{"type":"doc","content":[{"type":"paragraph","content":[{"text":"Goal","type":"text","marks":[{"type":"strong"}]},{"text":": Provide an exercise to work through that puts together the various concepts covered within this workshop.","type":"text"}]},{"type":"extension","attrs":{"layout":"default","extensionType":"com.atlassian.confluence.macro.core","extensionKey":"toc","parameters":{"macroParams":{"minLevel":{"value":"1"},"maxLevel":{"value":"1"},"outline":{"value":"false"},"style":{"value":"none"},"type":{"value":"list"},"printable":{"value":"true"}},"macroMetadata":{"macroId":{"value":"fa65aafd-9af3-48be-8735-1f99ace1ade6"},"schemaVersion":{"value":"1"},"title":"Table of Contents"}},"localId":"f4abd64a-70cb-4924-805c-692a3db74321"}},{"type":"rule"},{"type":"heading","attrs":{"level":1},"content":[{"text":"Exercise: Put Things Together","type":"text"}]},{"type":"panel","attrs":{"panelType":"warning"},"content":[{"type":"paragraph","content":[{"text":"Exercise","type":"text","marks":[{"type":"strong"}]},{"text":": Create a self contained Conda environment that would allow a user to run a python file that uses ","type":"text"},{"text":"PyTorch","type":"text","marks":[{"type":"link","attrs":{"href":"https://pytorch.org/get-started/locally/"}}]},{"text":". ","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"The environment should be created within a location that can be ","type":"text"},{"text":"shared","type":"text","marks":[{"type":"em"}]},{"text":" with other users within a project.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Make sure the environment can utilize GPUs running with ","type":"text"},{"text":"cuda","type":"text","marks":[{"type":"code"}]},{"text":" 12.4.","type":"text"}]}]}]}]},{"type":"panel","attrs":{"panelType":"warning"},"content":[{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Using the the following ","type":"text"},{"text":"pytorch_gpu_test.py","type":"text","marks":[{"type":"code"}]},{"text":" script, create a bash script to submit a job to the cluster that allows this to be run, using the created Conda environment, that utilizes one H100 NVidia GPU device on a single compute node, with 8 cores and 16G of memory.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Any output should be written into a ","type":"text"},{"text":"slurms","type":"text","marks":[{"type":"code"}]},{"text":" sub folder, with a filename of the form ","type":"text"},{"text":"pytorch_.out","type":"text","marks":[{"type":"code"}]}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Send email notification to yourself regards the status of the submission.","type":"text"}]}]}]}]},{"type":"expand","attrs":{"title":"pytorch_gpu_test.py"},"content":[{"type":"codeBlock","content":[{"text":"import torch\nimport math\n\nprint(\"PyTorch Version: \" + str(torch.__version__))\nprint(\"Cuda Available: \" + str(torch.cuda.is_available()))\nprint(\"Device Count: \" + str(torch.cuda.device_count()))\nprint(\"Device Name: \" + str(torch.cuda.get_device_name(0)))\n\ndtype = torch.float\ndevice = torch.device(\"cuda:0\") # Uncomment this to run on GPU\n\n# Create random input and output data\nx = torch.linspace(-math.pi, math.pi, 2000, device=device, dtype=dtype)\ny = torch.sin(x)\n\n# Randomly initialize weights\na = torch.randn((), device=device, dtype=dtype)\nb = torch.randn((), device=device, dtype=dtype)\nc = torch.randn((), device=device, dtype=dtype)\nd = torch.randn((), device=device, dtype=dtype)\n\nlearning_rate = 1e-6\nfor t in range(2000):\n # Forward pass: compute predicted y\n y_pred = a + b * x + c * x ** 2 + d * x ** 3\n\n # Compute and print loss\n loss = (y_pred - y).pow(2).sum().item()\n if t % 100 == 99:\n print(t, loss)\n\n # Backprop to compute gradients of a, b, c, d with respect to loss\n grad_y_pred = 2.0 * (y_pred - y)\n grad_a = grad_y_pred.sum()\n grad_b = (grad_y_pred * x).sum()\n grad_c = (grad_y_pred * x ** 2).sum()\n grad_d = (grad_y_pred * x ** 3).sum()\n\n # Update weights using gradient descent\n a -= learning_rate * grad_a\n b -= learning_rate * grad_b\n c -= learning_rate * grad_c\n d -= learning_rate * grad_d\n\nprint(f'Result: y = {a.item()} + {b.item()} x + {c.item()} x^2 + {d.item()} x^3')","type":"text"}]}]},{"type":"panel","attrs":{"panelType":"warning"},"content":[{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Ignoring any cached packages, how much storage space does this Conda environment use?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Which version of Python is being used?","type":"text"}]}]}]}]},{"type":"rule"},{"type":"heading","attrs":{"level":1},"content":[{"text":"Exercise: Extend with Pandas","type":"text"}]},{"type":"panel","attrs":{"panelType":"warning"},"content":[{"type":"paragraph","content":[{"text":"Exercise","type":"text","marks":[{"type":"strong"}]},{"text":": I want to also use the Python ","type":"text"},{"text":"Pandas","type":"text","marks":[{"type":"link","attrs":{"href":"https://pandas.pydata.org/pandas-docs/stable/index.html"}}]},{"text":" package, specifically version ","type":"text"},{"text":"2.1.4","type":"text","marks":[{"type":"code"}]},{"text":" to assist with some further analysis.","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Can I extend my existing Conda environment?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"If so how?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Should I?","type":"text"}]}]}]}]},{"type":"rule"},{"type":"heading","attrs":{"level":1},"content":[{"text":"Exercise: Suggested Answers: Create the Environment","type":"text"}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Considerations","type":"text","marks":[{"type":"strong"}]},{"text":":","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Asking for a shared environment: Install within a ","type":"text"},{"text":"/project/","type":"text","marks":[{"type":"code"}]},{"text":" location - a project you and users you want to share with are part of.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Read the documentation!","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"The installation process offers both conda and pip install directions. Since we want a self contained environment we would suggest using the ","type":"text"},{"text":"conda install","type":"text","marks":[{"type":"code"}]},{"text":" approach. If you use ","type":"text"},{"text":"pip install","type":"text","marks":[{"type":"code"}]},{"text":" you would need to set the ","type":"text"},{"text":"PYTHONUSERBASE","type":"text","marks":[{"type":"code"}]},{"text":" environment variable.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Directions indicate how to install ","type":"text"},{"text":"cuda","type":"text","marks":[{"type":"code"}]},{"text":" version 12.4.","type":"text"}]}]}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Look at the Linux ","type":"text"},{"text":"du","type":"text","marks":[{"type":"code"}]},{"text":" command to calculate the storage a folder takes.","type":"text"}]}]}]}]},{"type":"expand","attrs":{"title":"Example process:"},"content":[{"type":"codeBlock","content":[{"text":"[~]$ cd /project//software/pytorch\n[]$ module purge\n[]$ module load miniconda3/24.3.0\n[]$ conda create -p pytorch_env\n...\n# To activate this environment, use\n#\n# $ conda activate /cluster/medbow/project//software/pytorch/pytorch_env\n#\n# To deactivate an active environment, use\n#\n# $ conda deactivate\n\n[]$ conda activate /cluster/medbow/project//software/pytorch/pytorch_env\n(/cluster/medbow/project//software/pytorch/pytorch_env) []$ conda install pytorch torchvision torchaudio pytorch-cuda=12.4 -c pytorch -c nvidia\n(/cluster/medbow/project//software/pytorch/pytorch_env) []$ conda deactivate\n[]$ ","type":"text"}]}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Regards the storage space taken up:","type":"text"}]}]},{"type":"codeBlock","content":[{"text":"[pytorch]$ du -d 1 -h\n6.3G ./pytorch_env\n6.3G ","type":"text"}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Notice within the installation of ","type":"text"},{"text":"pytorch","type":"text","marks":[{"type":"code"}]},{"text":" the following (or something similar):","type":"text"}]},{"type":"codeBlock","content":[{"text":"...\npython-3.12.3 |hab00c5b_0_cpython 30.5 MB conda-forge\n...","type":"text"}]},{"type":"paragraph","content":[{"text":"You can confirm within the activate environment:","type":"text"}]},{"type":"codeBlock","content":[{"text":"(/cluster/medbow/project//software/pytorch/pytorch_env) []$ python --version\nPython 3.12.3","type":"text"}]}]},{"type":"rule"},{"type":"heading","attrs":{"level":1},"content":[{"text":"Exercise: Suggested Answers: Run the Code","type":"text"}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Considerations","type":"text","marks":[{"type":"strong"}]},{"text":":","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Decide on your working directory where you’ll have the python test script and the bash script to submit. Do you want to share this?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Look at the required resources:","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"One H100 - which partition do you need to define? Remember: ","type":"text"},{"text":"if you don’t ask, you don’t get","type":"text","marks":[{"type":"strong"},{"type":"em"}]},{"text":".","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"How do you request 8 cores and 16G of memory?","type":"text"}]}]}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"What are all the steps you need to perform? You can test via an interactive ","type":"text"},{"text":"salloc","type":"text","marks":[{"type":"code"}]},{"text":" session, and then copy into your bash submission script.","type":"text"}]}]}]}]},{"type":"expand","attrs":{"title":"Suggested Submission Script: run.sh"},"content":[{"type":"codeBlock","content":[{"text":"#!/bin/bash\n#SBATCH --account=\n#SBATCH --time=10:00\n#SBATCH --job-name=pytorch_test\n#SBATCH --cpus-per-task=8\n#SBATCH --mem=16G\n#SBATCH --mail-type=ALL\n#SBATCH --mail-user=\n#SBATCH --output=slurms/pytorch_%A.out\n#SBATCH --partition=mb-h100\n#SBATCH --gres=gpu:1\n\nmodule purge\nmodule load miniconda3/24.3.0\n\nconda activate /cluster/medbow/project/arcc/software/pytorch/pytorch_env\npython pytorch_gpu_test.py\nconda deactivate\necho \"Done.\"","type":"text"}]}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"If successful, your output should take the following form:","type":"text"}]}]},{"type":"expand","attrs":{"title":"cat slurms/pytorch_2331610.out"},"content":[{"type":"codeBlock","content":[{"text":"The following modules were not unloaded:\n (Use \"module --force purge\" to unload all):\n 1) slurm/latest 2) arcc/1.0\n\nPyTorch Version: 2.4.0\nCuda Available: True\nDevice Count: 1\nDevice Name: NVIDIA H100 80GB HBM3\n99 176.02212524414062\n199 119.40426635742188\n299 81.9596939086914\n399 57.194664001464844\n499 40.815223693847656\n599 29.981685638427734\n699 22.816429138183594\n799 18.07718276977539\n899 14.942441940307617\n999 12.869009017944336\n1099 11.497486114501953\n1199 10.590255737304688\n1299 9.990135192871094\n1399 9.59315013885498\n1499 9.330546379089355\n1599 9.156827926635742\n1699 9.041902542114258\n1799 8.965866088867188\n1899 8.915557861328125\n1999 8.882278442382812\nResult: y = -0.0008966787136159837 + 0.8645414710044861 x + 0.0001546924322610721 x^2 + -0.09443996101617813 x^3\n\nDone.","type":"text"}]}]},{"type":"rule"},{"type":"heading","attrs":{"level":1},"content":[{"text":"Exercise: Suggested Answers: Extend with Pandas","type":"text"}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Considerations","type":"text","marks":[{"type":"strong"}]},{"text":": ","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"Can you add Pandas to the environment? Yes. You can always go back to an existing environment, activate, and update.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"How will you install this package? Using ","type":"text"},{"text":"conda install","type":"text","marks":[{"type":"code"}]},{"text":" or ","type":"text"},{"text":"pip install","type":"text","marks":[{"type":"code"}]},{"text":"?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"You’re explicitly asked for version ","type":"text"},{"text":"2.1.4","type":"text","marks":[{"type":"code"}]},{"text":". Is this version available? How can you check?","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"If you ","type":"text"},{"text":"conda install","type":"text","marks":[{"type":"code"}]},{"text":", is there anything I need to take note off during the solving stage?","type":"text"}]}]}]}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"Because we want to ","type":"text"},{"text":"try","type":"text","marks":[{"type":"em"}]},{"text":" and keep are environment self contained, we’d suggest first looking at using Conda.","type":"text"}]},{"type":"paragraph","content":[{"text":"Perform the following to check if it is available as a Conda package:","type":"text"}]},{"type":"codeBlock","content":[{"text":"(/cluster/medbow/project//software/pytorch/pytorch_env) []$ conda search pandas\nLoading channels: done\n# Name Version Build Channel\n...\npandas 2.1.4 py39hddac248_0 conda-forge\n...","type":"text"}]},{"type":"paragraph","content":[{"text":"Notice","type":"text","marks":[{"type":"strong"}]},{"text":": This package is within the ","type":"text"},{"text":"conda-forge","type":"text","marks":[{"type":"code"}]},{"text":" ","type":"text"},{"text":"channel","type":"text","marks":[{"type":"strong"}]},{"text":". Do you have this configured?","type":"text"}]}]},{"type":"panel","attrs":{"panelType":"info"},"content":[{"type":"paragraph","content":[{"text":"What is details during the solving stage of the ","type":"text"},{"text":"conda install","type":"text","marks":[{"type":"code"}]},{"text":"?","type":"text"}]},{"type":"codeBlock","content":[{"text":"(/cluster/medbow/project//software/pytorch/pytorch_env) []$ conda install pandas==2.1.4\nThe following packages will be downloaded:\n package | build\n ---------------------------|-----------------\n numpy-1.26.4 | py312heda63a1_0 7.1 MB conda-forge\n pandas-2.1.4 | py312hfb8ada1_0 14.0 MB conda-forge\n ------------------------------------------------------------\n Total: 21.1 MB\n\nThe following NEW packages will be INSTALLED:\n pandas conda-forge/linux-64::pandas-2.1.4-py312hfb8ada1_0\n ...\n\nThe following packages will be DOWNGRADED:\n numpy 2.1.0-py312h1103770_0 --> 1.26.4-py312heda63a1_0","type":"text"}]}]},{"type":"panel","attrs":{"panelType":"warning"},"content":[{"type":"paragraph","content":[{"text":"Notice","type":"text","marks":[{"type":"strong"}]},{"text":": The ","type":"text"},{"text":"numpy","type":"text","marks":[{"type":"code"}]},{"text":" package was installed as part of the original ","type":"text"},{"text":"torch","type":"text","marks":[{"type":"code"}]},{"text":" install, but is going to be ","type":"text"},{"text":"downgraded","type":"text","marks":[{"type":"strong"},{"type":"em"}]},{"text":" from ","type":"text"},{"text":"2.1.0","type":"text","marks":[{"type":"code"}]},{"text":" to ","type":"text"},{"text":"1.26.4","type":"text","marks":[{"type":"code"}]},{"text":"! ","type":"text"}]},{"type":"bulletList","content":[{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"This is a reason to ","type":"text"},{"text":"maybe not","type":"text","marks":[{"type":"em"}]},{"text":" set the ","type":"text"},{"text":"always_yes","type":"text","marks":[{"type":"code"}]},{"text":" option in the ","type":"text"},{"text":"~/.condarc","type":"text","marks":[{"type":"code"}]},{"text":" file. If you have this set to ","type":"text"},{"text":"yes","type":"text","marks":[{"type":"code"}]},{"text":", then the installation would have continued regardless.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"How does this downgraded version potentially affect torch? With out testing, we don’t know.","type":"text"}]}]},{"type":"listItem","content":[{"type":"paragraph","content":[{"text":"We would suggest not downgrading. Instead create a separate Conda environment for Pandas so we do not run into potential dependency issues.","type":"text"}]}]}]}]},{"type":"rule"},{"type":"table","attrs":{"layout":"default","width":1011.0,"localId":"4cd25e95-1600-4d5c-bb99-471fe0139aa1"},"content":[{"type":"tableRow","content":[{"type":"tableCell","attrs":{"colspan":1,"rowspan":1,"colwidth":[333.0]},"content":[{"type":"paragraph","content":[{"text":"Prev","type":"text","marks":[{"type":"strong"}]}]},{"type":"paragraph","content":[{"text":"Python, Conda and Pip: Suggested Best Practices","type":"text","marks":[{"type":"link","attrs":{"href":"https://arccwiki.atlassian.net/wiki/spaces/DOCUMENTAT/pages/2182348869"}}]}]}]},{"type":"tableCell","attrs":{"colspan":1,"rowspan":1,"colwidth":[326.0]},"content":[{"type":"paragraph","marks":[{"type":"alignment","attrs":{"align":"center"}}],"content":[{"text":"Workshop Home","type":"text","marks":[{"type":"strong"}]}]},{"type":"paragraph","marks":[{"type":"alignment","attrs":{"align":"center"}}],"content":[{"text":"Using Python, Conda and Pip on the Cluster","type":"text","marks":[{"type":"link","attrs":{"href":"https://arccwiki.atlassian.net/wiki/spaces/DOCUMENTAT/pages/2136342529"}}]}]}]},{"type":"tableCell","attrs":{"colspan":1,"rowspan":1,"colwidth":[349.0]},"content":[{"type":"paragraph","marks":[{"type":"alignment","attrs":{"align":"end"}}],"content":[{"text":"Next","type":"text","marks":[{"type":"strong"}]}]},{"type":"paragraph","marks":[{"type":"alignment","attrs":{"align":"end"}}],"content":[{"text":"Python, Conda and Pip: Summary","type":"text","marks":[{"type":"link","attrs":{"href":"https://arccwiki.atlassian.net/wiki/spaces/DOCUMENTAT/pages/2282029089"}}]}]}]}]}]},{"type":"paragraph","content":[{"text":" ","type":"text"}]}],"version":1}

Browser not supported