Apollo 8000 System
Achieve incredible insights in science and technology—through an innovative, water-cooled supercomputer that actually reduces your carbon footprint.
The Ecology of Supercomputing
The HPE Apollo 8000 System advances the science of supercomputing like never before. It’s the world’s first warm-water liquid cooling system—1,000 times more efficient than air-cooled systems.
This cooling technology helps deliver a faster, more sustainable infrastructure for research workloads. With groundbreaking speed, density, efficiency, and economy, you can run world-changing analytics and simulations while reducing your impact on the world’s ecosystems.
Breakthrough Power Meets Energy Efficiency
The HPE Apollo 8000 System leads the industry in teraflops per rack. This performance is the result of highly efficient liquid cooling that’s delivered risk-free, supporting data center energy recycling and lowering energy costs.
- Up to 80 kW of power (4 x 30A 3ph 480AC) supporting up to 144 servers per rack
- Liquid cooling allows system to consistently stay in turbo mode
- Novel heat pipes provide liquid-cooling without tray liquid
- Up to 20% reduction in job time vs. air-cooled systems
Lower Ownership Costs
- Chillerless cooling, for lower CapEx facility costs
- Warm water cooling and waste heat reuse, for reduced OpEx costs
- Higher performance lowers license fees
- Hundreds of teraflops of compute power in a very compact space; 4X teraflops per square foot
- 72 trays per rack
- Each tray contains either, 2X2P Intel® Xeon Nodes or 1X2P Intel® Xeon Node plus 2 Accelerators
The Right Configurations for the Best Research
Get started with basic building blocks—one HPE Apollo f8000 Rack and one HPE Apollo 8000 iCDU Rack. The system offers up to 144 x 2P servers per rack.
HPE Apollo f8000 Rack
Move to industry-leading performance density with the HPE Apollo f8000 Rack, which holds up to 144 ProLiant servers. You get all the CPU power, accelerator, memory, and networking options you need, in a standard rack footprint.
HPE Apollo 8000 iCDU Rack
Get 320kW of efficient and resilient cooling, using the lower half of a standard rack footprint. The HPE Apollo 8000 Intelligent Cooling Distribution Unit (iCDU) Rack connects to a standard 2.5 inch facility water pipe, and cools up to four HPE Apollo f8000 racks.
HPE Apollo in Action
World’s Most Energy-Efficient Data Center
The National Renewable Energy Laboratory uses HPE Apollo 8000 servers to gain five times more computational capability than their previous systems, while saving $1 million dollars per year in operating expenses.
Gaining 1,700 Servers on One Platform
The Cyfronet Academic Computing Center AGH in Poland is home to Prometheus, the world’s largest HPE Apollo 8000 installation, with 41,500 processor cores and 1,728 nodes in only 13 square meters.
Over 400 Nodes Plus “Green” Features
Hikari, an HPE Apollo 8000 supercomputer, will be powered by solar energy with a battery backup, creating the most energy-efficient system at the Texas Advanced Computing Center (TACC).
Technology That Sets HPE Apollo Apart
We offer complete solutions, starting with Intel processors and purpose-built platforms, then layering innovative software, management tools, and fabric.
HPC-Specific Software Solutions
HPE works with leading industry partners to provide software that maximizes application performance, optimize resources, and improve operational productivity. The first, HPE Core HPC Software Stack, is a complete software set for creating, optimizing, and running HPC applications. The second, HPE Insight Cluster Management Utility (CMU), an integrated, easy-to-use point solution for provisioning, management, and monitoring clusters of any scale.
A Complete Portfolio of Management Components
HPE rack- and system-level management solutions help you get the most out of your infrastructure—and your budget—and improve efficiency without affecting performance. They provide extensive monitoring, cluster administration, lifecycle management, power control and measurement, pooled power, and system management.
High-Performance Fabric Interconnects
All HPE Apollo Systems support the latest high-bandwidth, low-latency interconnect technologies. They include InfiniBand, which provides the performance and scalability required by high-performance compute (HPC) clusters and data center applications. HPE Apollo also supports Intel® Omni-Path Fabric, which delivers HPC workloads and scales to tens of thousands of servers.
Tap into our extended ecosystem of partners, integrators and ISV’s ready to deliver the right HPE Apollo system for your unique needs.