Differences

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

--- users:hardware [2016/03/15 14:49] – ecalore
+++ users:hardware [2017/09/04 13:20] (current) – ecalore
@@ Line 1: / Line 1: @@
 ====== Available Hardware ======
-The COKA Cluster is made up of 4 nodes named:
+The COKA Cluster is made up of one master node acting as a login front-end and 4 compute nodes named:
+// node01, node02, node03, node04 //
+In the compute nodes hyper-threading is disabled, although it is enabled in the login node.
-// node01, node02, node03, node04 //
+The computing hardware of one node can be summarized by this schema:
+{{:users:nodo-cluster-schema.png?direct&350 | Topology of one node of the Cluster}}
-{{:users:node-topology.png?direct&280 | Topology of the hardware of one node of the Cluster}}
+Compute nodes are Supermicro [[http://www.supermicro.com/products/system/4u/4028/SYS-4028GR-TR.cfm|SYS-4028GR-TR]].
 <WRAP center round box 40%>
-Each node embed:
+Each node embeds:
   * 2 x Intel Xeon CPUs
   * 8 x NVIDIA K80 GPUs
   * 2 x Mellanox ConnectX-3 Single FDR 56Gb/s Infiniband cards
+  * 256 GB of RAM memory
+  * 512 GB SSD mounted in /tmp
 </WRAP>
-===== Intel Xeon E5- =====
+A more detailed view is given by the lstopo output:
+{{ :users:node-topology.png?direct&600 | Detailed topology of the hardware of one node of the Cluster}}
+===== Intel Xeon E5-2630v3 =====
+Each node has two sockets, each equipped with an:
+[[http://ark.intel.com/products/83356/Intel-Xeon-Processor-E5-2630-v3-20M-Cache-2_40-GHz|Intel® Xeon® Processor E5-2630 v3 (20M Cache, 2.40 GHz)]]
 ===== NVIDIA K80 =====
 Each Cluster Node hosts 8 x NVIDIA K80 boards and each board hosts 2 actual GPUs, for a total of 16 CUDA devices.
+[[http://www.nvidia.com/object/tesla-k80.html | NVIDIA Tesla K80]]
 ===== Mellanox ConnectX-3 =====
+Each node has two MT27500 Family [ConnectX-3] Infiniband HCA:
+[[http://www.mellanox.com/page/products_dyn?product_family=119&mtag=connectx_3_vpi|Mellanox Technologies MT27500 Family [ConnectX-3]]]