National University of Singapore

Gold GPU cluster

We are pleased to announce the release of gold, a new GPU cluster, complementing the HPC resources at Computer Centre.  The cluster has 16 nodes, with each node comprising a dual-socket INTEL hexa-core X5650 2.66GHz processor, 48 GB of memory and two NVDIA Tesla M2090 GPUs:

Number	Model	Processors	Memory
16	HP SL390 G7	Two INTEL X5650 @ 2.66 GHz, 6 cores	48 GB
		Two NVDIA Tesla M2090 @ 650 MHz, 512 GPU cores	6 GB

 

Please note the following:

1. Logging in

To access gold, you can either SSH directly to gold-c01 (172.25.192.196), or through HPC portal: after logging in to HPC portal, click gold-c01 cluster to start a xterm.

2. Batch jobs and queues

Only codes that use the GPU should be run on the gold cluster. The batch queues for gold cluster are goldgpu and goldgpu_test.  Codes for GPU written in CUDA may be compiled using the nvcc compiler (currently version 4.2):

prompt>  nvcc -o Helloworld.exe Helloworld.cu

To submit a GPU-enabled code:

prompt>  bsub –q goldgpu –o Helloworld.out  ./Helloworld.exe

 

Job description	Queue Name	Memory Limit	Runtime Limit
GPU-enabled	goldgpu	48 GB per node	Unlimited
GPU-enabled test	goldgpu_test	48 GB	100 minutes

 

The high performance parallel GPFS file system (/hpctmp) is accessible from the gold cluster. Users are advised to submit and run their batch jobs from this file system.

Please refer to the HPC GPU website for information on the GPU programming environment and GPU-enabled applications on gold cluster

For problems or queries, please contact us at .