You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 76 Next »

Policies

LCLS users are responsible for complying with the data management and curation policies of their home institutions and funding agents and authorities. To enhance the scientific productivity of the LCLS user community, LCLS supplies on-site disk, tape and compute resources for prompt analysis of LCLS data, and software to access those resources consistent with the published data retention policy. Compute resources are preferentially allocated to recent and running experiments.

Getting an Account

You will need a valid SLAC UNIX account in order to use the LCLS computing system. The instructions for getting a SLAC UNIX account are here.

Getting Access to the System

You can get into the LCLS photon computing system by ssh'ing to:

pslogin.slac.stanford.edu

From the pslogin pool you can then reach the analysis nodes (see below).

Each control room has a number of nodes for local login. These nodes have access to the Internet and are named psusr<id>.

The controls and DAQ nodes used for operating an instrument work in kiosk mode so you don't need a personal account to run an experiment from the control room. Remote access to these nodes is not allowed for normal users.

Data Management

LCLS provides space for all your experiment's data at no cost for you. This includes the raw data from the detectors as well as the data derived from your analysis. Your raw data are available as XTC files or, on demand, as HDF5 files. The path to the experimental data is:

/reg/d/psdm/<instrument>/<experiment>

The LCLS data policy is described here. The tools for managing files are described here.

Data Export

You can use the psexport pool for copying your data:

psexport.slac.stanford.edu

The recommended tools for exporting the data offsite are bbcp and Globus Online. The former, bbcp, is slightly simpler to setup. On the other hand Globus Online is more efficient when transferring large amount of data because it babysits the overall process by, for example, automatically restarting a failed or stalled transfer. The performance of the two tools is very similar.

All control rooms and the overflow room in FEH have one or more taps on the Visitor Data Network. These taps can be used to transfer data to a laptop or a storage device. These devices will be automatically assigned an IP address through DHCP as soon as they are connected to the network tap.

Running the Analysis

The analysis framework is documented in the Data Analysis page. This section describes the resources available for running the analysis. 

The following figure shows a logic diagram of the LCLS data flow and indicates the different stages where data analysis can be performed in LCLS: 

  1. Online monitoring nodes - These nodes get detector data over the network and place it in shared memory on each node.  The detector data are received over the network by snooping the multicast traffic between the DAQ readout nodes and the data cache layer. There is a set of monitoring nodes allocated for each instrument.
  2. Fast feedback nodes - These nodes read the detector data from dedicated disks that each experimental hutch has. It is possible to read the data in quasi-real time as they are written to the fast feedback (FFB) storage layer by the DAQ without waiting for the run to end. There is a farm of FFB nodes for each experimental hall.
  3. Offline nodes - These nodes read the detector data from disk. These nodes include both interactive nodes (psana pool) and batch farms (psanaq queue, see below).  These nodes are the main resource for doing data analysis and are shared across all LCLS instruments.

 

 

Interactive Pool

In order to get access to the interactive nodes, connect to psana:

ssh psana

The psana pool is currently made of 5 servers with the following specifications:

  • 12-cores, Xeon E5-2620 v3 2.4GHz, 128GB memory, 10Gb/s, no IB

LCLS provides a few Matlab licenses and one IDL license. Instructions on how to use these tools are in the Matlab and IDL pages.

Two of the psana interactive nodes (psanagpu101 and 102) have GPUs. The other three (psanaphi101, 102 and 103) have Xeon Phi coprocessors.

Batch Farms

There are a number of batch farms (i.e. collections of compute nodes) located in the central computing building and in the NEH and FEH experimental halls. Instructions describing how to submit jobs can be found on the Submitting Batch Job page.

The main batch farm currently consist of 80 nodes with the following general specifications:

  • 12 cores, Xeon X5675, 24GB memory, QDR IB

In addition, each experimental hall provides 20 nodes with access to the fast feedback storage layer (FFB). These nodes provide the resources for the high priority queues for the running experiments and have the following specifications:

  • 16 cores, Xeon E5-2640 v3, 128GB memory, QDR IB

Remote Visualization

When performing analysis on the psana interactive nodes, it is useful to display plots on your host machine. For host machines near SLAC, using ssh with X-windows forwarding (the -X or -Y options) suffices. X  windows forwarding can get slow for host machines in Europe. Some users have found better performance with technology called nomachine, this is documented on the Remote Visualization page.

Prompt Analysis

There are various ways to get real time information about the data acquired by an LCLS experiment. Users should be aware of the different possibilities and choose the approach that works best for their experiment. The methods for doing prompt analysis are described here.

  • No labels