paddle 1.2.2__tar.gz → 1.2.4__tar.gz

{paddle-1.2.2 → paddle-1.2.4}/.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 1.2.2
+current_version = 1.2.4
 commit = True
 tag = True
 

paddle-1.2.4/Makefile

@@ -0,0 +1,111 @@
+ENV_FILE := .env
+UID := $$(id -u)
+GID := $$(id -g)
+GITCONFIG := $${HOME}/.gitconfig
+GITCREDENTIALS := $${HOME}/.git-credentials
+DATE_STRING := $(shell date "+%Y-%m-%d")
+JOB := canoe$(date +%Y%m%d_%H%M%S)
+
+.PHONY: help env up down ps start build deploy finish log status mint upload node resource
+
+# Show help for each target
+help: ## Show this help message
+	@echo "Available commands:"
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | \
+		awk 'BEGIN {FS = ":.*?## "}; {printf "  \033[36m%-12s\033[0m %s\n", $$1, $$2}'
+
+env: ## Generate the .env file with git configs first, then user info
+	@# Check if either variable is missing
+	@if ! grep -q '^GITCONFIG=' $(ENV_FILE) 2>/dev/null || \
+	    ! grep -q '^GITCREDENTIALS=' $(ENV_FILE) 2>/dev/null; then \
+		if [ -f "$${HOME}/.gitconfig" ]; then \
+			echo "GITCONFIG=$${HOME}/.gitconfig" > $(ENV_FILE); \
+		else \
+			touch "$${HOME}/.gitconfig"; \
+			echo "GITCONFIG=$${HOME}/.gitconfig" > $(ENV_FILE); \
+		fi; \
+		if [ -f "$${HOME}/.git-credentials" ]; then \
+			echo "GITCREDENTIALS=$${HOME}/.git-credentials" >> $(ENV_FILE); \
+		else \
+			touch "$${HOME}/.git-credentials"; \
+			echo "[credential]" >> "$${HOME}/.gitconfig"; \
+			echo "    helper = store" >> "$${HOME}/.gitconfig"; \
+			echo "GITCREDENTIALS=$${HOME}/.git-credentials" >> $(ENV_FILE); \
+		fi; \
+		echo "USER=$$(id -un)"    >> $(ENV_FILE); \
+		echo "USER_UID=$$(id -u)" >> $(ENV_FILE); \
+		echo "USER_GID=$$(id -g)" >> $(ENV_FILE); \
+		echo "Created $(ENV_FILE):"; cat $(ENV_FILE); \
+	fi
+
+up: env ## Start docker containers in the background
+	@docker compose up -d
+	@docker compose exec dev bash -c '\
+		USER_HOME=$$(eval echo ~$$USERNAME); \
+		if [ -f /etc/git-credentials ] && [ ! -f $$USER_HOME/.git-credentials ]; then \
+			cp /etc/git-credentials $$USER_HOME/.git-credentials; \
+			chown $${USER_UID}:$${USER_GID} $$USER_HOME/.git-credentials; \
+			git config --global credential.helper store; \
+		fi \
+	'
+
+down: env ## Stop and remove docker containers
+	@docker compose down
+
+ps: env ## Show container status
+	@docker compose ps
+
+start: env ## Open a bash shell inside the 'dev' container as the host user, exit without error
+	@docker compose exec --user $(UID):$(GID) dev \
+		bash -c 'git config --global --add safe.directory /paddle; source /opt/venv/bin/activate; exec bash'
+
+build: env ## Build (or rebuild) the 'dev' container and start it
+	@docker compose up -d --build dev
+
+deploy: env ## Deploy a multi-node job to cluster
+	@USER_UID="$$(id -u)" USER_GID="$$(id -g)" docker stack deploy -c deploy.yaml ${JOB}
+	@echo -e "\033[32m[OK]\033[0m ${JOB} deployed"
+
+finish: ## Clean up the deployed job
+	@docker stack rm ${JOB}
+
+log: ## Show the job log file
+	docker service logs ${JOB}_captain
+
+status: ## Show the job status
+	docker service ps ${JOB}_captain
+
+mint: ## Mint the current environment
+	# Remove any git credential files from the dev container before snapshotting
+	docker exec paddle-dev-1 bash -lc 'rm -f /etc/git-credentials /root/.git-credentials /home/*/.git-credentials 2>/dev/null || true'
+	docker commit paddle-dev-1 ubuntu22.04-cuda12.9-py3.10-canoe:latest
+	docker tag ubuntu22.04-cuda12.9-py3.10-canoe:latest docker.io/luminoctum/ubuntu22.04-cuda12.9-py3.10-canoe:${DATE_STRING}
+
+upload: ## Upload the minted image to docker hub
+	# Refuse to push if the image still contains git credential files
+	docker run --rm docker.io/luminoctum/ubuntu22.04-cuda12.9-py3.10-canoe:${DATE_STRING} sh -c '\
+		if ls /home/*/.git-credentials >/dev/null 2>&1; then \
+			echo "Refusing to push image: git credential files detected inside the image." >&2; \
+			exit 1; \
+		fi \
+	'
+	docker push docker.io/luminoctum/ubuntu22.04-cuda12.9-py3.10-canoe:${DATE_STRING}
+
+node: ## Show nodes in cluster
+	docker node ls
+
+resource: ## Print out cluster resource
+	@printf "%-35s %-10s %-15s %-10s\n" "NODE" "CPUs" "MEMORY (GB)" "GPUs"
+	@printf "%-35s %-10s %-15s %-10s\n" "----" "----" "-----------" "----"
+	@for node in $$(docker node ls --format "{{.Hostname}}"); do \
+		RESOURCES=$$(docker node inspect $$node --format ' \
+			{{.Description.Resources.NanoCPUs}} \
+			{{.Description.Resources.MemoryBytes}} \
+			{{if .Description.Resources.GenericResources}}{{range .Description.Resources.GenericResources}}{{if .NamedResourceSpec}}{{.NamedResourceSpec.Kind}} {{else if .DiscreteResourceSpec}}{{.DiscreteResourceSpec.Kind}} {{end}}{{end}}{{else}}0{{end}}'); \
+		\
+		CPUS=$$(echo $$RESOURCES | awk '{print $$1 / 1000000000}'); \
+		MEM=$$(echo $$RESOURCES | awk '{print $$2 / 1024 / 1024 / 1024}'); \
+		GPU_COUNT=$$(echo $$RESOURCES | grep -o "gpu" | wc -l); \
+		\
+		printf "%-35s %-10s %-15.2f %-10s\n" $$node $$CPUS $$MEM $$GPU_COUNT; \
+	done

{paddle-1.2.2 → paddle-1.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: paddle
-Version: 1.2.2
+Version: 1.2.4
 Summary: Python Atmospheric Dynamics: Discovery and Learning about Exoplanets. An open-source, user-friendly python frontend of canoe
 Project-URL: Homepage, https://github.com/elijah-mullens/paddle
 Project-URL: Repository, https://github.com/elijah-mullens/paddle
@@ -158,3 +158,23 @@ Follow the steps below to set up your development environment.
 ## Troubleshooting
 
 1. If you have Docker installed but do not have Docker Compose, remove your current Docker installation, which could be docker or docker.io, and re-install it following the guide provided in the [Develop with Docker](#develop-with-docker) section above.
+
+2. If you run out of disk space while building the image. You can relocate the default location for container images:
+Here is a simple recipe:
+
+    1. Generate config (if you don’t already have one):
+    ```
+    sudo mkdir -p /etc/containerd
+    containerd config default | sudo tee /etc/containerd/config.toml >/dev/null
+    ```
+
+    2. Edit `/etc/containerd/config.toml` and set:
+    ```
+    root = "/home/containerd"   # location where you have space
+    state = "/run/containerd"
+    ```
+
+    3. Restart
+    ```
+    sudo systemctl restart containerd
+    ```

{paddle-1.2.2 → paddle-1.2.4}/README.md

@@ -127,3 +127,23 @@ Follow the steps below to set up your development environment.
 ## Troubleshooting
 
 1. If you have Docker installed but do not have Docker Compose, remove your current Docker installation, which could be docker or docker.io, and re-install it following the guide provided in the [Develop with Docker](#develop-with-docker) section above.
+
+2. If you run out of disk space while building the image. You can relocate the default location for container images:
+Here is a simple recipe:
+
+    1. Generate config (if you don’t already have one):
+    ```
+    sudo mkdir -p /etc/containerd
+    containerd config default | sudo tee /etc/containerd/config.toml >/dev/null
+    ```
+
+    2. Edit `/etc/containerd/config.toml` and set:
+    ```
+    root = "/home/containerd"   # location where you have space
+    state = "/run/containerd"
+    ```
+
+    3. Restart
+    ```
+    sudo systemctl restart containerd
+    ```

paddle-1.2.4/deploy.yaml

@@ -0,0 +1,83 @@
+version: "3.8"
+
+networks:
+  raynet:
+    driver: overlay
+    attachable: true
+
+services:
+  captain:
+    image: luminoctum/ubuntu22.04-cuda12.9-py3.10-canoe:2026-02-03
+    entrypoint: ["/bin/bash"]
+    user: "${USER_UID}:${USER_GID}"
+    networks:
+      - raynet
+    # args: run_command, master_addr, node_rank, proc_per_node, nodes
+    command: ['/work/run.sh', 'captain', '0', '1', '2']
+    deploy:
+      restart_policy:
+        condition: none
+      resources:
+        reservations:
+          generic_resources:
+            - discrete_resource_spec:
+                kind: gpu
+                value: 1
+      replicas: 1
+      placement:
+        constraints:
+          - node.labels.role == captain
+    environment:
+      - USERNAME=${USER}
+      - USER_UID=${USER_UID}
+      - USER_GID=${USER_GID}
+      - NVIDIA_VISIBLE_DEVICES=all
+      #- NCCL_DEBUG=INFO
+      - NCCL_SOCKET_IFNAME=eth0
+      #- NCCL_ASYNC_ERROR_HANDLING=1
+    volumes:
+      - .:/work
+      #- nfs:/data
+    working_dir: /work
+
+  crew1:
+    image: luminoctum/ubuntu22.04-cuda12.9-py3.10-canoe:2026-02-03
+    entrypoint: ["/bin/bash"]
+    user: "${USER_UID}:${USER_GID}"
+    networks:
+      - raynet
+    # args: run_command, master_addr, node_rank, proc_per_node, nodes
+    command: ['/work/run.sh', 'captain', '1', '1', '2']
+    deploy:
+      restart_policy:
+        condition: none
+      resources:
+        reservations:
+          generic_resources:
+            - discrete_resource_spec:
+                kind: gpu
+                value: 1
+      replicas: 1
+      placement:
+        constraints:
+          - node.labels.role == crew
+    environment:
+      - USERNAME=${USER}
+      - USER_UID=${USER_UID}
+      - USER_GID=${USER_GID}
+      - NVIDIA_VISIBLE_DEVICES=all
+      #- NCCL_DEBUG=INFO
+      - NCCL_SOCKET_IFNAME=eth0
+      #- NCCL_ASYNC_ERROR_HANDLING=1
+    volumes:
+      - .:/work
+      #- nfs:/data
+    working_dir: /work
+
+volumes:
+  nfs:
+    driver: local
+    driver_opts:
+      type: "nfs"
+      o: "addr=coe-chengcli.turbo.storage.umich.edu,nfsvers=3,rw,nolock,soft"
+      device: ":/coe-chengcli"

{paddle-1.2.2 → paddle-1.2.4}/docker/Dockerfile.cuda

@@ -24,8 +24,11 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
       ripgrep fd-find unzip neovim \
       clangd \
       nodejs npm \
+      libnetcdf-dev nco ncview \
       && rm -rf /var/lib/apt/lists/*
 
+RUN apt-get install libnccl2 libnccl-dev
+
 # Make fd available as `fd` (Ubuntu calls it fdfind)
 RUN update-alternatives --install /usr/bin/fd fd /usr/bin/fdfind 50
 

paddle-1.2.4/docker/requirements.txt

@@ -0,0 +1,7 @@
+netcdf4
+ninja
+numpy
+paddle
+pre-commit
+ray
+scipy

paddle-1.2.4/docker-compose.override.yaml.tmp1

@@ -0,0 +1,19 @@
+services:
+  dev:
+    image: paddle
+    build:
+      dockerfile: docker/Dockerfile.cuda
+      args:
+        CUDA_VER: "12.9.1"
+    gpus: all
+    volumes:
+      - /home/${USER}/Model:/Model
+      - dart_nfs:/data
+
+volumes:
+  dart_nfs:
+    driver: local
+    driver_opts:
+      type: "nfs"
+      o: "addr=coe-chengcli.turbo.storage.umich.edu,nfsvers=3,rw"
+      device: ":/coe-chengcli"