Commit 02106f44 authored by Michael Krause's avatar Michael Krause 🎉
Browse files

Merge branch 'master' of gitlab.mpib-berlin.mpg.de:lip-support/tardis-doc

parents 277bef1d 87034b99
#!/bin/bash
set -e
SRC=$1
DST=$2
cd $1
make html
cd -
mkdir -p $2
cp -r $1/_build/html/. $2
exit 0
......@@ -19,11 +19,11 @@ Tardis
:width: 60%
Some technical facts:
+ **608** Intel® Xeon® CPU E5-2670 CPU cores inside 32 Dell m620 blade servers
+ **R**:sub:`max` = 5.1 TFlops, **R**:sub:`peak` = 12.6 TFlops
+ **3.2TB** total amount of memory
+ **256 GB** RAM on 6 of those nodes
+ **32 TB** of attached NFS storage
+ **712** Intel® Xeon® CPU E5-2670 CPU cores(no HT) inside 42 Dell m6x0 blade servers
+ **R**:sub:`max` = 9.9 TFlops, **R**:sub:`peak` = 14.8 TFlops
+ **6.4TB** total amount of memory
+ **32TB** of attached NFS storage
+ **4.1TB** experimental glusterfs storage
+ simple 2-level fat tree network with a theoretical bisection bandwidth of **40GB/s**
......
......@@ -48,7 +48,7 @@ in Nautilus:
Open the following URL in your file manager (CTRL-L) ::
ssh://<YOUR_USERNAME>@tardis.mpib-berlin.mpg.de/home/mpib/``
ssh://YOUR_USERNAME@tardis.mpib-berlin.mpg.de/home/mpib/
Mac OSX
^^^^^^^
......
......@@ -14,3 +14,4 @@ while true ; do
WATCHSOURCE="**/*.rst *.rst"
$MAKEFUNC
done
FSL
===
We are automatically following the FSL 5 release history of `Neurodebian <http://neuro.debian.net>`_.
To use FSL in your jobs, activate it like this:
.. code-block:: bash
source /etc/fsl/5.0/fsl.sh
Afterwards the :file:`PATH` variable should be set correctly.
FIX
---
We have set up a running version of [FSL
Fix](http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FIX). Currently it is not using
FSL's internal qsub so it only makes sense to parallelize over a number of
subjects. Example:
We have set up a running version of `FSL Fix <http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FIX>`_.
Currently it is not using FSL's internal qsub so it only makes sense to
parallelize over a number of subjects. Example:
.. code-block:: bash
......
......@@ -105,8 +105,23 @@ Timings:
user system elapsed
107.442 0.052 7.20
Alternatively you can use `%dopar%` and `foreach`:
If you want to unroll the vector operation of the whole cluster, have a look at
.. code-block:: r
library(doMC)
library(gmp)
cores <- Sys.getenv("PBS_NUM_PPN")
cores <- if (cores == '') 1 else as.numeric(cores)
registerDoMC(cores)
foreach(i=seq(1,16)) %dopar% {
factorize(matrix(as.bigz(2**521)-1))
}
If you want to unroll the vector operation to the whole cluster, have a look at
the `batchjobs`_ package.
.. _batchjobs: https://cran.r-project.org/web/packages/BatchJobs/index.html
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment