SOSCIP GPU
| SOSCIP GPU | |
|---|---|
| Installed | September 2017 |
| Operating System | Ubuntu 16.04 le |
| Number of Nodes | 14x Power 8 with 4x NVIDIA P100 |
| Interconnect | Infiniband EDR |
| Ram/Node | 512 GB |
| Cores/Node | 2 x 10core (20 physical, 160 SMT) |
| Login/Devel Node | sgc01 |
| Vendor Compilers | xlc/xlf, nvcc |
SOSCIP
The SOSCIP GPU Cluster is a Southern Ontario Smart Computing Innovation Platform (SOSCIP) resource located at theUniversity of Toronto's SciNet HPC facility. The SOSCIP multi-university/industry consortium is funded by the Ontario Government and the Federal Economic Development Agency for Southern Ontario [1].
Support Email
Please use <soscip-support@scinet.utoronto.ca> for SOSCIP GPU specific inquiries.
Specifications
The SOSCIP GPU Cluster consists of of 14 IBM Power 822LC "Minsky" Servers each with 2x10core 3.25GHz Power8 CPUs and 512GB Ram. Similar to Power 7, the Power 8 utilizes Simultaneous MultiThreading (SMT), but extends the design to 8 threads per core allowing the 20 physical cores to support up to 160 threads. Each node has 4x NVIDIA Tesla P100 GPUs each with 16GB of RAM with CUDA Capability 6.0 (Pascal) connected using NVlink.
Compile/Devel/Test
Access is provided through the BGQ login node, bgqdev.scinet.utoronto.ca using ssh, and from there you can proceed to the GPU development node sgc01-ib0.
Filesystem
The filesystem is shared with the BGQ system. See here for details.
Job Submission
The SOSCIP GPU cluster uses SLURM as a job scheduler and jobs are scheduled by node, ie 20 cores and 4 GPUs each. Jobs are submitted from the development node sgc01. The maximum walltime per job is 12 hours (except in the 'long' queue, see below) with up to 8 nodes.
$ sbatch myjob.script
Where myjob.script is
#!/bin/bash #SBATCH --nodes=1 #SBATCH --ntasks=20 # MPI tasks (needed for srun) #SBATCH --time=00:10:00 # H:M:S #SBATCH --gres=gpu:4 # Ask for 4 GPUs per node cd $SLURM_SUBMIT_DIR hostname nvidia-smi
You can queury job information using
squeue
To cancel a job use
scancel $JOBID
Longer jobs
If your job takes more than 12 hours, the sbatch command will not let you submit your job. There is, however, a way to have jobs up to 24 hours long, by specifying "-p long" as an option (i.e., add #SBATCH -p long to your job script). The priority of such jobs may be throttled in the future if we see that the 'long' queue is having a negative efffect on turnover time in the queue.
Interactive
For an interactive session use
salloc --gres=gpu:4
Automatic Re-submission and Job Dependencies
Commonly you may have a job that you know will take longer to run than what is permissible in the queue. As long as your program contains checkpoint or restart capability, you can have one job automatically submit the next. In the following example it is assumed that the program finishes before the time limit requested and then resubmits itself by logging into the development nodes. Job dependencies and a maximum number of job re-submissions are used to ensure sequential operation.
#!/bin/bash
#SBATCH --nodes=1
#SBATCH --ntasks=20 # MPI tasks (needed for srun)
#SBATCH --time=00:10:00 # H:M:S
#SBATCH --gres=gpu:4 # Ask for 4 GPUs per node
cd $SLURM_SUBMIT_DIR
: ${job_number:="1"} # set job_nubmer to 1 if it is undefined
job_number_max=3
echo "hi from ${SLURM_JOB_ID}"
#RUN JOB HERE
# SUBMIT NEXT JOB
if [[ ${job_number} -lt ${job_number_max} ]]
then
(( job_number++ ))
next_jobid=$(ssh sgc01-ib0 "cd $SLURM_SUBMIT_DIR; /opt/slurm/bin/sbatch --export=job_number=${job_number} -d afterok:${SLURM_JOB_ID} thisscript.sh | awk '{print $4}'")
echo "submitted ${next_jobid}"
fi
sleep 15
echo "${SLURM_JOB_ID} done"
Software Installed
IBM PowerAI
The PowerAI platform contains popular open machine learning frameworks such as Caffe, Tensorflow, and Torch. Run the module avail command for a complete listing. More information is available at this link: https://developer.ibm.com/linuxonpower/deep-learning-powerai/releases/. Release 4.0 is currently installed.
GNU Compilers
More recent versions of the GNU Compiler Collection (C/C++/Fortran) are provided in the IBM Advanced Toolchain with enhancements for the POWER8 CPU. To load the newer advance toolchain version use:
Advanced Toolchain V10.0
module load gcc/6.3.1
Advanced Toolchain V11.0
module load gcc/7.2.1
More information about the IBM Advanced Toolchain can be found here: https://developer.ibm.com/linuxonpower/advance-toolchain/
IBM XL Compilers
To load the native IBM xlc/xlc++ and xlf (Fortran) compilers, run
module load xlc/13.1.5 module load xlf/15.1.5
Information about the IBM XL Compilers can be found at the following links:
NVIDIA GPU Driver Version
The current NVIDIA driver version is 384.66
CUDA
The current installed CUDA Tookits is are version 8.0 and version 9.0
module load cuda/8.0
or
module load cuda/9.0
The CUDA driver is installed locally, however the CUDA Toolkit is installed in:
/usr/local/cuda-8.0 /usr/local/cuda-9.0
Note that the /usr/local/cuda directory is linked to the /usr/local/cuda-9.0 directory.
OpenMPI
Currently OpenMPI has been setup on the 14 nodes connected over EDR Infiniband.
$ module load openmpi/2.1.1-gcc-5.4.0 $ module load openmpi/2.1.1-XL-13_15.1.5