
Maximizing GPU usage of your DGX systems throughout your AI journey
This is an all-flash NVMe Storage Server with very high performances and ultra-low latency.
Direct access to the data from the DGX-1 GPU as if it was internally stored with the RDMA protocol.
30% FASTER TRAINING | 50% LOWER COST | 100% SCALABILITY |
Real life deep learning projects show a massive 30% improvement in training times when compared to other solutions.Excellent performance with the standard storage synthetic benchmarks, with bandwidth, latency and IOPS, leaving others behind. | Cost and affordability are a key design focus. By removing the need for expensive storage controllers, costs are dramatically reduced, and more of your investment is spent on GPU and NVMe resource providing greater productivity and ROI. | With Scale-UP and Scale-OUT, both capacity and performance are not limited. Regardless of where you start, scaling is simple, fast and on demand.Providing a future proof solution and benefiting from the lowering trend of flash costs. |
NVMe over Fabric Low Latency Storage
Excelero delivers low-latency distributed block storage for web-scale applications. NVMesh enables shared NVMe across any network and supports any local or distributed file system. The solution features an intelligent management layer that abstracts underlying hardware with CPU offload, creates logical volumes with redundancy, and provides centralised, intelligent management and monitoring.
Applications can enjoy the latency, throughput and IOPs of a local NVMe device with the convenience of centralised storage while avoiding proprietary hardware lock-in and reducing the overall storage TCO. NVMesh features a distributed block layer that allows unmodified applications to utilize pooled NVMe storage devices across a network at local speeds and latencies. Distributed NVMe storage resources are pooled with the ability to create arbitrary, dynamic block volumes that can be utilised by any host running the NVMesh block client.
Start Small. Scale Only When Needed
Ensuring your project’s funds are better spent
Today’s AI servers consume and analyse data at much higher rates than traditional storage solutions an deliver. Resulting in low GPU utilisation and dramatically extending training times as well as decreasing productivity. PNY 3S-2400 has been developed from the ground up for AI workloads and NVIDIA DGX optimisation. Delivering ultra-low latency and tremendous bandwidth at a price which allows more investment to be made on GPU resource and less on expensive, slower storage.
Unlike most storage solutions where the initial investment dictates future growth and performance, often resulting in the need to overspend for potential future growth.
PNY’s NVmesh design can scale in stages to suit your project without any limitation. Just purchase the capacity and performance you need today and feel secure that as you scale, so can your capacity and performance.
Regardless of your resilience needs, PNY 3S-2400 storage server will maintain ultra-low latency and high performance for AI / DL workflows. Resilience options range from RAID0 to full mirrored solutions with No Single Point of Failure.
With full support and a range of on-site options, you can choose the package most effective for your organisation or project.
The capacity extends : from 32 TB (4 x 8 TB) to 360 TB (24 x 15 TB) per node
12GB/sec and > 2 million IOPS with only 4 drives
Latency < 90uSec
Connectivity : 2 x QSP28 EDR Inifinband /100Gb/s Ethernet
CPU Type : 2nd Gen. Intel Xeon Scalable Processors
Architecture : Cascade Lake-SP (Purley)
Memory : 96GB DDR4 ECC Registered
Front Panel connectivity :
Back Panel connectivity :
Power : Hot-swap redundant PSUs
Form Factor : 2U, 710mm deep
Dimensions : 438 x 87.3 x 704.9 (mm WxHxD)
Excelero Software License : 3 years. 5-years license available on request
Warranty : 3 Years