remove more outdated SLURM access restrictions (#124)

bih-cluster/docs/how-to/connect/gpu-nodes.md

@@ -6,34 +6,24 @@ Connecting to a node with GPUs is easy.
 You simply request a GPU using the `--gres=gpu:$CARD:COUNT` (for `CARD=tesla` or `CARD=a40`) argument to `srun` and `batch`.
 This will automatically place your job in the `gpu` partition (which is where the GPU nodes live) and allocate a number of `COUNT` GPUs to your job.
 
-!!! note
-
-    Recently, `--gres=gpu:tesla:COUNT` was often not able to allocate the right partion on it's own.
-    If scheduling a GPU fails, consider additionally indicating the GPU partion explicitely with `--partition gpu` (or `#SBATCH --partition gpu` in batch file).
-
-!!! hint
+!!! info
+    Fair use rules apply.
+    As GPU nodes are a limited resource, excessive use by single users is prohibited and can lead to mitigating actions.
+    Be nice and cooperative with other users.
+    Tip: `getent passwd USER_NAME` will give you a user's contact details.
 
-    Make sure to read the FAQ entry "[I have problems connecting to the GPU node! What's wrong?](../../help/faq.md#i-have-problems-connecting-to-the-gpu-node-whats-wrong)".
-
-!!! important "Interactive Use of GPU Nodes is Discouraged"
+!!! warning "Interactive Use of GPU Nodes is Discouraged"
 
     While interactive computation on the GPU nodes is convenient, it makes it very easy to forget a job after your computation is complete and let it run idle.
     While your job is allocated, it blocks the **allocated** GPUs and other users cannot use them although you might not be actually using them.
     Please prefer batch jobs for your GPU jobs over interactive jobs.
 
-    Further, interactive GPU jobs are currently limited to 24 hours.
+    Furthermore, interactive GPU jobs are currently limited to 24 hours.
     We will monitor the situation and adjust that limit to optimize GPU usage and usability.
-
-!!! important "Allocation of GPUs through Slurm is mandatory"
 
-    In other word: using GPUs from SSH sessions is prohibited.
+    Please also note that allocation of GPUs through Slurm is mandatory, in other words: Using GPUs via SSH sessions is prohibited.
     The scheduler is not aware of manually allocated GPUs and this interferes with other users' jobs.
 
-## Prequisites
-
-You have to register with [[email protected]](mailto:[email protected]) for requesting access.
-Afterwards, you can connect to the GPU nodes as shown below.
-
 ## Preparation
 
 We will setup a miniconda installation with `pytorch` testing the GPU.
@@ -96,6 +86,8 @@ True
     Recently, `--gres=gpu:tesla:COUNT` was often not able to allocate the right partion on it's own.
     If scheduling a GPU fails, consider additionally indicating the GPU partion explicitely with `--partition gpu` (or `#SBATCH --partition gpu` in batch file).
 
+    Also make sure to read the FAQ entry "[I have problems connecting to the GPU node! What's wrong?](../../help/faq.md#i-have-problems-connecting-to-the-gpu-node-whats-wrong)" if you encounter problems.   
+
 ## Bonus #1: Who is using the GPUs?
 
 Use `squeue` to find out about currently queued jobs (the `egrep` only keeps the header and entries in the `gpu` partition).

bih-cluster/docs/how-to/connect/high-memory.md

@@ -1,11 +1,14 @@
 # How-To: Connect to High-Memory Nodes
 
-## Prequisites
-
-You have to register with [[email protected]](mailto:[email protected]) for requesting access.
-
-Afterwards, you can connect to the High-Memory using the `highmem` SLURM partition (see below).
-Jobs allocating more than 200GB of RAM should be routed automatically to the `highmem` nodes.
+The cluster has 4 high-memory nodes with 1.5 TB of RAM.
+You can connect to these nodes using the `highmem` SLURM partition (see below).
+Jobs allocating more than 200 GB of RAM are automatically routed to the `highmem` nodes.
+
+!!! info
+    Fair use rules apply.
+    As high-memory nodes are a limited resource, excessive use by single users is prohibited and can lead to mitigating actions.
+    Be nice and cooperative with other users.
+    Tip: `getent passwd USER_NAME` will give you a user's contact details.
 
 ## How-To
 

bih-cluster/docs/how-to/software/tensorflow.md

@@ -6,7 +6,7 @@ You can find the original TensorFlow installation instructions [here](https://ww
 This article describes how to set up TensorFlow with GPU support using Conda.
 This how-to assumes that you have just connected to a GPU node via `srun --mem=10g --partition=gpu --gres=gpu:tesla:1 --pty bash -i` (for Tesla V100 GPUs, for A400 GPUs use `--gres=gpu:a40:1`).
 Note that you will need to allocate "enough" memory, otherwise your python session will be `Killed` because of too little memory.
-You should read the [How-To: Connect to GPU Nodes](../../how-to/connect/gpu-nodes/) tutorial on an explanation of how to do this and to learn how to register for GPU usage.
+You should read the [How-To: Connect to GPU Nodes](../../how-to/connect/gpu-nodes/) tutorial on an explanation of how to do this.
 
 This tutorial assumes, that conda has been set up as described in [Software Management]((../../best-practice/software-installation-with-conda.md).