This shows you the differences between two versions of the page.

Link to this comparison view

io_odds_and_ends [2017/11/29 19:35]
— (current)
Line 1: Line 1:
-====== Input and Output on HPC Systems ====== 
-This page comments on general aspects related to workloads on Mogon I/II, more information with respect to filesystems can be found [[filesystems|here]]. 
-===== Issues which may arise ===== 
-Scientific applications perform I/O to parallel file system in primarily one of two ways: 
-  * Shared-­file (N-­to-­1): A single file is created, and all application tasks write to that file (usually to completely disjoint regions) 
-    * This increases usability: There is only one file to keep track of by the application 
-    * It may create lock contention and hinder performance  
-  * File-­per‐process (N‐to‐N): Each instance of an application / each task creates or reads a separate file and writes to that file, only.     
-    * It may avoid lock contention on the application level, but increases the risk of file system stress when to writing to one destination, thereby triggering locking on the file system level  
-    * It is impossible to restart these applications with a different number of tasks 
-===== How may/can I - as a user - analyze issues? ===== 
-Currently, when suspecting I/O problems you should [[|address the HPC-team]]. There is no straight forward method available to analyze I/O problems on the user level (of third party applications). 
-<WRAP center round todo 90%> 
-We may provide more tools in the foreseeable future.  
-===== Which solution may solve which issue? ===== 
-The statements above may seem a little abstract, particularly when third-party applications have to be used and no decision can be made about the application architecture. 
-However, a few rules of thumb can be given: 
-  * [[node_local_scheduling|Pooling short jobs]] is generally a good idea with respect to scheduling and organizing your work flow. If this involves reading identical input files by all those application instances, the stage-in to a [[slurm_localscratch|node local scratch]] or even into [[ramdisk|RAM]] may solve performance issues: By creating a temporary input resource, the need to keep track of file accesses for these particular files for the global file system is dropped. 
-  * Avoid keeping open file handles by many processes (within a directory). Violating this rule may cause delays, because the global file system needs to coordinate every writing process. A possible solution is to write into the job directory (see [[slurm_localscratch|node local scratch]]) and to copy this output to the global file system after a writing process is finished / releases a file. 
-  * Avoid writing too many small files: The overhead in keeping track of the meta information for millions of small files can be bigger than the file size. The global file system is not optimized for this.  
  • io_odds_and_ends.1511980505.txt.gz
  • Last modified: 2017/11/29 19:35
  • (external edit)