We invite you to try
PERCEUS and find out why so many organizations, from private
enterprise to government research labs trust PERCEUS every day to
provision their mission critical infrastructure. PERCEUS is free,
open source, and has been tested beyond petascale deployments. We
thank the DoD and DoE for utilizing our software on systems bigger
than we ever initially designed for, and most of all we thank our
users for taking PERCEUS where it is today.
What is PERCEUS, and what
does the acronym stand for?
PERCEUS stands for 'Provision Enterprise Resources and Clusters
Enabling Uniform Systems', and it minimally says what the program
PERCEUS is the original 100% GPL, open source, free, infrastructure
provisioning solution proven at petascale (deployments at this scale
require multiple masters with proper fabric and filesystems) you can
always reach us for help. PERCEUS is an alrtuistic project and we
have given all away to further science.
PERCEUS has many firsts:
- It was the first automated and heterogeneous stateless
provisioning solution (not just PXE booting an ISO image or for a
single flavor of Linux),
- the first to offer advanced IPMI integration,
- the first to control C and P states of iCore7 CPUs (even before
- the first 1 wire InfiniBand only provisioning system,
- the first to release integrated cloud installations (Caos-NSA
and GravityOS had this years before others who call themselves the
- the first to integrate XFS, ZFS and GravityFS,
- the first provisioned HPC “Burst Buffers” (GravityFS Cache with
Intel SSD’s even before SSD’s were on the market).
PERCEUS has been described before
as a two stage rocket, a Swiss army knife, and even a super hero
cape. These descriptions acknowledge PERCEUS' provisioning
mechanism, modular versatility, and an ability to deliver a powerful
administrative solution to clustered environments. Without using
metaphor, PERCEUS can be described as a program that deploys
consistent operating system and application software images to
computing infrastructure over the network. It simplifies the
management of scalable systems built with commodity or customized
server hardware by providing a new way to boot the entire
deployment, no matter the scale or growth requirements. Flexible
configuration and module options give PERCEUS the means to enable a
range of workload solutions, from classic high performance and high
throughput computing clusters to content delivery networks, render
farms, or shared hosting systems such as private clouds or
enterprise compute grids.
PERCEUS doesn't just enable
full-throttle High Performance Computing. Green scalable
infrastructure deployments can take advantage of stateless and
hybrid provisioning, deep CPU control (such as C and P state control
of Intel processors), Greening technology for memory, APU, and GPU,
efficiency and performance tuned kernels, advanced node data center
distribution in response to power consumption levels and energy
rates, bare metal or virtual machine provisioning, and support for
low power ARM/Atom/MIPS/Cell and embedded processors.
Thanks to our contributors and users for helping evolve and improve
PERCEUS. We really appreciate the loyalty and greatly enjoy the
GravityPark | Infiscale - Help with development and
Intel - For providing hardware access and demonstrations, helping us
guarantee our users are supported, and early opportunities to enable
new advanced features.
Numerous companies, .gov agencies, organizations, user groups, and
amazing individuals - For greatly appreciated feedback on massive
HPC deployments, specialty devices, and uses of PERCEUS. We love you
guys and love to see you succeed with perceus.
DVLABS - Gracious hosting on their media delivery network.