meta data for this page
  •  
Error loading plugin include
ParseError: syntax error, unexpected 'include' (T_INCLUDE), expecting identifier (T_STRING) or '{'
More info is available in the error log.

Differences

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

Link to this comparison view

Next revision		Previous revision
en:computing:cluster:fronty:start [2024/02/15 17:51] – initial translation David Celnýen:computing:cluster:fronty:start [2026/02/09 14:48] (current) – [Parameters for the ''srun'' and ''sbatch'' commands] Jan Pech
Line 65: Line 65:
 ===== Running the tasks ===== ===== Running the tasks =====
  
-Tasks can be run on multiple nodes, but always on one of the parts of the server:+Tasks can be run on multiple nodes:
  
-  * part **M** - kraken machines-m1 to m10 (all users) +  * nodes **L**, 4 pcs, 8 cores/node, kraken-l1 to l4 
-  * part **L** - kraken-l1 to l4 machines (limited access)+  * nodes **M**, 6 pcs, 32 cores/node, kraken-m1 to m6 
 +  * nodes **X**, 5 pcs, 96 cores/node, kraken-x1 to x5 
 + 
 +It is possible to run one job over multiple nodes using MPI, but it is usually faster to run one job on one node, due to the limited connection between nodes.
  
 Tasks can be run on: Tasks can be run on:
Line 81: Line 84:
   * <wrap hi> A job must always run under a queue (partition).</wrap> If no queue is specified, ''Mexpress'' is used. A list of defined queues is given below.   * <wrap hi> A job must always run under a queue (partition).</wrap> If no queue is specified, ''Mexpress'' is used. A list of defined queues is given below.
   * By specifying a queue, a run time limit is defined.    * By specifying a queue, a run time limit is defined. 
-  * Tasks in the express and short queues cannot be given a longer run time using ''-''''-time''. The default queue long time is set to 1 week, but they allow running up to 2 weeks, e.g. 9 days and 5 hours by specifying '''-p Llong''' ''-''''-time=9-05:00:0''.+  * Tasks in the express and short queues cannot be given a longer run time using ''-''''-time''. The default queue long time is set to 1 week, but they allow running up to 2 weeks, e.g. 9 days and 5 hours by specifying '''-p long''' ''-''''-time=9-05:00:0''.
   * Slurm will prioritize the task and user that the cluster uses less when queuing pending tasks. Therefore, it is not advantageous to declare a longer computation time than strictly necessary.   * Slurm will prioritize the task and user that the cluster uses less when queuing pending tasks. Therefore, it is not advantageous to declare a longer computation time than strictly necessary.
  
Line 88: Line 91:
  
 There are 6 queues ("partitions") on the Kraken cluster, divided by job run length (express, short, long) and cluster partition ("Mxxx" and "Lxxx"). If the user does not specify a queue with the ``-````-partition`` switch, the default value (Mexpress) is used: There are 6 queues ("partitions") on the Kraken cluster, divided by job run length (express, short, long) and cluster partition ("Mxxx" and "Lxxx"). If the user does not specify a queue with the ``-````-partition`` switch, the default value (Mexpress) is used:
-^  cluster part  ^  partition  ^  node   ^  time limit  ^ +^  partition  ^  node   ^  time limit  ^ 
-|  M (nodes kraken-m[1-10])  |  **Mexpress**  |  kraken-m[1-10]  |  6 hours  | +|  **Mexpress**  |  all  |  6 hours  | 
-|  :::  |  Mshort  |  kraken-m[1-10]  |  **2 days**  | +|  short  |  all  |  **2 days**  | 
-|  :::  |  :::  |  :::  |  3 days  | +|  :::  |  :::  |  3 days  | 
-|  :::  |  Mlong   |  kraken-m[3-6], kraken-m8  |  **1 week**  | +|  long   |  kraken-l4, m[5-6], x[2-4]  |  **1 week**  | 
-|  :::  |  :::  |  :::  |  2 weeks  | +|  :::  |  :::  |  2 weeks  | 
-|  L (nodes kraken-l[1-4])  |  Lexpress  |  kraken-l[1-4]  |  6 hours  | +|  NoCompute  |  kraken  |  **1 hour**  | 
-|  :::  |  Lshort  |  kraken-l[1-4]  |  2 days  | +|  :::  |  :::  |  8 hours  |
-|  :::  |  Llong  |  kraken-l[1-4]  |  **1 week**  | +
-|  :::  |  :::  |  :::  |  2 months (max)  | +
-|  admin node only  |  NoCompute  |  kraken  |  **1 hour**  | +
-|  :::  |  :::  |  :::  |  8 hours  |+
 *bold=default *bold=default
  
Line 125: Line 124:
 |  ``-t``, ``-````-time=<time>``  |  Set a limit on the total run time of the job allocation (days-hours:minutes:seconds)  |  ``-t 1:12``  | |  ``-t``, ``-````-time=<time>``  |  Set a limit on the total run time of the job allocation (days-hours:minutes:seconds)  |  ``-t 1:12``  |
 |  ``-w``, ``-````-nodelist=<node_name_list>``  |  Request a specific list of hosts  |  ``-w kraken-m2,kraken-m[5-6]``  | |  ``-w``, ``-````-nodelist=<node_name_list>``  |  Request a specific list of hosts  |  ``-w kraken-m2,kraken-m[5-6]``  |
-|  ``-x``, ``-````-exclude={<host1[,<host2>...]}``  |  Request that a specific list of hosts not be included in the resources allocated to this job  |  ``-````-exclude=kraken-m[7-9]``  |+|  ``-x``, ``-````-exclude={<host1[,<host2>...]}``  |  Request that a specific list of hosts not be included in the resources allocated to this job  |  ``-````-exclude=kraken-x[1-5]``  |