Differences

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

--- strutture:cnaf:cnaf_rd:infrastruttura:cluster_calcolo [2013/07/25 10:55] – caberletti@infn.it
+++ strutture:cnaf:cnaf_rd:infrastruttura:cluster_calcolo [2013/10/15 13:20] (current) – [Descrizione Software] caberletti@infn.it
@@ Line 1: / Line 1: @@
+====== Cluster Calcolo ======
+Il cluster R&D del CNAF si compone attualmente delle seguenti macchine:
+  * [[cluster_calcolo#rd-ui|rd-ui]]: VM per il login degli utenti e per la sottomissione dei job
+  * [[cluster_calcolo#rd-lsfmaster|rd-lsfmaster]]: VM con il master daemon e scheduler del batch system LSF
+  * [[cluster_calcolo#rd-slurm|rd-slurm]]: VM con il master daemon e scheduler del batch system SLURM
+  * [[cluster_calcolo#rd-coka-01|rd-coka-01]]: server con 1 MIC e 2 GPU K20
+  * [[cluster_calcolo#rd-coka-02|rd-coka-02]]: server con 2 MIC
+  * [[cluster_calcolo#rd-gpu-02|rd-gpu-02]]: server con 2 GPU C2050
+  * [[cluster_calcolo#carma-devkit|carma-devkit]]: host con processore ARM e GPU nVidia
+  * [[cluster_calcolo#rd-arm-compiler|rd-arm-compiler]]: host per la cross compilazione
+\\
+{{:strutture:cnaf:cnaf_rd:infrastruttura:map.png}}
+\\
+==== Descrizione Hardware ====
+^ Name         ^ Chassis ^ MotherBoard         ^ Cores                               ^ Memory         ^ GPU/Co-processor                       ^ OS           ^
+| rd-coka-01   | E4      | Supermicro X9DRG-HF | SandyBridge 2 x 12 cores @ 2.00 GHz | 32GB @ 1333MHz | 2 x nVidia K20m + Intel Xeon Phi 3110P | CentOS 6.4   |
+| rd-coka-02   | E4      | Supermicro X9DRG-HF | SandyBridge 2 x 12 cores @ 2.00 GHz | 32GB @ 1333MHz | 2 x Intel Xeon Phi 5110P               | SL 6.4       |
+| rd-gpu-02    | E4      | Supermicro X8DTG-QF | Westmere 2 x 8 cores @ 2.40 GHz     | 24GB @ 1066MHz | 3 x nVidia C2050                       | CentOS 6.4   |
+| carma-devkit | Seco    | cardhu              | ARM A9 1 x 4 cores @ 1.4GHz         | 2GB DDR3       | nVidia Quadro 1000M                    | Ubuntu 12.04 |
+\\
+== Dettagli Hardware ==
+^    GPU   K20                          ^^
+| Modello           | NVIDIA Tesla K20m |
+| Cores             | 13 (Multiprocessors) x 192 (CUDA cores) = 2496 CUDA cores |
+| GPU clock rate    | 706 MHz |
+| Memory            | 5 GB GDDR5 |
+| Memory clock rate | 2600 MHz |
+| Memory bandwidth  | 208 GB/sec |
+| Max power cons.   | 225 Watt |
+| Peak SP perf.     | 3.52 Tflops |
+| Peak DP perf.     | 1.17 Tflops |
+^    GPU   C2050                          ^^
+| Modello           | NVIDIA Tesla C2050 |
+| Cores             | 14 (Multiprocessors) x 32 (CUDA cores) = 448 CUDA cores |
+| GPU clock rate    | 1147 MHz |
+| Memory            | 3 GB GDDR5 |
+| Memory clock rate | 1500 MHz |
+| Memory bandwidth  | 144 GB/sec |
+| Max power cons.   | 238 Watt |
+| Peak SP perf.     | 1.03 Tflops |
+| Peak DP perf.     | 515 Gflops |
+^    GPU  Quadro 1000M                   ^^
+| Modello           | NVIDIA Quadro 1000M |
+| Cores             | 2 (Multiprocessors) x 48 (CUDA cores) = 96 CUDA cores |
+| GPU clock rate    | 1400 MHz |
+| Memory            | 2048 MBytes |
+| Memory clock rate | 900 MHz |
+| Memory bandwidth  | 28.8 GB/sec |
+| Max power cons.   | 45 Watt |
+| Peak SP perf.     | 268.8 Gflops |
+| Peak DP perf.     | ... Tflops |
+^    MIC 3110P                          ^^
+| Modello           | Intel Xeon Phi 3110P |
+| Cores             | 57 (Multiprocessors) x 4 (hardware threads) = 228 cores |
+| CPU clock rate    | 1100 MHz |
+| Memory            | 3 GBytes |
+| Memory clock rate | 2500 MHz |
+| Memory bandwidth  | 240 GB/s |
+| Max power cons.   | 225 Watt |
+| Peak SP perf.     | ... Tflops |
+| Peak DP perf.     | 1.0 Tflops |
+^    MIC 5110P                          ^^
+| Modello           | Intel Xeon Phi 5110P |
+| Cores             | 60 (Multiprocessors) x 4 (hardware threads) = 240 cores |
+| CPU clock rate    | 1053 MHz |
+| Memory            | 8 GBytes |
+| Memory clock rate | 2500 MHz |
+| Memory bandwidth  | 320 GB/s |
+| Max power cons.   | 225 Watt |
+| Peak SP perf.     | ... Tflops |
+| Peak DP perf.     | 1.01 Tflops |
+==== Descrizione Software ====
+== GPU ==
+Per utilizzare le GPU sono disponibili:
+  * CUDA 5.0
+  * CUDA 5.5
+  * Intel OpenCL 1.2-3.0.67279
+== Intel Xeon Phi ==
+Per l'uso dei co-processori Phi sono disponibili:
+  * Intel Parallel Studio XE 2013 (icc 13.1.1)
+  * Intel Parallel Studio XE 2013 SP1 (icc 14.0.0)
+== MPI ==
+  * Intel MPI
+  * MVAPICH 2
+  * OpenMPI