====== GPGPU compute servers ====== ===== Service description ===== These servers are intended for GPU computing (GPGPU), aimed at compute-intensive tasks, machine learning, data processing and scientific simulation that require acceleration by graphics hardware. ==== Free-access servers ==== Any researcher at the center may request access to these servers. Access is granted upon request and validation. ^ Node ^ Server ^ CPU ^ RAM ^ GPUs ^ Operating System ^ Job management ^ | ''ctgpgpu4'' | PowerEdge R730 | 2 × [[https://ark.intel.com/products/92980/Intel-Xeon-Processor-E5-2623-v4-10M-Cache-2_60-GHz|Intel Xeon E5-2623 v4]] | 128 GB | 2 × Nvidia GP102GL 24GB (Tesla P40, 2016) | AlmaLinux 9.1 \\ • CUDA 12.0 | **Slurm (mandatory use)** | * Servers in the HPC compute cluster: [[ centro:servizos:hpc | HPC compute cluster ]] * Servers at CESGA: [[ centro:servizos:cesga | Request access ]] ==== Restricted-access servers ==== Access to these servers is restricted to a specific group, a particular project, or is more tightly controlled for resource management and planning reasons. It is essential to check the updated information in Xici at the time of requesting the service, where the particular circumstances of each server are detailed (access criteria, priorities, usage conditions, etc.). ^ Node ^ Server ^ CPU ^ RAM ^ GPUs ^ Operating System ^ Job management ^ | ''ctgpgpu5'' | PowerEdge R730 | 2 × [[https://ark.intel.com/products/92980/Intel-Xeon-Processor-E5-2623-v4-10M-Cache-2_60-GHz|Intel Xeon E5-2623 v4]] | 128 GB | 2 × Nvidia GP102GL (Tesla P40) | Ubuntu 22.04 \\ • Nvidia driver 590 \\ • CUDA Toolkit 12.5 and 13.1 (default) | n/a | | ''ctgpgpu6'' | SIE LADON 4214 | 2 × [[https://ark.intel.com/content/www/us/en/ark/products/193385/intel-xeon-silver-4214-processor-16-5m-cache-2-20-ghz.html|Intel Xeon Silver 4214]] | 192 GB | Nvidia Quadro P6000 24GB (2018) \\ Nvidia Quadro RTX8000 48GB (2019) \\ 2 × Nvidia A30 24GB (2020) | CentOS 7.9 \\ • Nvidia driver 535.86.10 (CUDA 12.2) | n/a | | ''ctgpgpu9'' | Dell PowerEdge R750 | 2 × [[https://ark.intel.com/content/www/es/es/ark/products/215274/intel-xeon-gold-6326-processor-24m-cache-2-90-ghz.html|Intel Xeon Gold 6326]] | 128 GB | 2 × NVIDIA Ampere A100 80GB | AlmaLinux 8.6 \\ • NVIDIA driver 515.48.07 (CUDA 11.7) | n/a | | ''ctgpgpu11'' | Gigabyte G482-Z54 | 2 × [[https://www.amd.com/es/products/cpu/amd-epyc-7413|AMD EPYC 7413 @2.65 GHz (24c)]] | 256 GB | 5 × NVIDIA Ampere A100 80GB | AlmaLinux 9.1 \\ • NVIDIA driver 520.61.05 (CUDA 11.8) | n/a | | ''ctgpgpu12'' | Dell PowerEdge R760 | 2 × [[https://ark.intel.com/content/www/xl/es/ark/products/232376.html|Intel Xeon Silver 4410Y]] | 384 GB | 2 × NVIDIA Hopper H100 80GB | AlmaLinux 9.2 \\ • NVIDIA driver 555.42.06 (CUDA 12.5) | n/a | | ''ctgpgpu13'' | Gigabyte G493-ZB1-AAP1 | 2x AMD EPYC 9474F (48c) | 1536 GB | Nvidia RTX Pro 6000 Blackwell Server Edition\\ Nvidia H100 NVL\\ Nvidia L40S | AlmaLinux 9.6\\ • NVIDIA driver 580.95.05 (CUDA 13.0) | gpuctl | | ''ctgpgpu14'' | Gigabyte R283-ZF0-AAL1 | 2 × [[https://www.amd.com/es/products/processors/server/epyc/4th-generation-9004-and-8004-series/amd-epyc-9554.html|AMD EPYC 9554 (128c)]] | 768 GB | 2 x Nvidia Blackwell Pro 6000 96GB | AlmaLinux 10.1 \\ • NVIDIA driver 595.71.05 (CUDA 13.2) | n/a | | ''ctgpgpu15'' ⚠️ | SIE LADON (Gigabyte) | 2x AMD EPYC 9474F (48c) | 768 GB | 4 × NVIDIA H200 NVL | AlmaLinux 9.6 | Slurm | | ''ctgpgpu16'' ⚠️ | SIE LADON (Gigabyte) | 2x AMD EPYC 9474F (48c) | 768 GB | 4 × NVIDIA H200 NVL | AlmaLinux 9.7 | gpuctl | | ''ctgpgpu17'' ⚠️ | SIE LADON (Gigabyte) | 2x AMD EPYC 9474F (48c) | 768 GB | 4 × NVIDIA H200 NVL | AlmaLinux 9.7 | gpuctl | | ''ctgpgpu18'' ⚠️ | SIE LADON (MegaRAC SP-X) | 2x AMD EPYC 9335 (24c) | 1536 GB | 4 × NVIDIA H200 | Ubuntu 22.04 | gpuctl | ⚠️ The ''ctgpgpu15'', ''ctgpgpu16'', ''ctgpgpu17'' and ''ctgpgpu18'' servers have a temporary installation and allocations, and their configuration and access may be changed in the future. ===== Service registration ===== Not all servers are available at all times for any use. To access the servers, you must request it beforehand through the [[https://citius.usc.es/uxitic/incidencias/add|incident form]]. Users who do not have access permission will receive an incorrect password message. ===== User manual ===== ==== Connecting to the servers ==== To connect to the servers, you must do so via SSH. The names and IP addresses of the servers are as follows: ^ Node ^ FQDN ^ IP ^ | ''ctgpgpu4'' | ctgpgpu4.inv.usc.es | 172.16.242.201 | | ''ctgpgpu5'' | ctgpgpu5.inv.usc.es | 172.16.242.202 | | ''ctgpgpu6'' | ctgpgpu6.inv.usc.es | 172.16.242.205 | | ''ctgpgpu9'' | ctgpgpu9.inv.usc.es | 172.16.242.94 | | ''ctgpgpu11'' | ctgpgpu11.inv.usc.es | 172.16.242.96 | | ''ctgpgpu12'' | ctgpgpu12.inv.usc.es | 172.16.242.97 | | ''ctgpgpu15'' | ctgpgpu15.inv.usc.es | 172.16.242.207 | | ''ctgpgpu16'' | ctgpgpu16.inv.usc.es | 172.16.242.212 | | ''ctgpgpu17'' | ctgpgpu17.inv.usc.es | 172.16.242.213 | | ''ctgpgpu18'' | ctgpgpu18.inv.usc.es | 172.16.242.208 | Connections are only available from the center's network. To connect from other locations or from the RAI network you must use the [[:en:centro:servizos:vpn:start|VPN]] or the [[:en:centro:servizos:pasarela_ssh|SSH gateway]]. ==== Job management with SLURM ==== On servers where the Slurm queue manager is present, its use is mandatory to submit jobs and thus avoid conflicts between processes, since two jobs should not be run at the same time. Users are limited to using two cores and 1GB of RAM outside Slurm. To submit a job to the queue use the "sbatch" command with a Slurm script or directly ''srun'': srun cuda_program cuda_program_arguments It is mandatory to request at least one GPU when submitting jobs to this server or the job will be rejected with the following message: "Job rejected: a GPU must be requested (e.g. --gres=gpu:1 or --gpus=H200:1)." The ''srun'' process waits for the job to run before returning control to the user. If you do not want to wait, you can use console session managers such as ''screen'', allowing you to leave the job queued and disconnect the session without worry and retrieve the console output later. Alternatively, you can use ''nohup'' and send the job to the background with ''&''. In this case the output is saved to the file ''nohup.out'': nohup srun cuda_program cuda_program_arguments & To see the queue status use the ''squeue'' command. The command shows an output similar to this: JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON) 9 servidore ca_water pablo.qu PD 0:00 1 (Resources) 10 servidore ca_water pablo.qu PD 0:00 1 (Priority) 11 servidore ca_water pablo.qu PD 0:00 1 (Priority) 12 servidore ca_water pablo.qu PD 0:00 1 (Priority) 13 servidore ca_water pablo.qu PD 0:00 1 (Priority) 14 servidore ca_water pablo.qu PD 0:00 1 (Priority) 8 servidore ca_water pablo.qu R 0:11 1 ctgpgpu2 You can also get an interactive view, updated every second, with the ''smap'' command: smap -i 1 ==== Reservations and jobs management with gpuctl ==== On servers that use ''gpuctl'', you must use the ''gpu'' command to request GPUs and be able to run jobs on them. The ''gpu'' command does not include a job queue or parallelism control. It can be used interactively or from scripts, but to queue multiple jobs or run them in parallel you should use external tools, such as ''task-spooler'' (''tsp'' command). There are two main workflows: === 1. Automatic GPU reservation === This method is best suited if you only want to run a single command. ''gpu exec'' acts as a wrapper: it takes care of claiming the GPU, preparing the environment needed for the command and releasing it when finished. gpu exec python ./train.py This is the recommended option for simple runs, as it avoids having to manually manage the reservation and release of the GPU. === 2. Manual GPU reservation === This method is best suited if you want to keep an active reservation across multiple commands, whether in an interactive session or in a workflow with several consecutive runs. gpu claim gpu exec python ./train.py gpu release Or, alternatively: gpu claim eval "$(gpu env --shell)" python ./train.py gpu release If there is already an active reservation, ''gpu exec'' will reuse it and automatically prepare the appropriate environment for the command. It is important to remember that, after creating or modifying a reservation, you must run ''eval "$(gpu env --shell)"'' again if you are going to launch commands directly from the shell. Alternatively, you can use ''gpu exec'', which prepares the environment on each execution. === Lifetime of reservations === Reservations are kept while real compute activity on the GPU is detected. If a GPU remains reserved without activity for a prolonged time, the reservation may be lost automatically due to inactivity. Therefore, if you later try to run a job assuming the reservation is still active, the job may fail. A reservation is not kept simply by having a terminal open, but by actual GPU activity. The inactivity expiration times are configured as follows: * Monday to Friday, from 10:00 to 18:00 → 1 hour * Other times → 2 hours * If the reservation seems abandoned (there is no user process on the system nor any shell) → 10 minutes === Checking reservation status === If you are working with a manual reservation and want to check if it is still active, you can use: gpu mine Or, if you want general information about all GPUs, the time when real usage was captured for each, and the currently configured expiration time, you can run: gpu status This is especially recommended if time has passed since the last execution or if the GPU may have gone unused for a long period. === Queue behavior of reservations === Although ''gpu'' does not implement a job queue as such, there is an implicit waiting system for GPU reservation. When a command that requires reserving a GPU is executed and none are available, the command does not fail but remains blocked waiting for a GPU to become free. The user is placed in an internal queue managed by ''gpuctl'' and the execution will continue automatically as soon as the reservation can be satisfied. This behavior applies to both manual and automatic reservations. For example: gpu exec python ./train.py If there are no GPUs available, this command will wait until one is released. Similarly: gpu claim will also remain blocked until it can obtain a GPU. === Selecting a specific GPU === By default, ''gpuctl'' automatically assigns an available GPU. However, it is possible to request a specific GPU by its index. Both ''gpu claim'' and ''gpu exec'' allow using the ''--gpu-index'' parameter for this purpose: gpu claim --gpu-index 0 gpu exec --gpu-index 1 python ./train.py In this case, the command will attempt to reserve the specified GPU. If that GPU is not available, the command will wait in queue until that particular GPU is released. This behavior allows fixing runs to specific GPUs, but it may increase waiting time if the chosen GPU is highly requested. === Opportunistic reservation of a second GPU === In addition to the main reservation, which is made with priority in the ''guaranteed'' queue, it is possible to reserve a second GPU in the ''burst_preemptible'' queue on some servers (check the reservation policy message when logging into the server; if applicable it will mention a second opportunistic GPU). This second GPU may be used whenever it is free, but it is not guaranteed: it may be lost even in the middle of a job if another user needs that GPU for their primary reservation. This allows taking advantage of additional free capacity when available, but jobs that depend on this second GPU must be prepared to tolerate its loss. To request this second GPU, just run ''gpu claim'' a second time: gpu claim gpu claim You can also explicitly specify the number of GPUs: gpu claim --numgpus 2 In this case, the command will wait until it can reserve the requested number of GPUs. Note that these two forms are not exactly equivalent: * with ''gpu claim'' followed by another ''gpu claim'', the first reservation is obtained earlier and the second is requested afterwards; * with ''gpu claim ---numgpus 2'', the request is made jointly and the command will wait until it can reserve both GPUs at once. It is important to remember that, after making any reservation, you must run ''eval "$(gpu env --shell)"'' again or perform executions with ''gpu exec''. If the second GPU is lost later for being ''preemptible'', the jobs using it may fail or be interrupted. Therefore: * the first GPU is the priority and guaranteed reservation; * the second GPU is opportunistic and will only be available while it is not needed for another priority reservation; * if any GPU cannot be reserved, the command will remain waiting and the user will enter the corresponding queue; * when reserving two GPUs at once, the command will wait until both can be reserved simultaneously. === Using task-spooler (tsp) === Since ''gpu'' does not incorporate a job queue or parallelism control, a practical option to queue runs is to use ''task-spooler'', with the ''tsp'' command, combined with ''gpu''. == Run several jobs in series with the same reservation == gpu claim tsp gpu exec python ./train1.py tsp gpu exec python ./train2.py tsp gpu exec python ./train3.py tsp gpu release This method allows reusing the same reservation for several consecutive jobs queued in ''tsp''. == Run several jobs in parallel == If you want to run several jobs at the same time, you can increase the number of ''slots'' of ''tsp'': gpu claim tsp -S 2 tsp gpu exec python ./train1.py tsp gpu exec python ./train2.py tsp gpu exec python ./train3.py tsp gpu exec python ./train4.py tsp -w tsp -S 1 gpu release In this case, be careful not to run ''gpu release'' before all jobs have finished. First make sure the queue has finished, and only then release the GPU. Note: running multiple jobs in parallel does not imply reserving multiple GPUs. If only one GPU is reserved, all processes will share that same GPU and compete for its resources, such as memory and compute time. Also remember that the reservation is only maintained while real compute activity on the GPU is detected. If the jobs go too long without using it, the reservation may expire automatically due to inactivity. === Using tmux === Using ''tmux'' allows keeping terminal sessions, disconnecting from them and recovering them later, thus allowing to maintain an interactive session without keeping the connection open. Disconnecting does not send ''HUP'' to the processes. To start a ''tmux'' session or reconnect to it if it already exists, it is recommended to use a fixed name: tmux new -A -s main This creates the ''main'' session if it does not exist, or reconnects to it if it was already created. Once inside ''tmux'', to detach while keeping the session active press ''Control+b'', release, and then press ''d''. At this point you can disconnect from the machine without interrupting the processes. To return to the same session, simply run the previous command again. If you want to completely terminate the session from outside: tmux kill-session -t main === GPU partitioning with MIG profiles === ''gpuctl'' allows partitioning a GPU into multiple MIG instances to isolate resources and run smaller workloads in parallel on a single physical GPU in a more controlled way. This functionality is only available on compatible GPUs; otherwise it cannot be used. For now, only NVIDIA H200 GPUs have this function enabled. To reserve a GPU and enable it in MIG mode: gpu claim --mig half gpu claim --mig third gpu claim --mig quarter This will split the GPU into two, three or four partitions, choosing the most appropriate MIG profiles according to the compatible card. Note that the ''--mig'' parameter can only be used when exactly one GPU is claimed. Once the MIG reservation is created, ''gpu exec'' and ''gpu env'' automatically prepare ''CUDA_VISIBLE_DEVICES'' with the corresponding MIG UUIDs. === Email notifications === ''gpuctl'' allows configuring the sending of email notifications related to reservation status and GPU usage. == View current configuration == gpu notify == Enable or disable notifications == gpu notify --mode off gpu notify --mode email gpu notify --mode telegram gpu notify --mode both When notifications are enabled, emails or Telegram messages are sent in the following cases: * when a GPU is automatically released for remaining unused; * when the enforcer kills a process for using a GPU without having an active reservation or for losing an opportunistic GPU; * when a reservation request is queued; * and subsequently when that reservation is granted. Notifications are not sent in the case that a reservation is made and the reservation is granted immediately without waiting, or if the GPU is manually released with ''release'' or when an ''exec'' that made an automatic reservation finishes. == Configure a custom email == gpu notify --email your.email@domain If a custom email is not configured, the address obtained from LDAP will be used by default. To return to the default email: gpu notify --clear-email == Pair Telegram account == After running the following command: gpu notify --pair-telegram Follow the instructions to finish the pairing process. To remove the pairing: gpu notify --unpair-telegram == Sending custom notifications == The user can send custom notifications at various points in their workflow using: gpu notify --send "message content" This can be useful to receive personal alerts at the start or end of an experiment, after completing a processing phase, or at any other relevant point in the workflow. The preferred notification method configured will be used, or email if neither is configured. === Usage reports === To consult a summary of recent GPU usage history by process, you can use: gpu report This shows the recorded PIDs, the number of samples available and aggregated metrics of GPU, CPU and memory usage. If you want to see the data of a specific process, with ASCII charts and the timeline of stored samples, you can use: gpu report --pid PID If there are many samples, the output may appear summarized. To see the full series: gpu report --pid PID --all-samples === Practical recommendations === * To run a single command, the simplest is usually ''gpu exec''. * To keep an active reservation across several commands, it is more appropriate to use ''gpu claim'' + ''gpu exec'' + ''gpu release''. * If you are going to work directly from the shell after making a manual reservation, you must run ''eval "$(gpu env --shell)"'' or prefix commands with ''gpu exec''. * To keep sessions active and recover them later, ''tmux'' is usually the most convenient option. * To queue multiple jobs, ''tsp'' is a practical alternative. * If ''tsp'' is used with parallelism, remember that this does not reserve additional GPUs: processes will share the GPUs visible at that time. * If a reservation remains too long without real GPU compute activity, it may be lost automatically due to inactivity.