Differences

This shows you the differences between two versions of the page.

--- public:service:loewe-csc [2018/04/04 14:01] – [Links] geier
+++ public:service:loewe-csc [2019/08/07 14:07] (current) – removed jankowiak
@@ Line 1: / Line 1: @@
-====== LOEWE-CSC ======
-The LOEWE-CSC is a general-purpose computer cluster based on several CPU and GPU architectures. The system was installed in Industriepark Höchst in late 2010 and has been constantly upgraded since then.
-{{ :public:loewe-3d-2.jpg?370 }}
-**Cluster performance:**
-  * CPUs performance (dp((double precision))):  226 TFlop/s (peak)
-  * GPUs performance (dp):  597 TFlop/s (peak)
-  * Cluster performance HPL (2010): 299.3 TFlop/s
-  * Energy efficiency Green500 (2010): 740.78 MFlop/s/Watt
-**Hardware:**
-  * 825 compute nodes in 36 water-cooled racks, two server rooms,
-  * thousands of CPU cores (Intel, AMD, see table below),
-  * GPGPU accelerators (AMD),
-  * 64-128 of RAM per node,
-  * over 2 PB aggregated storage capacity,
-  * QDR and FDR InfiniBand interconnects,
-  * a parallel scratch filesystem with a capacity of 764 TB and an aggregated bandwidth of 10 GB/s.
-^  Nodes  ^  CPU  ^  Cores \\ per CPU  ^  Cores  \\ per Node  ^  HT((Intel Hyper-Threads per Node)) ^  RAM \\ per Node  ^  GPU  ^
-|  438 |2xAMD Opteron 6172 |    	12 |  24|   -|	 64 GB |AMD HD 5800 (1GB) |
-|  198 |2xIntel Xeon E5-2670 v2 |   	10 |  20|  40|	128 GB |n/a |
-|  139 |2xIntel Xeon E5-2640 v4 |   	10 |  20|  40|	128 GB |n/a |
-|   50 |2xIntel Xeon E5-2630 v2 |   	 6 |  12|  24|	128 GB |2xAMD S10000 \\ (12GB)|
-\\
-{{..:loewe1.jpg?240 }} {{..:loewe2.jpg?128 }} {{..:loewe-csc-04b.jpg?300}}
-===== InfiniBand Network =====
-Most of the compute nodes (the ones in the main server room) are connected through a mixed 4X QDR-FDR InfiniBand 2:1 blocking fat-tree network (FDR at the spine level, FDR and QDR at the edge level). In the second (smaller) server room there are two non-blocking InfiniBand "islands" (one per cabinet), one running at 4X QDR and one at 4X FDR. The connection between the rooms is a 20 Gb/s Ethernet. Compute jobs are scheduled in a way that avoids mixed allocations, i.e. jobs are either kept within the fat-tree or on an island.
-===== Links =====
-[[https://www.vertivco.com/de-emea/products-catalog/critical-power/power-distribution/vertiv-mph2-mphb1243/]]
-{{:public:service:pdu.zip}}