Guillimin Hardware Information

The Guillimin system (see Figure 1) is actually made of five main components that are connected together:

All these main components are described in the following sections.

Figure 1 - Architecture Schematic of the Combined Phase 1 and 2 of Guillimin

Phase 1 and Phase 2


Icon Guillimin - High Performance Computing Cluster

The Guillimin general purpose compute cluster resource is one of the largest in Canada providing access more than 21000 processing cores running on the latest Intel Sandy Bridge and Westmere processesors. The cluster makes use of a Non-blocking and 2:1 blocking QDR InfiniBand network capable of 40 Gbps to each node and for fast access to the large scale storage systems operating with an optimized parallel file system.

Specifications of Compute Nodes

PhaseNode TypeCountProcessorsTotal coresMemory (GB/core)Total memory (GB)Network
Phase 1 SW 600 Dual Intel Westmere EP Xeon X5650
(6-core, 2.66 GHz, 12MB Cache, 95W)
7,200 3 21,600 InfiniBand QDR 2:1 blocking
HB 400 4,800 2 9,600 InfiniBand QDR 1:1 non-blocking
LM 188 2,256 6 13,536
Phase 2 SW2 8 Dual Intel Ivy Bridge EP E5-2650v2
(8-core, 2.6 GHz, 20MB Cache, 115W)
128 4 512 InfiniBand QDR 2:1 blocking
XLM2 4 Dual Intel Sandy Bridge EP E5-2670
(8-core, 2.6 GHz, 20MB Cache, 115W)
64 32 2,048
SW2 216 3,456 4 13,824
LM2 146 2,336 8 18,688 InfiniBand QDR 1:1 blocking
XLM2 6 96 16 1,536
1 Quad Intel Sandy Bridge EP E5-4620
(8-core, 2.2 GHz, 16MB Cache, 95W)
32 12 384
1 32 32 1,024
Total or
Average
- 1,570 - 20,400 4.06 82,752 29.5 Gbps

 


Icon Guillimin - Accelerator Technologies Cluster (Guillimin - ATC)

The Guillimin Accelerator Technologies Resource consists of leading-edge accelerator systems making use of both Graphical Processing Units (GPUs) from NVIDIA and the Many-Integated Core (MIC) Phi architecture from Intel. The accelerator systems provide a platform for developing new programming models that can harness novel computational hardware that can provide a significant boost in compute capability and performance.

Specifications of the Accelerator Worker nodes

PhaseNode TypeCountProcessorsTotal coresMemory (GB/core)Total memory (GB)CardsTotal Peak SP FP (TFlops)Total Peak DP FP (TFlops)Network
Phase 2 AW nVidia 8 Dual Intel Ivy Bridge EP E5-2650v2
(8-core, 2.6 GHz, 20MB Cache, 115W)
128 4 512 Dual nVidia K20,
Peak SP FP: 3.52 TFLops, Peak DP FP: 1.17 TFlops
56.3 18.7 InfiniBand QDR 2:1 blocking
25 Dual Intel Sandy Bridge EP E5-2670 (8-core, 2.6 GHz, 20MB Cache, 115W) 400 8 3,200 176 58.5
25 400 4 1,600 176 58.5
AW Intel 50 800 4 3,200 Dual Intel Phi 5110P,
60 cores, 1.053 GHz, 30 MB cache, 8 GB memory, Peak SP FP: 2.0 TFlops, Peak DP FP: 1.0 TFlops
200 100
Total or
Average
- 108 - 1,728 5 8,512 - 608 236 20 Gbps

 


Icon Guillimin - Data Storage and Archival Resource (Guillimin - DataSTAR)

The Guillimin HPC Centre provides access to significant data storage and archival facilities. At present 5 PetaBytes (PBs) of disk and 2.5 PB of tape archive are available to researchers and industry clients who have large data storage need requiring both large capacity and high-speed access to data. The facility makes use of the GPFS parallel file system specifically tuned for high-throughput access and configured with the ability to manage the full life cycle of data management through an attached Hierarchical Storage Management (HSM) system.

Technology Highlights

 


Icon Guillimin - Grid and Cloud Computing Resource (Guillimin - GrCloud)

The Guillimin Grid and Cloud Computing environment provides customizable and secure entry portals through which users can tap into the Guillimin facilities for their computing and data management needs. Resources can be provisioned through dedicated and reliable Grid and Cloud computing services that enable clients to securely compute, store and manage their data.