Commit 1269e0ac authored by Garhan Attebury's avatar Garhan Attebury
Browse files

Update facilities.md

parent dd8050c3
......@@ -2,23 +2,23 @@
title: "Facilities of the Holland Computing Center"
---
This document details the equipment resident in the Holland Computing Center (HCC) as of October 2019.
This document details the equipment resident in the Holland Computing Center (HCC) as of December 2020.
HCC has two primary locations directly interconnected by a 100 Gbps primary link with a 10 Gbps backup. The 1800 sq. ft. HCC machine room at the Peter Kiewit Institute (PKI) in Omaha can provide up to 500 kVA in UPS and genset protected power, and 160 ton cooling. A 2200 sq. ft. second machine room in the Schorr Center at the University of Nebraska-Lincoln (UNL) can currently provide up to 100 ton cooling with up to 400 kVA of power. Dell S4248FB-ON edge switches and Z9264F-ON core switches provide high WAN bandwidth and Software Defined Networking (SDN) capability for both locations. The Schorr and PKI machine rooms both have 100 Gbps paths to the University of Nebraska, Internet2, and ESnet as well as backup 10 Gbps paths. HCC uses multiple data transfer nodes as well as a FIONA (flash IO network appliance) to facilitate end-to-end performance for data intensive workflows.
HCC's resources at UNL include two distinct offerings: Rhino and Red. Rhino is a linux cluster dedicated to general campus usage with 7,040 compute cores interconnected by low-latency Mellanox QDR InfiniBand networking. 360 TB of BeeGFS storage is complemented by 50 TB of NFS storage and 1.5 TB of local scratch per node. Each compute node is a Dell R815 server with at least 192 GB RAM and 4 Opteron 6272 / 6376 (2.1 / 2.3 GHz) processors.
The largest machine on the Lincoln campus is Red, with 14,160 job slots interconnected by a mixture of 1, 10, and 40 Gbps ethernet. More importantly, Red serves up over 11 PB of storage using the Hadoop Distributed File System (HDFS). Red is integrated with the Open Science Grid (OSG), and serves as a major site for storage and analysis in the international high energy physics project known as CMS (Compact Muon Solenoid).
The largest machine on the Lincoln campus is Red, with 15,984 job slots interconnected by a mixture of 1, 10, 25, 40, and 100 Gbps Ethernet. More importantly, Red serves up over 11 PB of storage using the Hadoop Distributed File System (HDFS). Red integrated primarily serves as a major site for storage and analysis in the international high energy physics project known as CMS (Compact Muon Solenoid) and is integrated with the Open Science Grid (OSG).
HCC's resources at PKI (Peter Kiewit Institute) in Omaha include Crane, Anvil, Attic, and Common storage.
HCC's resources at PKI (Peter Kiewit Institute) in Omaha include the Crane and Anvil clusters along with the Attic and Common storage services.
Crane debuted at 474 on the Top500 list with an HPL benchmark or 121.8 TeraFLOPS. Intel Xeon chips (8-core, 2.6 GHz) provide the processing with 4 GB RAM available per core and a total of 12,236 cores. The cluster shares 1.5 PetaBytes of Lustre storage and contains HCC's GPU resources. We have since expanded the existing cluster: 96 nodes with new Intel Xeon E5-2697 v4 chips and 100GB Intel Omni-Path interconnect were added to Crane. Moreover, Crane has 43 GPU nodes with 110 NVIDIA GPUs in total which enables the most state-of-art research, from drug discovery to deep learning.
Anvil is an OpenStack cloud environment consisting of 1,520 cores and 400TB of CEPH storage all connected by 10 Gbps networking. The Anvil cloud exists to address needs of NU researchers that cannot be served by traditional scheduler-based HPC environments such as GUI applications, Windows based software, test environments, and persistent services. In addition, a project to expand Ceph storage by 1.1 PB is in progress.
Anvil is an OpenStack cloud environment consisting of 1,520 cores and 400TB of CEPH storage all connected by 10 Gbps networking. The Anvil cloud exists to address needs of NU researchers that cannot be served by traditional scheduler-based HPC environments such as GUI applications, Windows based software, test environments, and persistent services.
Attic and Silo form a near line archive with 1.0 PB of usable storage. Attic is located at PKI in Omaha, while Silo acts as an online backup located in Lincoln. Both Attic and Silo are connected with 10 Gbps network connections.
In addition to the cluster specific Lustre storage, a shared common storage space exists between all HCC resources with 1.9PB capacity.
In addition to the cluster specific Lustre storage, a shared storage space known as Common exists between all HCC resources with 1.9PB capacity.
These resources are detailed further below.
......@@ -38,6 +38,8 @@ These resources are detailed further below.
## 1.2 Red
* USCMS Tier-2 resource, available opportunistically via the Open Science Grid
* 18 2-socket Xeon Gold 6248R (3.00GHz) (96 slots per node)
* 1x 2-socket AMD EPYC 7402 (2.8GHz) with 1x V100S GPU (96 slots)
* 46 2-socket Xeon Gold 6126 (2.6GHz) (48 slots per node)
* 24 2-socket Xeon E5-2660 v4 (2.0GHz) (56 slots per node)
* 16 2-socket Xeon E5-2640 v3 (2.6GHz) (32 slots per node)
......@@ -51,12 +53,14 @@ These resources are detailed further below.
* 40 2-socket Opteron 6128 (2.0GHz) (32 slots per node)
* 40 4-socket Opteron 6272 (2.1GHz) (64 slots per node)
* 11 PB HDFS storage
* Mix of 1, 10, and 40 GbE networking
* Mix of 1, 10, 25, 40, and 100 GbE networking
* 2x Dell Z9264F-ON switches
* 1x Dell S5248F-ON switch
* 1x Dell S6000-ON switch
* 3x Dell S4048-ON switch
* 5x Dell S3048-ON switches
* 2x Dell S4810 switches
* 2x Dell N3048 switches
* 5x Dell N3048 switches
## 1.3 Silo (backup mirror for Attic)
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment