Difference between revisions of "P8"

From SciNet Users Documentation
Jump to: navigation, search
Line 18: Line 18:
 
== Compile/Devel/Test ==
 
== Compile/Devel/Test ==
  
First login via ssh with your CC/scinet account at '''<tt>niagara.scinet.utoronto.ca</tt>''', see [https://docs.scinet.utoronto.ca/index.php/Niagara_Quickstart#Using_Niagara:_Logging_in] and from there you can ssh to <tt>p8t01</tt> or <tt>p8t02</tt> for the K80 GPUs and to <tt>p8t03</tt> or <tt>p8t04</tt> for the Pascal GPUs.
+
Access through the Niagara login nodes '''<tt>niagara.scinet.utoronto.ca</tt>''' using your [https://docs.scinet.utoronto.ca/index.php/Niagara_Quickstart#Using_Niagara:_Logging_in | CC/SciNet account] and from there you can ssh to <tt>p8t01</tt> or <tt>p8t02</tt> for the K80 GPUs and to <tt>p8t03</tt> or <tt>p8t04</tt> for the Pascal GPUs.
  
 
== Software for  ==
 
== Software for  ==

Revision as of 16:13, 9 July 2018

P8
P8 s822.jpg
Installed June 2016
Operating System Linux RHEL 7.2 le / Ubuntu 16.04 le
Number of Nodes 2x Power8 with 2x NVIDIA K80, 2x Power 8 with 4x NVIDIA P100
Interconnect Infiniband EDR
Ram/Node 512 GB
Cores/Node 2 x 8core (16 physical, 128 SMT)
Login/Devel Node p8t0[1-2] / p8t0[3-4]
Vendor Compilers xlc/xlf, nvcc

Specifications

The P8 Test System consists of of 4 IBM Power 822LC Servers each with 2x8core 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 16 physical cores to support up to 128 threads. 2 nodes have two NVIDIA Tesla K80 GPUs with CUDA Capability 3.7 (Kepler), consisting of 2xGK210 GPUs each with 12 GB of RAM connected using PCI-E, and 2 others have 4x NVIDIA Tesla P100 GPUs each wit h 16GB of RAM with CUDA Capability 6.0 (Pascal) connected using NVlink.

Compile/Devel/Test

Access through the Niagara login nodes niagara.scinet.utoronto.ca using your | CC/SciNet account and from there you can ssh to p8t01 or p8t02 for the K80 GPUs and to p8t03 or p8t04 for the Pascal GPUs.

Software for

GNU Compilers

To load the newer advance toolchain version use:

For p8t0[1-2]

module load gcc/5.3.1

For p8t0[3-4]

module load gcc/6.4.1

IBM Compilers

To load the native IBM xlc/xlc++ compilers

For p8t0[1-2]

module load xlc/13.1.4
module load xlf/13.1.4

For p8t0[3-4]

module load xlc/13.1.5_b2
module load xlf/13.1.5_b2

Driver Version

The current NVIDIA driver version is 361.93 for p8t0[1-2], 396.26 for p8t0[3-4]

CUDA

The current installed CUDA Tookit is 8.0 for p8t0[1-2], 9.2 for p8t[3-4].

module load cuda/8.0
OR
module load cuda/9.2

The CUDA driver is installed locally, however the CUDA Toolkit is installed in:

/usr/local/cuda-8.0
OR
/usr/local/cuda-9.2

OpenMPI

Currently OpenMPI has been setup on the four nodes connected over QDR Infiniband.

For p8t0[1-2]

$ module load openmpi/1.10.3-gcc-5.3.1
$ module load openmpi/1.10.3-XL-13_15.1.4

For p8t0[3-4]

$ module load openmpi/1.10.3-gcc-6.2.1
$ module load openmpi/1.10.3-XL-13_15.1.5

PE

IBM's Parallel Environment (PE), is available for use with XL compilers using the following

$ module pe/xl.perf
mpiexec -n 4 ./a.out

documentation is here