en:hpc
Differences
This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revision | |||
| en:hpc [2026/11/19 14:20] – eduardas | en:hpc [2026/11/19 14:21] (current) – old revision restored (2026/10/02 12:41) eduardas | ||
|---|---|---|---|
| Line 1: | Line 1: | ||
| - | ====== | + | ====== |
| - | This guide walks you through the basic HPC workflow: | + | A High Performance Computing (HPC) is a specially designed network of computers capable of running applications that can exchange data efficiently. |
| - | - Connect to the HPC system | + | VU MIF HPC consists of a supercomputer from the clusters (the first number is the actual |
| - | - Understand | + | |
| - | - Create and submit your first job | + | |
| - | - Monitor job execution | + | |
| - | - View job results | + | |
| - | - Request CPU and memory resources | + | |
| - | - Learn how project allocations work | + | |
| - | - Run a multi-core Python application | + | |
| - | - Verify CPU utilization | + | |
| - | - Submit a GPU job | + | |
| - | If you can complete this guide, you are ready to start running your own workloads on the HPC cluster. | + | ^Title ^Nodes ^CPU ^GPU ^RAM ^HDD ^Network ^Notes| |
| + | ^main ^35/ | ||
| + | ^gpu | ||
| + | ^power | ||
| - | ===== Before You Start ===== | + | Total **40/ |
| - | Make sure that: | + | The processor below = CPU = core - a single core of the processor (with all hyperthreads if they are turned on). |
| - | * You have completed [[hpc: | + | ====== Software ====== |
| - | * You have an active HPC account | + | |
| - | * You have an SSH client installed | + | |
| - | * You can connect to the HPC login node | + | |
| - | Need access first? | + | In **main** and **gpu** partitions there are installed [[https:// |
| - | | + | You can check the list of OS package with the command '' |
| - | ===== Step 1: Connect to the Cluster | + | ===== Singularity |
| - | Open a terminal and connect to the HPC login node: | + | With the command [[https:// |
| - | < | + | You can prepare your container with singularity, |
| - | ssh username@hpc.mif.vu.lt | + | < |
| + | $ singularity build --sandbox /tmp/python docker:// | ||
| + | $ singularity exec -w /tmp/python pip install package | ||
| + | $ singularity build python.sif / | ||
| + | $ rm -rf /tmp/python | ||
| </ | </ | ||
| - | Replace '' | ||
| - | Verify that you are connected: | + | To use this container, it is advised to use different directory instead of home directory (for python packages don't mixup with installed to home directory). |
| - | + | < | |
| - | < | + | $ mkdir ~/workdir |
| - | hostname | + | $ singularity exec -H ~/ |
| - | pwd | + | |
| </ | </ | ||
| - | You should see the hostname of the login node and your home directory path. | + | Similarly, you can use R, Julia or other containers that do not require root privileges to install packages. |
| - | <note important> | + | If you want to add OS packages |
| - | Your SLURM account is created automatically after your first successful login. This process may take up to 5 minutes. | + | <code shell> |
| - | + | $ singularity build --sandbox /tmp/python docker:// | |
| - | Until the account is created, job submission may fail with account-related errors. | + | $ cp / |
| - | </note> | + | $ fakeroot -l / |
| - | + | $ fakeroot -l / | |
| - | For detailed instructions see: | + | $ fakeroot -l / |
| - | + | $ rm -rf / | |
| - | * [[hpc: | + | $ singularity build python.sif |
| - | + | $ rm -rf /tmp/python | |
| - | ===== Step 2: Understanding Login Nodes and Compute Nodes ===== | + | |
| - | + | ||
| - | Unlike a personal computer, computational workloads are not normally executed directly from the login node. Instead, jobs are submitted to SLURM, which schedules | + | |
| - | + | ||
| - | The login node is intended | + | |
| - | + | ||
| - | * Editing files | + | |
| - | * Compiling | + | |
| - | * Managing data | + | |
| - | * Preparing and submitting jobs | + | |
| - | + | ||
| - | The compute nodes are used for: | + | |
| - | + | ||
| - | * CPU-intensive calculations | + | |
| - | * Simulations | + | |
| - | * Data analysis | + | |
| - | * Machine learning and AI workloads | + | |
| - | + | ||
| - | <note important> | + | |
| - | Do not run computationally intensive applications directly on the login node. Submit them through SLURM so they can be executed on the appropriate compute resources. | + | |
| - | </note> | + | |
| - | + | ||
| - | ===== Step 3: Create a Working Directory ===== | + | |
| - | + | ||
| - | Create a directory for this tutorial: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | mkdir getting_started | + | |
| - | cd getting_started | + | |
| </ | </ | ||
| - | Verify your location: | + | ===== Hadoop ===== |
| - | <code bash> | + | There are ready-made scripts to run your **hadoop** tasks using the [[https:// |
| - | pwd | + | |
| - | </code> | + | |
| - | ===== Step 4: Create Your First Job Script | + | ===== JupyterHub |
| - | Create | + | With [[https:// |
| - | < | + | < |
| - | nano hello.sh | + | $ python3.7 -m pip install --upgrade pip setuptools wheel |
| + | $ python3.7 -m pip install --ignore-installed batchspawner jupyterlab | ||
| </ | </ | ||
| - | Paste the following content: | + | Alternatively, |
| - | + | < | |
| - | < | + | #!/bin/sh |
| - | #!/bin/bash | + | exec singularity exec --nv myjupyterlab.sif batchspawner-singleuser |
| - | # | + | |
| - | #SBATCH --partition=main | + | |
| - | #SBATCH --output=hello.out | + | |
| - | # | + | |
| - | + | ||
| - | echo "Hello from HPC!" | + | |
| - | hostname | + | |
| - | date | + | |
| </ | </ | ||
| - | Save the file and exit the editor. | + | ====== Registration ====== |
| - | ===== Step 5: Submit the Job ===== | + | * **For VU MIF network users** - HPC can be used without additional registration if the available resources are enough (monthly limit - **2000 CPU-h and 100 GPU-h**). Once this limit has been reached, you can request more by filling in [[https:// |
| - | Submit | + | * **For users of the VU computer network** - you must fill in the [[https:// |
| - | <code bash> | + | * **For other users (non-members of the VU community)** - you must fill in the [[https:// |
| - | sbatch hello.sh | + | |
| - | </code> | + | |
| - | Example output: | + | ====== Connection ====== |
| - | < | + | You need to use SSH applications (ssh, putty, winscp, mobaxterm) and Kerberos or SSH key authentication to connect to **HPC**. |
| - | Submitted batch job 123456 | + | |
| - | </ | + | |
| - | The returned number | + | If **Kerberos** |
| - | ===== Step 6: Monitor | + | * Log in to the Linux environment in a VU MIF classroom or public terminal with your VU MIF username and password or login to **uosis.mif.vu.lt** with your VU MIF username and password using **ssh** or **putty**. |
| + | * Check if you have a valid Kerberos key (ticket) with the **klist** command. If the key is not available or has expired, the **kinit** command must be used. | ||
| + | * Connect to the **hpc** node with the command **ssh hpc** (password must not be required). | ||
| - | Check the status of your jobs: | + | If **SSH keys** are used (e.g. if you need to copy big files): |
| + | * If you don't have SSH keys, you can find instructions on how to create them in a Windows environment **[[en: | ||
| + | * | ||
| + | * | ||
| - | <code bash> | + | The **first time** you connect, you **will not** be able to run **SLURM jobs** for the first **5 minutes**. After that, SLURM account will be created. |
| - | squeue -u $USER | + | |
| - | </ | + | |
| - | When the job disappears from the queue, it has finished. | + | ====== Lustre - Shared File System ====== |
| - | ===== Step 7: View the Results ===== | + | VU MIF HPC shared file system is available in the directory ''/ |
| - | Display | + | The system creates directory ''/ |
| - | <code bash> | + | The files in this file system are equally accessible on all compute nodes and on the **hpc** node. |
| - | cat hello.out | + | |
| - | </ | + | |
| - | Example output: | + | Please use these directories only for their purpose and clean them up after calculations. |
| - | < | + | ====== |
| - | Hello from HPC! | + | |
| - | compute01 | + | |
| - | Mon Nov 17 14:32:05 EET 2026 | + | |
| - | </ | + | |
| - | Congratulations! You have successfully executed your first HPC job. | + | ^Partition ^Time limit ^RAM |
| + | ^main | ||
| + | ^gpu ^48h | ||
| + | ^power | ||
| - | ===== Step 8: Request Resources | + | The time limit for tasks is **2h** in all partitions if it has not been specified. The table shows the maximum time limit. |
| - | Real HPC workloads usually require specific | + | The **RAM** column gives the amount of RAM allocated to each reserved **CPU** core. |
| - | Create a new script: | + | ====== Batch Processing of Tasks (SLURM) ====== |
| - | <code bash> | + | To use computing |
| - | nano resources.sh | + | |
| - | </ | + | |
| - | Paste the following content: | + | Example: |
| - | < | + | < |
| #!/bin/bash | #!/bin/bash | ||
| - | #SBATCH --job-name=resources-demo | + | #SBATCH -p main |
| - | #SBATCH --partition=main | + | #SBATCH -n4 |
| - | #SBATCH --cpus-per-task=4 | + | module load openmpi |
| - | #SBATCH --mem=4G | + | mpicc -o mpi-test mpi-test.c |
| - | # | + | mpirun mpi-test |
| - | # | + | |
| - | + | ||
| - | echo "Job ID:" | + | |
| - | echo $SLURM_JOB_ID | + | |
| - | + | ||
| - | echo "" | + | |
| - | echo " | + | |
| - | echo $SLURM_JOB_PARTITION | + | |
| - | + | ||
| - | echo "" | + | |
| - | echo " | + | |
| - | echo $SLURM_CPUS_PER_TASK | + | |
| - | + | ||
| - | echo "" | + | |
| - | echo " | + | |
| </ | </ | ||
| - | Submit | + | After submission and confirmation of your application to the ITOAC services, you need to create a user at https:// |
| - | < | + | < |
| - | sbatch resources.sh | + | |
| - | </ | + | |
| - | + | ||
| - | ===== Understanding Resource Requests ===== | + | |
| - | + | ||
| - | ^ Directive ^ Description ^ | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | + | ||
| - | <note tip> | + | |
| - | Most CPU workloads should use the '' | + | |
| - | + | ||
| - | GPU workloads should use the '' | + | |
| - | </ | + | |
| - | + | ||
| - | <note tip> | + | |
| - | If no runtime limit is specified, SLURM uses a default limit of 2 hours. However, specifying a realistic runtime is recommended because it helps the scheduler start jobs more efficiently. | + | |
| - | </ | + | |
| - | + | ||
| - | <note tip> | + | |
| - | Request only the resources you actually need. Smaller jobs generally start faster and help ensure fair sharing of cluster resources. | + | |
| - | </ | + | |
| - | + | ||
| - | ===== Check the Job Output ===== | + | |
| - | + | ||
| - | Once the job completes: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | cat resources.out | + | |
| - | </ | + | |
| - | + | ||
| - | ===== Step 9: Using Project Allocations ===== | + | |
| - | + | ||
| - | Some users are assigned project allocations that provide additional CPU, GPU, memory, or storage resources. | + | |
| - | + | ||
| - | If your project administrator provides an allocation name, add it to your job script: | + | |
| - | + | ||
| - | <code bash> | + | |
| #!/bin/bash | #!/bin/bash | ||
| - | #SBATCH --job-name=project-demo | + | #SBATCH --account=alloc_xxxx_projektas |
| - | #SBATCH --partition=main | + | #SBATCH -p main |
| - | #SBATCH --account=alloc_xxxx_project | + | #SBATCH -n4 |
| - | #SBATCH --cpus-per-task=4 | + | #SBATCH --time=minutes |
| - | # | + | module load openmpi |
| - | # | + | mpicc -o mpi-test mpi-test.c |
| - | + | mpirun mpi-test | |
| - | hostname | + | |
| </ | </ | ||
| - | The allocation name determines which project' | ||
| - | <note important> | + | It contains instructions for the task performer as special comments. |
| - | VU MIF users using the default allocation do NOT need to specify the '' | + | |
| - | Only users associated with a project allocation should use the '' | + | -p short - which queue to send to (main, gpu, power). |
| - | </ | + | |
| - | Example: | + | -n4 - how many processors to reserve (**NOTE:** if you set the number of cores to be used to x, but actually use fewer cores programmatically, |
| - | <code bash> | + | The initial running directory of the task is the current directory (**pwd**) on the login node from where the task is run, unless it was changed to another directory by the -D parameter. For the initial running directory, use the HPC shared filesystem directories **/scratch/ |
| - | # | + | |
| - | </code> | + | |
| - | ===== Step 10: Run a Multi-Core Python Job ===== | + | The generated script is sent with the command //sbatch//, |
| - | Python is one of the most commonly used programming languages on HPC systems. This example performs a matrix multiplication, | + | '' |
| - | ==== Create a Python Script ==== | + | which returns the number of the submitted job **JOBID**. |
| - | Create | + | The status of a pending or ongoing task can be checked with the command //squeue// |
| - | <code bash> | + | '' |
| - | nano matrix.py | + | |
| - | </ | + | |
| - | Paste: | + | With the //scancel// command it is possible to cancel the running of a task or to remove it from the queue |
| - | <code python> | + | '' |
| - | import os | + | |
| - | import numpy as np | + | |
| - | from time import perf_counter | + | |
| - | cpus = os.environ.get(" | + | If you forgot your tasks **JOBID**, you can check them with the command //squeue// |
| - | print(f" | + | '' |
| - | size = 2000 | + | Completed tasks are no longer displayed in **squeue**. |
| - | print(f" | + | If the specified number of processors is not available, your task is added to the queue. It will remain in the queue until a sufficient number of processors become available or until you remove it with **scancel**. |
| - | a = np.random.rand(size, size) | + | The **output** of the running job is recorded in the file **slurm-JOBID.out**. The error output is written to the same file unless you specified somewhere else. The file names can be changed with the **sbatch** command parameters -o (specify the output file) and -e (specify the error file). |
| - | b = np.random.rand(size, size) | + | |
| - | start = perf_counter() | + | More about SLURM opportunities you can read [[https:// |
| - | c = a @ b | + | |
| - | end = perf_counter() | + | |
| - | print(" | + | ====== Interactive Tasks (SLURM) ====== |
| - | print(f" | + | |
| - | print(f" | + | |
| - | </ | + | |
| - | ==== Create a SLURM Job Script ==== | + | Interactive tasks can be done with the //srun// command: |
| - | Create a new file: | + | < |
| - | + | $ srun --pty $SHELL | |
| - | < | + | |
| - | nano python-job.sh | + | |
| </ | </ | ||
| - | Paste: | + | The above command will connect you to the compute node environment assigned to SLURM and allow you to directly run and debug programs on it. |
| - | <code bash> | + | After the commands are done disconnect from the compute node with the command |
| - | # | + | |
| - | #SBATCH --job-name=python-demo | + | |
| - | #SBATCH --partition=main | + | |
| - | #SBATCH --cpus-per-task=4 | + | |
| - | #SBATCH --mem=2G | + | |
| - | #SBATCH --time=00: | + | |
| - | #SBATCH --output=python-demo.out | + | |
| - | export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK | + | < |
| - | export OPENBLAS_NUM_THREADS=$SLURM_CPUS_PER_TASK | + | $ exit |
| - | export MKL_NUM_THREADS=$SLURM_CPUS_PER_TASK | + | |
| - | + | ||
| - | python3 matrix.py | + | |
| - | </code> | + | |
| - | + | ||
| - | Submit the job: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | sbatch python-job.sh | + | |
| - | </ | + | |
| - | + | ||
| - | ==== Check the Results ==== | + | |
| - | + | ||
| - | After the job finishes: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | cat python-demo.out | + | |
| </ | </ | ||
| - | Example output: | + | If you want to run graphical programs, you need to connect to **ssh -X** to **uosis.mif.vu.lt** and **hpc**: |
| < | < | ||
| - | CPU cores allocated by SLURM: 4 | + | $ ssh -X uosis.mif.vu.lt |
| - | Creating two 2000x2000 matrices... | + | $ ssh -X hpc |
| - | Matrix multiplication completed | + | $ srun --pty $SHELL |
| - | Result shape: (2000, 2000) | + | |
| - | Execution time: 1.82 seconds | + | |
| </ | </ | ||
| - | ==== Optional: Load a Python Module ==== | + | In **power** |
| - | + | ||
| - | Depending on the cluster | + | |
| - | + | ||
| - | <code bash> | + | |
| - | module avail Python | + | |
| - | </ | + | |
| - | + | ||
| - | Load one of the available versions shown by the system: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | module load Python/< | + | |
| - | </ | + | |
| - | + | ||
| - | Verify: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | python3 --version | + | |
| - | </ | + | |
| - | + | ||
| - | ===== Step 11: Verify CPU Usage ===== | + | |
| - | + | ||
| - | While a job is running, you can monitor resource utilization | + | |
| - | + | ||
| - | <code bash> | + | |
| - | sstat -j JOBID.batch --format=JobID, | + | |
| - | </ | + | |
| - | + | ||
| - | Example: | + | |
| < | < | ||
| - | | + | $ srun -p power --mpi=none |
| - | ------------ ---------- ---------- ---------- | + | |
| - | 123456.batch | + | |
| </ | </ | ||
| - | Where: | + | ====== GPU Tasks (SLURM) ====== |
| - | ^ Field ^ Description ^ | + | To use GPU you need to specify additionally < |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | You can identify your Job ID with: | + | With '' |
| - | + | ||
| - | <code bash> | + | |
| - | squeue | + | |
| - | </ | + | |
| - | + | ||
| - | ==== Compare Elapsed Time and CPU Time ==== | + | |
| - | + | ||
| - | After the job finishes: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | sacct -j JOBID --format=JobID, | + | |
| - | </ | + | |
| - | + | ||
| - | Example: | + | |
| + | Example of an interactive task with 1 GPU: | ||
| < | < | ||
| - | JobID AllocCPUS | + | $ srun -p gpu --gres gpu --pty $SHELL |
| - | ------------ ---------- ---------- ---------- | + | |
| - | 123456 | + | |
| </ | </ | ||
| - | ^ Field ^ Description ^ | + | ====== Introduction |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | + | ||
| - | ==== How to Interpret the Results | + | |
| - | + | ||
| - | Single-threaded application: | + | |
| + | Ubuntu 18.04 LTS is the packet of **2.1.1** OpenMPI version. | ||
| + | To use the newer version **4.0.1** you need to use | ||
| < | < | ||
| - | AllocCPUS = 4 | + | module load openmpi/4.0 |
| - | Elapsed | + | |
| - | CPUTime | + | |
| </ | </ | ||
| + | before running MPI commands. | ||
| - | The application used approximately one CPU core, even though four cores were allocated. | + | ===== MPI Compiling Programs ===== |
| - | Multi-threaded application: | + | An example of a simple MPI program is in the directory ''/ |
| < | < | ||
| - | AllocCPUS = 4 | + | $ mpicc -o foo foo.c |
| - | Elapsed | + | $ mpif77 -o foo foo.f |
| - | CPUTime | + | $ mpif90 -o foo foo.f |
| </ | </ | ||
| + | ===== Implementation of MPI Programmes ===== | ||
| - | The application actively used multiple CPU cores. | + | MPI programs are started with **mpirun** or **mpiexec**. You can learn more about them with the **man mpirun** or **man mpiexec** command. |
| - | In an ideal case, a fully utilized four-core job running for ten seconds would show: | + | A simple (SPMD) program can be started with the following mpirun command line. |
| < | < | ||
| - | Elapsed | + | $ mpirun foo |
| - | CPUTime | + | |
| </ | </ | ||
| - | <note tip> | + | All allocated processors will be used according |
| - | A quick way to verify that your application is using multiple CPU cores is to compare '' | + | |
| - | </ | + | |
| - | <note tip> | + | **ATTENTION** It is strictly forbidden to use more CPU than you have reserved, as this may affect the performance of other tasks. |
| - | Requesting multiple | + | |
| - | # | + | Find more information on [[https:// |
| - | does not automatically make an application run in parallel. The application must also support multithreading and be configured to use the allocated cores. | + | ====== Task Efficiency ====== |
| - | </ | + | |
| - | ===== Step 12: Submit a GPU Job ===== | + | * Please use at least 50% of the ordered CPU quantity. |
| + | * Using more CPUs than ordered will not improve performance, | ||
| + | * If you use the '' | ||
| - | To run a job on a GPU node: | + | ====== The Limits of Resources ====== |
| - | <code bash> | + | If your tasks don't start because of **AssocGrpCPUMinutesLimit** or **AssocGrpGRESMinutes**, |
| - | nano gpu-demo.sh | + | |
| - | </code> | + | |
| - | Paste: | + | //The first way to see how much resources are used:// |
| - | < | + | < |
| - | # | + | sreport |
| - | #SBATCH --job-name=gpu-demo | + | |
| - | #SBATCH --partition=gpu | + | |
| - | #SBATCH --gres=gpu:1 | + | |
| - | #SBATCH --cpus-per-task=4 | + | |
| - | #SBATCH --mem=16G | + | |
| - | #SBATCH --time=01:00:00 | + | |
| - | + | ||
| - | nvidia-smi | + | |
| </ | </ | ||
| - | Submit | + | Where the **USERNAME** - is your MIF user name. **Start** and **End** show the start and end days of the current month. You can specify them also by '' |
| - | <code bash> | + | **NOTE** Usage of resources is given in minutes, divide the number by 60 to get hours. |
| - | sbatch gpu-demo.sh | + | |
| - | </ | + | |
| - | The command '' | + | //The second way to see how much resources are used:// |
| - | Users working within a project allocation may also need: | + | < |
| - | + | sshare -l -A USERNAME_mif | |
| - | < | + | |
| - | # | + | |
| </ | </ | ||
| - | ==== Interactive GPU Session ==== | + | Where **USERNAME** is your MIF user name. Or specify the account whose usage you want to see in **-A**. The data is also displayed in minutes: |
| - | + | * **GrpTRESRaw** - how much is used. | |
| - | To start an interactive shell on a GPU node: | + | * **GrpTRESMins** - what is the limit. |
| - | + | * **GGRTRESRunMins** - the remaining | |
| - | <code bash> | + | |
| - | srun -p gpu --gres=gpu: | + | |
| - | </ | + | |
| - | + | ||
| - | After the session starts: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | nvidia-smi | + | |
| - | </ | + | |
| - | + | ||
| - | Interactive sessions are useful for testing software, debugging, and experimenting before submitting batch jobs. | + | |
| - | + | ||
| - | ===== Useful Commands ===== | + | |
| - | + | ||
| - | ^ Command ^ Purpose ^ | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | | '' | + | |
| - | + | ||
| - | ===== Typical Resource Requests ===== | + | |
| - | + | ||
| - | ^ Workload ^ Partition ^ CPUs ^ Memory ^ GPUs ^ | + | |
| - | | Learning and testing | main | 1-2 | 1-4 GB | None | | + | |
| - | | Data analysis | main | 4-8 | 8-32 GB | Optional | | + | |
| - | | Scientific simulations | main | 8-64 | 16-128 GB | Optional | | + | |
| - | | Machine learning and AI | gpu | 4-16 | 16-64 GB | 1 or more | | + | |
| - | + | ||
| - | ===== Next Steps ===== | + | |
| - | + | ||
| - | You now know how to: | + | |
| - | + | ||
| - | | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | * Verify resource utilization | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | + | ||
| - | Continue with: | + | |
| - | + | ||
| - | | + | |
| - | * [[hpc: | + | |
| - | | + | |
| - | + | ||
| - | ===== Common Problems ===== | + | |
| - | + | ||
| - | ==== Permission denied ==== | + | |
| - | + | ||
| - | Verify that: | + | |
| - | + | ||
| - | | + | |
| - | | + | |
| - | | + | |
| - | + | ||
| - | ==== sbatch: command not found ==== | + | |
| - | + | ||
| - | You are probably running the command on your local computer instead of the HPC login node. | + | |
| - | + | ||
| - | ==== Invalid account or account/ | + | |
| - | + | ||
| - | Possible causes: | + | |
| - | + | ||
| - | * Your SLURM account is still being created. | + | |
| - | | + | |
| - | | + | |
| - | + | ||
| - | ==== Job remains in PENDING state ==== | + | |
| - | + | ||
| - | Resources may currently be unavailable. | + | |
| - | + | ||
| - | Check your jobs: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | squeue -u $USER | + | |
| - | </ | + | |
| - | + | ||
| - | ==== GPU job does not start ==== | + | |
| - | + | ||
| - | Verify that: | + | |
| - | + | ||
| - | | + | |
| - | | + | |
| - | * GPU resources are available | + | |
| - | * Your allocation includes GPU access | + | |
| - | + | ||
| - | ==== Python command not found ==== | + | |
| - | + | ||
| - | Verify that Python is installed in the environment or load the appropriate Python module: | + | |
| - | + | ||
| - | <code bash> | + | |
| - | module avail Python | + | |
| - | </ | + | |
| - | ===== Where to Go Next ===== | + | ** |
| + | How are monthly limits assigned? | ||
| - | Recommended learning path: | + | Once an application has been submitted to get more CPU or GPU hours, a decision is usually made within 5 business days. The monthly limit is increased on the day the decision is made in the current month, minus the limit already used up that month. For example, if a user who has used 100 GPU-h requests that the limits be increased to 600 GPU-h per month, they will be granted an additional 500 GPU-h for that month. The following month, the user will again have 600 GPU-h, so the unused balance from the previous month is not carried over to the following month. |
| + | ====== The Links ====== | ||
| - | | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | | + | |
| - | ---- | ||
| - | < | ||
| - | This guide was prepared and updated by the VU MIF HPC team with the assistance of Microsoft Copilot. All procedures and examples should be validated against the current HPC environment, | ||
| - | </ | ||
en/hpc.txt · Last modified: by eduardas
