sarman/dgx-spark-playbooks
1
0
Fork 0
mirror of https://github.com/NVIDIA/dgx-spark-playbooks.git synced 2026-04-23 10:33:51 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

chore: Regenerate all playbooks

Browse Source
This commit is contained in:
GitLab CI 2025-12-15 20:19:01 +00:00
parent 5472c97a8c
commit b4e7892d2c
2 changed files with 15 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
nvidia/nccl/README.md
View File

@ -44,14 +44,14 @@ and proper GPU topology detection.
* **Duration**: 30 minutes for setup and validation * **Duration**: 30 minutes for setup and validation
* **Risk level**: Medium - involves network configuration changes * **Risk level**: Medium - involves network configuration changes
* **Rollback**: The NCCL & NCCL Tests repositories can be deleted from DGX Spark * **Rollback**: The NCCL & NCCL Tests repositories can be deleted from DGX Spark
* **Last Updated:** 10/12/2025 * **Last Updated:** 12/15/2025
* First publication * Use nccl latest version v2.28.9-1
## Run on two Sparks ## Run on two Sparks
## Step 1. Configure network connectivity ## Step 1. Configure network connectivity
Follow the network setup instructions from the Connect two Sparks playbook to establish connectivity between your DGX Spark nodes. Follow the network setup instructions from the [Connect two Sparks](https://build.nvidia.com/spark/connect-two-sparks/stacked-sparks) playbook to establish connectivity between your DGX Spark nodes.
This includes: This includes:
- Physical QSFP cable connection - Physical QSFP cable connection
@ -67,7 +67,7 @@ architecture support:
```bash ```bash
## Install dependencies and build NCCL ## Install dependencies and build NCCL
sudo apt-get update && sudo apt-get install -y libopenmpi-dev sudo apt-get update && sudo apt-get install -y libopenmpi-dev
git clone -b v2.28.3-1 https://github.com/NVIDIA/nccl.git ~/nccl/ git clone -b v2.28.9-1 https://github.com/NVIDIA/nccl.git ~/nccl/
cd ~/nccl/ cd ~/nccl/
make -j src.build NVCC_GENCODE="-gencode=arch=compute_121,code=sm_121" make -j src.build NVCC_GENCODE="-gencode=arch=compute_121,code=sm_121"
@ -80,7 +80,7 @@ export LD_LIBRARY_PATH="$NCCL_HOME/lib:$CUDA_HOME/lib64/:$MPI_HOME/lib:$LD_LIBRA
## Step 3. Build NCCL test suite ## Step 3. Build NCCL test suite
Compile the NCCL test suite to validate communication performance: Compile the NCCL test suite on **both nodes**:
```bash ```bash
## Clone and build NCCL tests ## Clone and build NCCL tests
@ -91,7 +91,7 @@ make MPI=1
## Step 4. Find the active network interface and IP addresses ## Step 4. Find the active network interface and IP addresses
Execute multi-node NCCL performance test using the active network interface. First, identify which network ports are available and up: First, identify which network ports are available and up:
```bash ```bash
## Check network port status ## Check network port status

25
nvidia/unsloth/README.md
View File

@ -55,9 +55,8 @@ The Python test script can be found [here on GitHub](https://github.com/NVIDIA/d
* CUDA toolkit configuration issues may prevent kernel compilation * CUDA toolkit configuration issues may prevent kernel compilation
* Memory constraints on smaller models require batch size adjustments * Memory constraints on smaller models require batch size adjustments
* **Rollback**: Uninstall packages with `pip uninstall unsloth torch torchvision`. * **Rollback**: Uninstall packages with `pip uninstall unsloth torch torchvision`.
* **Last Updated:** 11/07/2025 * **Last Updated:** 12/15/2025
* Add required python dependencies * Upgrade pytorch container and python dependencies to the latest version
* Fix broken commands to access files on GitHub
## Instructions ## Instructions
@ -77,28 +76,22 @@ The output should show a summary of GPU information.
## Step 2. Get the container image ## Step 2. Get the container image
```bash ```bash
docker pull nvcr.io/nvidia/pytorch:25.09-py3 docker pull nvcr.io/nvidia/pytorch:25.11-py3
``` ```
## Step 3. Launch Docker ## Step 3. Launch Docker
```bash ```bash
docker run --gpus all --ulimit memlock=-1 -it --ulimit stack=67108864 --entrypoint /usr/bin/bash --rm nvcr.io/nvidia/pytorch:25.09-py3 docker run --gpus all --ulimit memlock=-1 -it --ulimit stack=67108864 --entrypoint /usr/bin/bash --rm nvcr.io/nvidia/pytorch:25.11-py3
``` ```
## Step 4. Install dependencies inside Docker ## Step 4. Install dependencies inside Docker
```bash ```bash
pip install transformers peft "datasets==4.3.0" "trl==0.19.1" pip install transformers peft hf_transfer "datasets==4.3.0" "trl==0.26.1"
pip install --no-deps unsloth unsloth_zoo pip install --no-deps unsloth unsloth_zoo bitsandbytes
pip install hf_transfer
``` ```
## Step 5. Build and install bitsandbytes inside Docker ## Step 5. Create Python test script
```bash
pip install --no-deps bitsandbytes
```
## Step 6. Create Python test script
Curl the test script [here](https://github.com/NVIDIA/dgx-spark-playbooks/blob/main/nvidia/unsloth/assets/test_unsloth.py) into the container. Curl the test script [here](https://github.com/NVIDIA/dgx-spark-playbooks/blob/main/nvidia/unsloth/assets/test_unsloth.py) into the container.
@ -109,7 +102,7 @@ curl -O https://raw.githubusercontent.com/NVIDIA/dgx-spark-playbooks/refs/heads/
We will use this test script to validate the installation with a simple fine-tuning task. We will use this test script to validate the installation with a simple fine-tuning task.
## Step 7. Run the validation test ## Step 6. Run the validation test
Execute the test script to verify Unsloth is working correctly. Execute the test script to verify Unsloth is working correctly.
@ -122,7 +115,7 @@ Expected output in the terminal window:
- Training progress bars showing loss decreasing over 60 steps - Training progress bars showing loss decreasing over 60 steps
- Final training metrics showing completion - Final training metrics showing completion
## Step 8. Next steps ## Step 7. Next steps
Test with your own model and dataset by updating the `test_unsloth.py` file: Test with your own model and dataset by updating the `test_unsloth.py` file: