Merge branch '1-add-leonardo-system' into 'main'

Update 6 files

Closes #1

See merge request !3

data/form-schema.json

@@ -13,7 +13,8 @@
                     "JUWELS",
                     "LUMI",
                     "MockHPCSystem",
-                    "VEGA"
+                    "VEGA",
+                    "LEONARDO"
                 ]
             }
         },
@@ -95,6 +96,15 @@
                                     "Pytorch-DDP",
                                     "DeepSpeed"
                                 ]
+                            },
+                            {
+                                "key": [
+                                    "LEONARDO"
+                                ],
+                                "value": [
+                                    "Pytorch-DDP",
+                                    "Horovod"
+                                ]
                             }
                         ]
                     }
@@ -195,6 +205,14 @@
                                     "dp-esb",
                                     "dp-dam"
                                 ]
+                            },
+                            {
+                                "key": [
+                                    "LEONARDO"
+                                ],
+                                "value": [
+                                    "boost_usr_prod"
+                                ]
                             }
                         ]
                     }
@@ -638,6 +656,14 @@
                         [
                             "DEEP",
                             "DeepSpeed"
+                        ],
+                        [
+                            "LEONARDO",
+                            "Horovod"
+                        ],
+                        [
+                            "LEONARDO",
+                            "Pytorch-DDP"
                         ]
                     ]
                 }
@@ -722,6 +748,14 @@
                         [
                             "DEEP",
                             "DeepSpeed"
+                        ],
+                        [
+                            "LEONARDO",
+                            "Horovod"
+                        ],
+                        [
+                            "LEONARDO",
+                            "Pytorch-DDP"
                         ]
                     ]
                 }
@@ -735,7 +769,8 @@
             "JUWELS": "https://apps.fz-juelich.de/jsc/hps/juwels/index.html",
             "LUMI": "https://docs.lumi-supercomputer.eu/software/",
             "VEGA": "https://doc.vega.izum.si",
-            "Cyclone": "https://hpcf.cyi.ac.cy/documentation/"
+            "Cyclone": "https://hpcf.cyi.ac.cy/documentation/",
+            "LEONARDO": "https://wiki.u-gov.it/confluence/display/SCAIUS/HPC+User+Guide"
         },
         "software": {
             "Pytorch-DDP": "https://pytorch.org/tutorials/intermediate/ddp_tutorial.html",

scripts/LEONARDO/Horovod/LEONARDO_Horovod_script.sh

+#!/usr/bin/bash
+
+#SBATCH --job-name=job
+#SBATCH --account=%account%
+#SBATCH --output=job.out
+#SBATCH --error=job.err
+#SBATCH --partition=%partition%
+#SBATCH --nodes=%nodes%
+#SBATCH --gpus=
+#SBATCH --gpus-per-node=4
+
+#MODULES BEGIN LEONARDO Horovod
+ml cudnn/8.9.7.29-12--gcc--12.2.0-cuda-12.1
+ml parallel-netcdf/1.12.3--openmpi--4.1.6--gcc--12.2.0
+ml hdf5/1.14.3--gcc--12.2.0 
+ml py-mpi4py/3.1.4--openmpi--4.1.6--gcc--12.2.0
+ml python/3.11.6--gcc--12.2.0-nlkgjki
+ml nccl/2.19.1-1--gcc--12.2.0-cuda-12.1
+ml cuda/12.1
+#MODULES END
+
+source your/env_path/bin/activate
+
+export CUDA_VISIBLE_DEVICES="0,1,2,3"
+
+srun --cpu-bind=none python3 -u %executable%
\ No newline at end of file

scripts/LEONARDO/Horovod/lamec.json

+{"template": "LEONARDO_Horovod_script.sh"}
\ No newline at end of file

scripts/LEONARDO/Pytorch-DDP/LEONARDO_Pytorch-DDP_script.sh

+#!/usr/bin/bash
+
+#SBATCH --job-name=job
+#SBATCH --account=%account%
+#SBATCH --output=job.out
+#SBATCH --error=job.err
+#SBATCH --time=01:00:00
+#SBATCH --partition=%partition%
+#SBATCH --nodes=%nodes%
+#SBATCH --gpus=
+#SBATCH --gpus-per-node=4
+
+#MODULES BEGIN LEONARDO Pytorch-DDP
+ml cudnn/8.9.7.29-12--gcc--12.2.0-cuda-12.1
+ml parallel-netcdf/1.12.3--openmpi--4.1.6--gcc--12.2.0
+ml hdf5/1.14.3--gcc--12.2.0 
+ml py-mpi4py/3.1.4--openmpi--4.1.6--gcc--12.2.0
+ml python/3.11.6--gcc--12.2.0-nlkgjki
+ml nccl/2.19.1-1--gcc--12.2.0-cuda-12.1
+ml cuda/12.1
+#MODULES END
+
+source your/env_path/bin/activate
+
+export CUDA_VISIBLE_DEVICES="0,1,2,3"
+export PSP_CUDA=1
+export PSP_UCP=1
+export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
+
+srun --cpu-bind=none bash -c "torchrun \
+   --nnodes=$SLURM_NNODES \
+   --nproc_per_node=$SLURM_GPUS_PER_NODE \
+   --rdzv_id=$SLURM_JOB_ID \
+   --rdzv_conf=is_host=\$(((SLURM_NODEID)) && echo 0 || echo 1) \
+   --rdzv_backend=c10d \
+   --rdzv_endpoint='$(scontrol show hostnames "$SLURM_JOB_NODELIST" | head -n 1)':29400 \
+   %executable%

scripts/LEONARDO/Pytorch-DDP/lamec.json

+{"template": "LEONARDO_Pytorch-DDP_script.sh"}
\ No newline at end of file

scripts/LEONARDO/sysinfo.json

+{
+    "partition": {
+        "boost_usr_prod": {
+            "nodes": 64
+        }
+    }
+}
\ No newline at end of file