managers the confidence to deploy this server knowing that it can be quickly set up to excel at a wide variety of CUDA-X accelerated applications. Optional direct enterprise grade support is also available from NVIDIA to further ensure peace of mind.
Optimized for High TDP
GIGABYTE systems are thermally designed with the highest rated CPU SKUs in mind so that top performance is enabled which ever Xeon Scalable CPU you choose.
Carefully Selected Components
GIGABYTE’s passive components are carefully selected to guarantee a stable operating environment and allow the processors and the memory to deliver their maximum performance. GIGABYTE’s server motherboards are built with long lasting components for a high degree of reliability and durability.
Ease of Deployment and Management
GIGABYTE continues to be responsive to market needs and to create user-friendly solutions, including:
- Tool-less rail kit included as a default across all new server models
- Supplying multiple connector options and tools where appropriate
[ Tool-less rail kit ]
Efficient, Reliable Power Supply
GIGABYTE SERVER Xeon Scalable servers are designed to guarantee high conversion efficiency regardless of load. Redundant 80 Plus ‘Titanium’ or ‘Platinum’ PSUs, combined with GIGABYTE’s precision engineering, translates to super efficient power usage which maximizes the system’s power : performance ratio and reduces OPEX for the user.
Intelligent Power Management Features
SCMP (Smart Crises Management / Protection)
SCMP is a GIGABYTE patented feature that will automatically force the system’s CPU to enter ULFM (ultra-low frequency mode for minimum power consumption) when the BMC is alerted to a PSU fault or error (such as power loss, power surge, overheating or a fan problem). This feature will prevent shutdown in systems with less than 1 + 1 PSU redundancy when one PSU is lost. In a 2U 4 Node server, only two nodes will be switched to low power mode, while two remaining nodes can continue normal operation. When the fault is resolved or the PSU is replaced, the system will automatically return to normal power mode.
System Loading Curve
GIGABYTE Server Management (GSM)
GIGABYTE Server Management (GSM) is GIGABYTE’s proprietary multiple server remote management software platform, available as a free download from each GIGABYTE server product page. GSM is compatible with either IPMI or Redfish (RESTful API) connection interfaces, and comprises the following sub-programs:
- GSM Server, a software program with an easy to use browser-based GUI to enable global remote monitoring and management of multiple GIGABYTE servers via each server node’s BMC.
- GSM CLI, a command-line interface program to enable global remote monitoring and management of multiple GIGABYTE servers via each server node’s BMC.
- GSM Agent*, a software program installed locally on each GIGABYTE server node that retrieves additional node information (CPU/Mem/HDD/PCI/…) from the OS and passes it to the BMC. This information can then be utilized by GSM Server or GSM CLI.
*GSM Agent is currently compatible with Avocent MergePoint IPMI 2.0 BMC firmware, but not yet compatible with Megarac SP-X BMC firmware.
- GSM Mobile, a remote server management mobile APP, available for both Android and iOS.
- GSM Plugin, a plugin that allows the user to manage and monitor GIGABYTE server nodes within VMware’s vCenter.
Intuitive & Informative User Interface
Starting from the GSM Server dashboard, the user can clearly understand the status of each node in the IT environment, including:
- Node connection status (system is online/offline)
- Node hardware sensor status (to detect hardware anomalies). Sensors measure voltage, fan speed and temperature
- Node usage rate status (allowing you to timely allocate more resources when they are needed), including:
- Power consumption
- Disk/RAID usage information
- Memory usage rate
- CPU usage rate
- Network card / PCI information
The user can also perform various management functions, including:
- Node remote access
- Power on / off / reset
- Launch iKVM
- Reboot BMC / BMC Account Configuration
- Network Configuration
- IPv4/IPv6 setting
- Alert Management
- SNMP trap setting
- Platform events filters
- Forwarding alerts by email
- BMC / BIOS / CPLD Update
- Power Consumption Limit Setting
- Group Management
- Create groups of nodes for BMC / BIOS update or manage power consumption settings
GSM Mobile for iOS and Android
A GSM Mobile app is available for download for both iOS and Android mobile phones for remote server management on the go.
Management through VMware vCenter
A GSM Plugin is available for VMware’s vCenter, allowing the user to perform remote monitoring and management of GIGABYTE server nodes without having to switch to a separate software platform.
GIGABYTE BMC Interface
Easy BIOS Upgrade
GIGABYTE has developed this integrated function (requiring no additional software) that allows users to update the BIOS of motherboards through the BMC controller:
- No need to install core components (CPUs/ memory/ drives/ OS)
- No need to power on the system (NB. PSU must be connected)
- Update individual (IPMI 2.0) or multiple boards simultaneously (CMD)
GIGABYTE Management Console (AMI MegaRAC SP-X)
This GIGABYTE server product utilizes a AMI MegaRAC SP-X platform for BMC server management, with a feature rich and easy to use browser-based graphical user interface.
Notable features include:
- RESTful API support (including the latest DMTF standards of Redfish) allows the administrator to integrate with 3rd party applications for server management
- HTML5-based iKVM remote management client included as a standard feature, no additional add-on license required to purchase
- Detailed FRU information from SMBIOS
- Pre-event automatic video recording feature from 10 to 30 seconds
- SAS / RAID controller monitoring feature
Easy BIOS Refresh
GIGABYTE has also developed this integrated function (requiring no additional software) that allows users to update the BIOS of motherboards through the BMC controller:
- No need to install core components such as CPUs, memory, drives and OS
- No need to power on the system (NB. PSU must be connected)
- Update individual (IPMI 2.0) or multiple boards simultaneously (Command Line)
Dynamic Fan Speed Control
GIGABYTE servers are enabled with dynamic fan speed adjustment. Individual fan speed will be automatically adjusted according to the system’s current temperature (according to CPU, DIMM, M.2, HDD, GPU temperature sensors), to achieve the best cooling and power efficiency. When the BMC detects a temperature change, fan speed will be automatically adjusted accordingly. Fan speed profiles may also be manually created and edited* when necessary according to the user’s custom requirements.
* Enabled on selected models & performed through BMC console management GUI
GIGABYTE maintains its commitment to green products, and adheres to the latest RoHS regulations (in effect July 2017).
VMware Technology Alliance Partner
Recognized as an Advanced Partner in VMware’s Technology Alliance Partner (TAP) program enables GIGABYTE to rapidly develop and certify joint solutions or go-to-market products that leverage VMware virtualization and cloud solutions, enabling our customers to modernize their data centers and implement IT infrastructure and application services with speed, agility, and cost optimization.
Enquire with your GIGABYTE sales representative for more information about combining or integrating this server product with VMware solutions.