Update installation and usage documentation for TritonParse

Summary:
- Revised prerequisites in README.md and installation guide to specify GPU requirements for Triton.
- Enhanced installation instructions for both NVIDIA and AMD GPUs, including PyTorch installation steps.
- Updated usage examples to dynamically select device based on GPU availability.
- Clarified FAQ section regarding GPU necessity for generating traces.

These changes improve clarity and ensure users are well-informed about the hardware requirements for TritonParse.

Author: FindHao

README.md

@@ -68,7 +68,7 @@ cd tritonparse
 pip install -e .
 ```
 
-**Prerequisites:** Python ≥ 3.10, Triton > 3.3.1 ([install from source](https://github.com/triton-lang/triton))
+**Prerequisites:** Python ≥ 3.10, Triton > 3.3.1 ([install from source](https://github.com/triton-lang/triton)), GPU required (NVIDIA/AMD)
 
 ## 📚 Complete Documentation
 

docs/wiki-pages/01.-Installation.md

@@ -20,20 +20,20 @@ This guide covers all installation scenarios for TritonParse, from basic usage t
 ### System Requirements
 - **Python** >= 3.10
 - **Operating System**: Linux, macOS, or Windows (with WSL recommended)
-- **CUDA** (for GPU tracing): Compatible NVIDIA GPU with CUDA support
+- **GPU Required** (Triton depends on GPU):
+  - **NVIDIA GPUs**: CUDA 11.8+ or 12.x
+  - **AMD GPUs**: ROCm 5.0+ (supports MI100, MI200, MI300 series)
 - **Node.js** >= 18.0.0 (for website development only)
 
-### Required: Triton Installation
-**Important**: You need Triton > 3.3.1 or compiled from source.
+> ⚠️ **Important**: GPU is required to generate traces because Triton kernels can only run on GPU hardware. The web interface can view existing traces without GPU.
 
-```bash
-# Install Triton from source (required)
-git clone https://github.com/triton-lang/triton.git
-cd triton
-pip install -e .
-```
+### Required Dependencies
+- **PyTorch** with GPU support (we recommend PyTorch nightly for best compatibility)
+  - For NVIDIA GPUs: PyTorch with CUDA support
+  - For AMD GPUs: PyTorch with ROCm support
+- **Triton** > 3.3.1 (must be compiled from source for TritonParse compatibility)
 
-For detailed Triton installation instructions, see the [official Triton documentation](https://github.com/triton-lang/triton?tab=readme-ov-file#install-from-source).
+> 💡 **Note**: Detailed installation instructions for these dependencies are provided in each installation option below.
 
 ---
 
@@ -47,16 +47,63 @@ git clone https://github.com/pytorch-labs/tritonparse.git
 cd tritonparse
 ```
 
-### Step 2: Install TritonParse
+### Step 2: Install PyTorch with GPU Support
+
+#### For NVIDIA GPUs (CUDA)
+```bash
+# Install PyTorch nightly with CUDA 12.8 support (recommended)
+pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128
+
+# Alternative: Install stable PyTorch with CUDA support
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
+
+# Verify PyTorch installation
+python -c "import torch; print(f'PyTorch version: {torch.__version__}')"
+python -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}')"
+```
+
+#### For AMD GPUs (ROCm)
+```bash
+# Install PyTorch nightly with ROCm support (recommended)
+pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.2
+
+# Alternative: Install stable PyTorch with ROCm support
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.1
+
+# Verify PyTorch installation
+python -c "import torch; print(f'PyTorch version: {torch.__version__}')"
+python -c "import torch; print(f'ROCm available: {torch.cuda.is_available()}')"
+python -c "import torch; print(f'GPU device: {torch.cuda.get_device_name(0) if torch.cuda.is_available() else \"No GPU\"}')"
+```
+
+### Step 3: Install Triton from Source
 ```bash
+# First, uninstall any existing PyTorch-bundled Triton
+pip uninstall -y pytorch-triton triton || true
+
+# Install Triton from source (required)
+git clone https://github.com/triton-lang/triton.git
+cd triton
+pip install -e .
+
+# Verify Triton installation
+python -c "import triton; print(f'Triton version: {triton.__version__}')"
+python -c "import triton; print(f'Triton path: {triton.__file__}')"
+```
+
+### Step 4: Install TritonParse
+```bash
+# Go back to tritonparse directory
+cd ../tritonparse
+
 # Install in development mode
 pip install -e .
 
 # Or install with test dependencies
 pip install -e ".[test]"
 ```
 
-### Step 3: Verify Installation
+### Step 5: Verify Installation
 ```bash
 # Test with the included example
 cd tests
@@ -85,7 +132,7 @@ INFO:tritonparse:Copying parsed logs from /tmp/tmp1gan7zky to /scratch/findhao/t
 ================================================================================
 ```
 
-### Step 4: Use the Web Interface
+### Step 6: Use the Web Interface
 1. Generate trace files using the Python API
 2. Visit [https://pytorch-labs.github.io/tritonparse/](https://pytorch-labs.github.io/tritonparse/)
 3. Load your trace files (.ndjson or .gz format)
@@ -101,7 +148,7 @@ For contributors working on the React-based web interface.
 - npm (comes with Node.js)
 
 ### Step 1: Basic Installation
-Follow [Option 1](#option-1-basic-user-installation) first.
+Follow [Option 1: Basic User Installation](#-option-1-basic-user-installation) first to install PyTorch, Triton, and TritonParse.
 
 ### Step 2: Install Website Dependencies
 ```bash
@@ -141,7 +188,7 @@ npm run preview
 For core contributors working on Python code, including formatting and testing.
 
 ### Step 1: Basic Installation
-Follow [Option 1](#option-1-basic-user-installation) first.
+Follow [Option 1: Basic User Installation](#-option-1-basic-user-installation) first to install PyTorch, Triton, and TritonParse.
 
 ### Step 2: Install Development Dependencies
 ```bash
@@ -167,47 +214,7 @@ python -m unittest tests.test_tritonparse -v
 ```
 
 ### Step 4: Website Development (Optional)
-```bash
-cd website
-npm install
-npm run dev
-```
-
----
-
-## 🛠️ Development Commands
-
-### Python Development
-```bash
-# Format code
-make format
-
-# Check formatting
-make format-check
-
-# Run linting
-make lint-check
-
-# Run tests
-python -m unittest tests.test_tritonparse -v
-
-# Run specific test
-python -m unittest tests.test_tritonparse.TestTritonparseCUDA.test_whole_workflow -v
-```
-
-### Website Development
-```bash
-cd website
-
-# Development server
-npm run dev
-
-# Build for production
-npm run build
-
-# Lint frontend code
-npm run lint
-```
+If you also need to work on the web interface, follow [Option 2: Website Development Setup](#-option-2-website-development-setup) for additional setup.
 
 ---
 
@@ -218,20 +225,34 @@ npm run lint
 #### 1. Triton Installation Issues
 ```bash
 # Error: "No module named 'triton'"
-# Solution: Install Triton from source
+# Solution: Uninstall existing Triton and install from source
+pip uninstall -y pytorch-triton triton || true
 git clone https://github.com/triton-lang/triton.git
 cd triton
 pip install -e .
 ```
 
-#### 2. CUDA Not Available
+#### 2. GPU Not Available
 ```bash
-# Error: "CUDA not available"
-# Check CUDA installation
-python -c "import torch; print(torch.cuda.is_available())"
-
-# If False, install CUDA-enabled PyTorch
+# Error: "CUDA not available" or "ROCm not available"
+# Check GPU installation
+python -c "import torch; print(f'GPU available: {torch.cuda.is_available()}')"
+python -c "import torch; print(f'Device count: {torch.cuda.device_count()}')"
+
+# For NVIDIA GPUs - Install CUDA-enabled PyTorch
+pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128
+# Alternative: stable version
 pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
+
+# For AMD GPUs - Install ROCm-enabled PyTorch
+pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.2
+# Alternative: stable version
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.1
+
+# Verify GPU support after installation
+python -c "import torch; print(f'GPU available: {torch.cuda.is_available()}')"
+python -c "import torch; print(f'GPU device: {torch.cuda.get_device_name(0) if torch.cuda.is_available() else \"No GPU\"}')"
+python -c "import torch; print(f'Backend version: {torch.version.cuda if torch.version.cuda else torch.version.hip}')"
 ```
 
 #### 3. Permission Issues
@@ -270,20 +291,23 @@ npm install
 
 Set these for development:
 ```bash
-# Enable debug logging
-export TRITONPARSE_DEBUG=1
-
-# Enable NDJSON output (default)
-export TRITONPARSE_NDJSON=1
-
-# Enable gzip compression
-export TRITON_TRACE_GZIP=1
-
-# Custom trace directory
-export TRITON_TRACE=/path/to/traces
-
-# Disable FX graph cache (for testing)
-export TORCHINDUCTOR_FX_GRAPH_CACHE=0
+# TritonParse specific
+export TRITONPARSE_DEBUG=1                    # Enable debug logging
+export TRITONPARSE_NDJSON=1                  # Enable NDJSON output (default)
+export TRITON_TRACE_GZIP=1                   # Enable gzip compression
+export TRITON_TRACE=/path/to/traces          # Custom trace directory
+
+# PyTorch/TorchInductor related
+export TORCHINDUCTOR_FX_GRAPH_CACHE=0        # Disable FX graph cache (for testing)
+export TORCH_LOGS="+dynamo,+inductor"        # Enable PyTorch debug logs
+export CUDA_VISIBLE_DEVICES=0                # Limit to specific NVIDIA GPU
+export ROCR_VISIBLE_DEVICES=0                # Limit to specific AMD GPU (ROCm)
+export HIP_VISIBLE_DEVICES=0                 # Alternative for AMD GPUs
+
+# GPU debugging (if needed)
+export CUDA_LAUNCH_BLOCKING=1                # Synchronous CUDA execution (NVIDIA)
+export HIP_LAUNCH_BLOCKING=1                 # Synchronous HIP execution (AMD)
+export TORCH_USE_CUDA_DSA=1                  # Enable CUDA device-side assertions (NVIDIA)
 ```
 
 ### Getting Help
@@ -306,10 +330,22 @@ After installation, verify everything works:
 
 ### 1. Python API Test
 ```python
+# Test PyTorch installation
+import torch
+print(f"PyTorch version: {torch.__version__}")
+print(f"GPU available: {torch.cuda.is_available()}")
+print(f"GPU device count: {torch.cuda.device_count()}")
+if torch.cuda.is_available():
+    print(f"GPU device: {torch.cuda.get_device_name(0)}")
+    print(f"Backend version: {torch.version.cuda if torch.version.cuda else torch.version.hip}")
+
+# Test Triton installation
+import triton
+print(f"Triton version: {triton.__version__}")
+
+# Test TritonParse installation
 import tritonparse.structured_logging
 import tritonparse.utils
-
-# Should not raise any errors
 print("TritonParse installed successfully!")
 ```
 

docs/wiki-pages/02.-Usage-Guide.md

@@ -76,8 +76,9 @@ def tensor_add(a, b):
 # Example usage
 if __name__ == "__main__":
     # Create test tensors
-    a = torch.randn(1024, 1024, device="cuda", dtype=torch.float32)
-    b = torch.randn(1024, 1024, device="cuda", dtype=torch.float32)
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    a = torch.randn(1024, 1024, device=device, dtype=torch.float32)
+    b = torch.randn(1024, 1024, device=device, dtype=torch.float32)
 
     # Execute kernel (this will be traced)
     c = tensor_add(a, b)
@@ -110,8 +111,9 @@ def simple_add(a, b):
 compiled_add = torch.compile(simple_add)
 
 # Create test data
-a = torch.randn(1024, 1024, device="cuda", dtype=torch.float32)
-b = torch.randn(1024, 1024, device="cuda", dtype=torch.float32)
+device = "cuda"
+a = torch.randn(1024, 1024, device=device, dtype=torch.float32)
+b = torch.randn(1024, 1024, device=device, dtype=torch.float32)
 
 # Execute compiled function (this will be traced)
 result = compiled_add(a, b)

docs/wiki-pages/06.-FAQ.md

@@ -26,7 +26,9 @@ This page addresses common questions and issues encountered when using TritonPar
 **A:** 
 - **Python** >= 3.10
 - **Triton** > 3.3.1 (must be compiled from source)
-- **CUDA** (for GPU tracing) - compatible NVIDIA GPU
+- **GPU Support** (for GPU tracing):
+  - **NVIDIA GPUs**: CUDA 11.8+ or 12.x
+  - **AMD GPUs**: ROCm 5.0+ (MI100, MI200, MI300 series)
 - **Modern browser** (Chrome 90+, Firefox 88+, Safari 14+, Edge 90+)
 
 ## 🔧 Installation and Setup
@@ -36,12 +38,16 @@ This page addresses common questions and issues encountered when using TritonPar
 **A:** Triton must be compiled from source for TritonParse to work:
 
 ```bash
+# First, uninstall any existing PyTorch-bundled Triton
+pip uninstall -y pytorch-triton triton || true
+
+# Install Triton from source
 git clone https://github.com/triton-lang/triton.git
 cd triton
 pip install -e .
 ```
 
-For detailed instructions, see the [Triton installation guide](https://github.com/triton-lang/triton?tab=readme-ov-file#install-from-source).
+For detailed instructions, see our [Installation Guide](01.-Installation) or the [official Triton installation guide](https://github.com/triton-lang/triton?tab=readme-ov-file#install-from-source).
 
 ### Q: I'm getting "No module named 'triton'" errors. What's wrong?
 
@@ -50,12 +56,13 @@ For detailed instructions, see the [Triton installation guide](https://github.co
 2. **Wrong Python environment** - Make sure you're in the right virtual environment
 3. **Installation failed** - Check for compilation errors during Triton installation
 
-### Q: Do I need CUDA to use TritonParse?
+### Q: Do I need a GPU to use TritonParse?
 
-**A:** 
-- **For CPU-only analysis**: No CUDA needed
-- **For GPU kernel tracing**: Yes, CUDA is required
-- **For web interface**: No CUDA needed (just to view existing traces)
+**A:** Yes, a GPU is required because Triton itself depends on GPU:
+- **For generating traces**: GPU is required (either NVIDIA with CUDA or AMD with ROCm)
+- **For web interface only**: No GPU needed (just to view existing trace files from others)
+
+Note: Triton kernels can only run on GPU, so you need GPU hardware to generate your own traces.
 
 ## 📊 Generating Traces
 

docs/wiki-pages/Home.md

@@ -112,4 +112,4 @@ This project is licensed under the BSD-3 License. See the [LICENSE](https://gith
 
 ---
 
-**Note**: This tool is designed for developers working with Triton kernels and GPU computing. Basic familiarity with CUDA, GPU programming concepts, and the Triton language is recommended for effective use. 
+**Note**: This tool is designed for developers working with Triton kernels and GPU computing. Basic familiarity with GPU programming concepts (CUDA for NVIDIA or ROCm/HIP for AMD), and the Triton language is recommended for effective use.