Differences

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

Link to this comparison view

Both sides previous revision Previous revision
Next revision		 Previous revision
Next revisionBoth sides next revision
doku:vsc5quickstart [2023/02/17 17:45] – msiegeldoku:vsc5quickstart [2023/05/17 14:54] – msiegel
Line 1: Line 1:
 ====== Quick start guide for VSC-5 ====== ====== Quick start guide for VSC-5 ======
- 
-**Status: 2023/01** 
- 
-This page is under construction. 
  
 ===== Connecting ===== ===== Connecting =====
Line 23: Line 19:
  
  
-===== Loading Modules & Spack Environments =====+==== New SPACK without environments ====
  
-Different CPUs come with different compilers, so we use +Having worked with spack environments for some time, we have encountered several severe issues which have convinced us that we need to find a more practical way of maintaining software packages at VSC.
-the new spack feature ''environment'' to make sure to choose the right +
-package. +
  
-On login the default ''spack environment'' (zen3) is loaded +There are now three separate spack installation trees corresponding to the CPU/GPU architectures on VSC:
-automatically, so only modules that run on AMD processors are visible +
-with ''spack find''.+
  
-On VSC5 no default modules are loaded. Please do that by yourself +  * skylake - Intel CPUs; works on Intel Skylake and Cascadelake CPUs 
-using ''spack load <module>'' or ''module load <module>''. +  * zen - AMD CPUs; works on Zen 2 and 3 CPUs 
 +  * cuda-zen - AMD CPUs + NVIDIA GPUs; works on all nodes equipped with graphics cards
  
-Find the official SPACK documentation at https://spack.readthedocs.io/ +By default the spack installation tree suitable for the current compute/login node is activated and will be indicated by a **prefix** on the command line, e.g.:
- +
- +
-==== List Spack Environments ==== +
- +
-Type ''spack env list'' to see which environments are available and +
-which one is active. +
- +
-<code> +
-$ spack env list +
-==> 2 environments +
-    cascadelake  zen3 +
-</code> +
- +
-The current ''spack environment'' is also shown in your prompt: +
- +
-<code> +
-(zen3) [myname@l55 ~]# +
-</code> +
- +
-Mind that if your prompt is changed later, like when loading a ''python +
-environment'' using ''conda'', the correct ''spack environment'' might +
-not be shown correctly in your prompt. +
- +
-When a spack environment is activated, the command ''spack find -l'' lists those packages available for the active +
-environment. +
- +
-The command ''module avail'' will also show only those modules that are compatible with the active +
-spack environment. +
- +
- +
-==== Change Spack Environment ==== +
- +
-If you want to look for a certain package that belongs to another +
-architecture, first change the spack environment: +
- +
-<code> +
-$ spacktivate <myenv> +
-$ spacktivate cascadelake +
-</code> +
- +
-Only then ''spack find'' will show the modules for the active environment (e.g. ''cascadelake''). +
- +
- +
-==== Save Spack Environment ==== +
- +
-The following creates a load script for your current spack environment +
-with all loaded modules:+
  
 <code> <code>
-$ spack env loads -r +zen [user@l51 ~]$
 </code> </code>
  
-This creates a file called ''loads'' in the environment +Read more about SPACK at: 
-directory. Sourcing that file in bash will make the environment +  * [[doku:spack-transition | Transition to new SPACK without Environments]] 
-available to the user. The ''source loads'' command can be included in +  * [[doku:spack]] 
-''.bashrc'' files. The loads file may also be copied out of the +  * [[https://spack.readthedocs.io/en/latest/basic_usage.html|Official documentation of SPACK]]
-environment, renamed, etc.+
  
  
-==== Load a Module ====+==== Load a module ====
  
 Please always use spack, see [[doku:spack|SPACK - a package manager for Please always use spack, see [[doku:spack|SPACK - a package manager for
 HPC systems]]. HPC systems]].
  
 +Of course ''module'' commands will still work, like ''module avail xyz''. 
  
-===== Compile Code =====+ 
 +===== Compile code =====
  
 A program needs to be compiled on the hardware it will later run A program needs to be compiled on the hardware it will later run
Line 157: Line 104:
  
 ===== SLURM ===== ===== SLURM =====
-**For the exact partition/queue setup see [[doku:vsc5_queue|Queue | Partition setup on VSC-5]]** +For the partition/queue setup see [[doku:vsc5_queue|Queue | Partition setup on VSC-5]]. 
- +type ''sinfo -o %P'' to see the available partitions.
-type ''sinfo -o %P'' to see the available partitions: +
- +
-^partition ^nodes ^description^ +
-| zen2_0256_a40x2 | | AMD CPU nodes with 2x AMD Epyc (Milan) and 2x NIVIDA A40 and 256GB RAM | +
-| jupyter | | reserved for the jupyterhub | +
-| login5 | | login nodes, not an actual slurm partition | +
-| zen3_2048 | | AMD CPU nodes with 2x AMD Epyc (Milan) and 2TB RAM | +
-| zen3_1024 | | AMD CPU nodes with 2x AMD Epyc (Milan) and 1TB RAM | +
-| zen3_0512* | | The default partition. AMD CPU nodes with 2x AMD Epyc (Milan) and 512GB RAM | +
-| cascadelake_0384 | | Intel CPU nodes with 2x Intel Cascadelake and 384GB RAM | +
-| zen3_0512_a100x2 | | AMD CPU nodes with 2x AMD Epyc (Milan) and 2x NIVIDA A100 and 512GB RAM | +
- +
- +
-==== QoS ==== +
- +
-The following QoS are available for normal (=non private) projects: +
- +
-^QOS name ^ gives access to partition ^description^ +
-|zen3_0512 | zen3_0512 | default | +
-|zen3_1024 | zen3_1024 | +
-|zen3_2048 | zen3_2048 | +
-|cascadelake_0384 | cascadelake_0384 | +
-|zen2_0256_a40x2 | zen2_0256_a40x2 | +
-|zen3_0512_a100x2 | zen3_0512_a100x2 | +
-|zen3_0512_devel | 5 nodes on zen3_0512 | +
  
 ==== Submit a Job ==== ==== Submit a Job ====
  • doku/vsc5quickstart.txt
  • Last modified: 2023/05/17 15:28
  • by msiegel