All Australian Company

Sales: 8am to 8pm 7 days

PH : 1300 00 1000

High Performance Compute Servers/HP Enterprise: Apollo, 6500/XL270D, High, Performance, Compute, system, with, dual, 8268, providing, 48, cores, at, 2.9GHZ, 1TB, RAM, 3, *, 1.6TB, SAS, S,

Apollo 6500/XL270D High Performance Compute system with dual 8268 providing 48 cores at 2.9GHZ, 1TB RAM, 3 * 1.6TB SAS S

Our Price (inc. GST):

$199,000

$4,214.84 /month*

Freight Cost:

Syd: 

$74.02

Melb: 

$103.99

Calculate freight to your postcode:

Full Australian Warranty

Australian Distributors

Product is in a sealed box

PRODUCT ID: AP6500HPC2

Description

Apollo 6500/XL270D High Performance Compute system with dual 8268 providing 48 cores at 2.9GHZ, 1TB RAM, 3 * 1.6TB SAS SSDs, dual 2200W power and 4 * V100 SXM2 32GB modules

The Apollo 6500/XL270D High Performance compute server is an AI powerhouse utilising Nvidia NVLINK and SXM2 bus to provide up to 300gbps transfer between GPUs. The 4 onboard GPU modules provide over 125TFLOPs (FP32) of processing power making this system one of the most powerful GPU-accelerated AI solutions available.

This configuration has dual Intel Platinum 8268 processors providing 48 cores at a fast 2.9ghz. It comes with 1TB preinstalled and 3 * 1.6TB SAS SSDs with fast 12gbps transfer rate. Dual 2200W power supplies support the 4 onboard V100 SXM2 modules each with 32GB high speed GPU RAM. The 4 GPUs provide a total of 20480 CUDA compute cores and 2560 AI specific Tensor cores.

Suitability
This system is recommended for intensive High Perfrormance compute and AI Machine and Deep Learning workloads. It is also suitable for render farm workloads and many financial market applications. The high performance interconnected GPUs make it a good choice for suitable for complex simulation and modeling workloads.

Expandability
RAM is easily expanded to 1.5TB (or 3TB with LRDIMMs) and the server can be expanded with an additional 9 drive bays ( 12 bays total - we recommend SAS SSDs) and with up to 4 NVME. There are connectivity options for Infiniband 100GBE and Omnipath available.

Software Stack

Nvidia's V100 performance GPUs are supported by many GPU-enabled HPC and AI applications.

For AI, we recommend Nvidia's software framework and containers. Nvidia provide a number of containerised solutions for AI, medical imaging and scientific application. In addition there are a number of approaches to converting existing applications to support GPU acceleration. Ask us for details.

We can provide this server with bare metal installation of Linux, docker (or alternative) and containers or as a virtualised solution under VMWare ESXi and vComputeServer (separate licensing costs apply). The virtualised solution allows for multiple workloads on multiple virtual machines with GPU resources being able to be dynamically allocated between VMs and workloads.

Warranty
This server is covered by a HP Enterprise 3 year warranty with 4 hour response to site 24 * 7. The Nvidia V100 modules are also covered by a 3 year warranty.

Last updated: Friday, 29 March 2024

Warranty

This product is sourced in Australia and carries a full Australian Manufacturers Warranty

Brochure

Brochure