HECToR User Guide
This guide explains how to make use of the HECToR service. It is designed with the needs of new users in mind, but will also be a useful reference for everyone.
Contents
1. Introduction
An introduction to the HECToR facility.
2. System hardware
Details of the hardware and architecture that make up the HECToR facility.
3. Connecting
How to access the HECToR facilities and transfer data to and from HECToR.
4. Resource Management
Information on using budgets (for CPU time), the HECToR filesystems and managing your resources using the HECToR SAFE.
- 4.1 The HECToR administration web site (SAFE)
- 4.2 Checking your CPU-time (kAU) quota
- 4.3 HECToR Filesystems
- 4.4 Disk quotas
- 4.4.1 Checking disk quotas
- 4.4.2 Quotas on the Lustre filesystem
- 4.4.3 Allocating quotas to specific groups and users
- 4.5 File permissions and security
- 4.6 Sharing data with other HECToR users
- 4.7 Sharing data between systems
- 4.8 File IO Performance Guidelines
- 4.9 Data archiving
- 4.10 Use of /tmp
- 4.11 Backup policies
5. Modules
Working with the module environment that manages software packages and libraries on HECToR.
- 5.1 Information on the available modules
- 5.2 Loading, unloading and swapping modules
- 5.3 Things to watch out for
6. Running Jobs
How to run jobs on HECToR facilities and more information on the batch queuing system.
- 6.1 Using PBSpro
- 6.2 Output from PBS jobs
- 6.3 bolt: Job submission script creation tool
- 6.4 Running Parallel Jobs
- 6.4.1 Parallel job launcher aprun
- 6.4.2 Task affinity for "unpacked" jobs
- 6.4.3 Job submission scripts for parallel jobs using MPI
- 6.4 4 Job submission scripts for parallel jobs using OpenMP
- 6.4.5 Job submission scripts for hybrid parallel jobs using MPI and OpenMP (Mixed Mode)
- 6.4.6 Running Coarray Fortran Jobs
- 6.4.7 Running Chapel Jobs
- 6.4.8 Interactive Jobs
- 6.5 checkScript job submission script validation tool
- 6.6 OOM (Out of Memory) Error Messages
- 6.7 Automatic re-running of jobs after system restart
- 6.8 Batch System Layouts and Limits
- 6.9 Reservations
- 6.10 Developing/debugging codes in the batch system - Interactive jobs
- 6.11 Serial Queues
7. Compiling
Information on compiling programs to run on HECToR facilities.
- 7.1 Cross-compiling for Compute Nodes
- 7.1.1 Compiling Fortran code
- 7.1.2 Compiling C code
- 7.1.3 Compiling C++ code
- 7.1.4 Selecting compiler suites and compiler versions
- 7.1.5 Compiling MPI code
- 7.1.6 Compiling OpenMP code
- 7.1.7 Compiling mixed(OpenMP/MPI) code
- 7.1.8 Compiling Coarray Fortran code
- 7.1.9 Useful compiler options
- 7.1.10 Using dynamic linking/libraries
- 7.2 Compiling for the frontend nodes
8. Libraries
The software libraries available on HECToR facilities. This includes high-performance mathematical libraries and others.
9. Tools
Performance analysis, debugging and visualisation tools available on HECToR facilities.
- 9.1 Cray Performance Analysis Tool (CrayPAT)
- 9.2 Cray Apprentice2
- 9.3 Totalview debugger
- 9.4 Paraview
- 9.5 VampirTrace
- 9.6 DDT Debugger
10. Software on HECToR
Information on the centrally supported scientific software packages on HECToR facilities.
11. Grid computing
Using grid computing tools on HECToR facilities.
12. Further reading
Links to other useful documents.