Please login or create an account to save your build.Login or create an account now
For any urgent queries please call us on 023 9232 2500 (mon-fri -9am - 5pm)
Enter a title and optional short description, then click the 'Generate links' button to view share options.
Description :150 Characters remaining
Configure from £50639.21 inc VAT
Description: Intel Xeon Gold 5120 Processor
Cache: 19.25MB L3
Clockspeed: 2.2GHz (3.2GHz Turbo)
No. of Cores: 14
Description: 32GB DDR4 2666MHz ECC Registered DIMM Module
Description: 1U 4x SXM2 NVLINK2 GPU Server, 4x 2.5” HDD Bays.
Dimensions: 437 x 43 x 997 (WxHxD)mm
Form Factor: 1U
Drive Bays 2 1/2": 4
Description: NVIDIA Tesla V100 32GB GDDR5 SXM2 NVLINK2 Graphics Card, 5120 CUDA Cores
GPU Clock speed: 1290 MHz
GPU Turbo Clock Speed: 1530 MHz
Memory Clock Speed: 876 MHz
1752 MHz Effective
Number of displays Supported:
Description: Supermicro X11DGQ Motherboard
Form Factor: Proprietary
CPU Support: Dual Intel Xeon Scalable
Chipset: Intel® C621
LAN: Dual LAN with 10GBase-T with Intel® X540 10GbE Controller
Socket Type: 2x LGA 3647
USB: Rear: 2x USB 3.0 Ports
RAID: 0, 1, 5, 10
SATA Ports: 4
Memory Slots: 12
Max Memory Capacity: 1.5TB
Xeon Scalable Chips are designed and optimised for businesses and professionals, improving a systems efficiency with features not found in the Core Series. Advanced technologies paired with one another work to minimise system error and downtime for a business environment, preventing severe costs and maintenance labour for professionals. This is the next step in reliability.
Intel Xeon Gold provide more cores and higher frequencies for the high-performance workloads that you would find in mainstream data centres, cloud compute nodes and network and storage controllers. These workloads benefit for the Gold family’s quad socket compatibility and the ability of some of these processors to address double the amount of RAM from the silver and bronzes, allowing up to 6TB of RAM in a 4-socket configuration.
The latest CPU architecture boasts up to 6 memory channels per socket and a bandwidth of up to 2666MHz, with memory capacity benefits available up to 12 DIMMs per socket. This represents a 50% generation-on-generation increase. The system’s increased memory performance is also future-proofed with the potential to support up to Apache Pass DIMM.
NVLink is the standard communication method for high performance GPU interconnects, demanded from 4 GPU and 8 GPU systems to solve their heavy, complex problems with compute-intensive applications. Fast and scalable multiprocessor interconnects have never been more important to a systems success. NVLink provides higher bandwidth, more links, and enhanced scalability for multi-GPU system configurations, allowing modern servers take advantage for greater scalability and ultrafast abilities.
Novatech servers are specifically optimised to use GPUDirect for the tremendous horsepower required from them. GPUDirect allows multiple GPUs, network adapters, SSDs and other devices utilising it to directly read and write CUDA host and device memory, ending unnecessary memory copies and drastically lowering CPU overhead and latency when using compute-intensive applications. This results in significant improvements of a system’s performance for an applications data transfer time.
A GIGABYTE Server supports 8 x GPU cards in a 1U form factor server chassis for boosted, unrivalled performance.