Delphi: Memristor Logic Synthesis Toolchain

██████╗ ███████╗██╗     ██████╗ ██╗  ██╗██╗
██╔══██╗██╔════╝██║     ██╔══██╗██║  ██║██║
██║  ██║█████╗  ██║     ██████╔╝███████║██║
██║  ██║██╔══╝  ██║     ██╔═══╝ ██╔══██║██║
██████╔╝███████╗███████╗██║     ██║  ██║██║
╚═════╝ ╚══════╝╚══════╝╚═╝     ╚═╝  ╚═╝╚═╝

A high-performance memristor-based logic synthesis toolchain with advanced scheduling and mapping capabilities.

Overview

Delphi is a modern implementation of a memristor-based logic synthesis toolchain, designed for optimizing digital circuits for implementation on memristor crossbar arrays. It converts NOT/NOR netlists into optimized crossbar configurations, providing both naive and compact mapping strategies.

Installation

Prerequisites

• Rust (1.67.0 or newer) - Install Rust
• Cargo package manager (included with Rust)

Building from Source

Windows

# Clone the repository
git clone https://github.com/CynicPixel/delphi.git
cd delphi

# Build in release mode
cargo build --release

# The executable will be available at
# .\target\release\delphi.exe

macOS/Linux

# Clone the repository
git clone https://github.com/CynicPixel/delphi.git
cd delphi

# Build in release mode
cargo build --release

# The executable will be available at
# ./target/release/delphi

Command Line Interface

USAGE:
    delphi [OPTIONS] <COMMAND>

COMMANDS:
    process     Process a single netlist file
    bench       Process an entire benchmark suite
    benchmark   Run performance comparison between sequential and parallel implementations
    help        Print this message or the help of the given subcommand(s)

OPTIONS:
    -h, --help       Print help information
    -V, --version    Print version information

Detailed Usage

Processing a Single Netlist

Windows

# Basic usage
.\delphi process <NETLIST>

# With options
.\delphi process <NETLIST> --output <DIR> --disable-parallel

macOS/Linux

# Basic usage
./delphi process <NETLIST>

# With options
./delphi process <NETLIST> --output <DIR> --disable-parallel

Options for process command:

• <NETLIST>: Path to the netlist file (required)
• -o, --output <DIR>: Output directory for results (default: ./Results)
• --disable-parallel: Disable parallel processing

Example (Windows):

# Process c17 benchmark with parallel processing
.\delphi process C:\path\to\BENCH\netlist\iscas85_c17.txt

# Process with custom output directory and sequential processing
.\delphi process C:\path\to\BENCH\netlist\iscas85_c17.txt -o .\my_results --disable-parallel

Processing a Benchmark Suite

Windows

# Basic usage
.\delphi.exe bench <DIR>

# With options
.\delphi bench <DIR> --output <OUTPUT_DIR> --pattern <PATTERN> --disable-parallel

macOS/Linux

# Basic usage
./delphi bench <DIR>

# With options
./delphi bench <DIR> --output <OUTPUT_DIR> --pattern <PATTERN> --disable-parallel

Options for bench command:

• <DIR>: Path to the benchmark directory (required)
• -o, --output <DIR>: Output directory for results (default: ./Results)
• -p, --pattern <PATTERN>: Only process files matching this pattern
• --disable-parallel: Disable parallel processing

Example (Windows):

# Process all benchmarks in a directory
.\delphi bench C:\path\to\BENCH\netlist\

# Process only iscas85 benchmarks
.\delphi bench C:\path\to\BENCH\netlist\ -p iscas85

Running Performance Benchmarks

Windows

# Basic usage
.\delphi benchmark <NETLIST>

# With options
.\delphi benchmark <NETLIST> --iterations <ITERATIONS>

macOS/Linux

# Basic usage
./delphi benchmark <NETLIST>

# With options
./delphi benchmark <NETLIST> --iterations <ITERATIONS>

Options for benchmark command:

• <NETLIST>: Path to the netlist file (required)
• -i, --iterations <ITERATIONS>: Number of iterations for accurate timing (default: 3)

Example (Windows):

# Benchmark c17 with default iterations
.\delphi benchmark C:\path\to\BENCH\netlist\iscas85_c17.txt

# Benchmark with 10 iterations
.\delphi benchmark C:\path\to\BENCH\netlist\iscas85_c17.txt -i 10

Output Files

Delphi generates several output files organized in subdirectories under the specified output directory:

Output Structure

Windows:

Results\
├── magic\
│   └── [benchmark]_magic.v       # NOR/NOT mapped Verilog module
├── micro_ins_compact\
│   └── [benchmark]_compact.txt   # Compact mapping micro-operations
├── micro_ins_naive\
│   └── [benchmark]_naive.txt     # Naive mapping micro-operations
└── schedule_stats\
    └── [benchmark]_stats.txt     # Scheduling statistics

macOS/Linux:

Results/
├── magic/
│   └── [benchmark]_magic.v       # NOR/NOT mapped Verilog module
├── micro_ins_compact/
│   └── [benchmark]_compact.txt   # Compact mapping micro-operations
├── micro_ins_naive/
│   └── [benchmark]_naive.txt     # Naive mapping micro-operations
└── schedule_stats/
    └── [benchmark]_stats.txt     # Scheduling statistics

File Descriptions

1. Magic Verilog (_magic.v)

   • NOR/NOT mapped module definition
   • Input, output, and wire declarations
   • Gate instantiations

2. Naive Mapping (_naive.txt)

   • Micro-operations for naive crossbar mapping
   • Simple, linear mapping strategy
   • Each level's operations listed separately

3. Compact Mapping (_compact.txt)

   • Micro-operations for optimized compact mapping
   • Efficiently utilizes crossbar space
   • Each level's operations listed separately

4. Schedule Statistics (_stats.txt)

   • ASAP, ALAP, and LIST scheduling metrics
   • Gate distribution across levels
   • Crossbar size and time step information
   • Performance comparisons

Viewing Output Files (Windows)

# View statistics
type Results\schedule_stats\iscas85_c17_stats.txt

# View Verilog mapping
type Results\magic\iscas85_c17_magic.v

# View micro-operations
type Results\micro_ins_naive\iscas85_c17_naive.txt

Advanced Features

Parallel Processing

Delphi supports parallel processing to speed up computations for larger circuits:

# Enable parallel processing (default)
.\delphi process <NETLIST>

# Disable parallel processing
.\delphi process <NETLIST> --disable-parallel

The parallel implementation:

• Automatically adjusts to the number of available CPU cores
• Only activates for circuits with 100+ gates
• Provides significant speedup for larger benchmarks
• Falls back to sequential processing for small circuits

Benchmark-Specific Processing

You can process specific benchmark types or patterns:

# Process only iscas85 benchmarks
.\delphi bench C:\path\to\benchmarks -p iscas85

# Process only c17 benchmark variants
.\delphi bench C:\path\to\benchmarks -p c17

Performance Benchmarking

Compare sequential and parallel implementations:

.\delphi benchmark <NETLIST> -i 5

This will:

1. Run both sequential and parallel versions multiple times
2. Measure execution time for each stage
3. Calculate speedup metrics
4. Display detailed performance comparisons

Troubleshooting

Windows-Specific Issues

Path Separators: Windows uses backslashes (\) in paths, while the command line might require either escaped backslashes (\\) or forward slashes (/).

File Access Permissions: If you encounter permission errors, try running the command prompt or PowerShell as Administrator.

Network Paths: If your files are on a network drive, use the full UNC path (e.g., \\server\share\path\to\file.txt).

Common Issues on All Platforms

File Not Found:

Error: Failed to parse netlist: No such file or directory

Solution: Check that the netlist file path is correct and accessible.

Invalid Benchmark Directory:

Error: Invalid benchmark directory

Solution: Ensure the benchmark directory exists and contains netlist files.

Parsing Errors:

Error: Failed to parse netlist: Invalid format at line X

Solution: Verify that the netlist file follows the required format for NOT/NOR gates.

Examples

Basic Processing (Windows)

# Process a single benchmark
.\delphi process C:\path\to\BENCH\netlist\iscas85_c17.txt

# View the statistics
type Results\schedule_stats\iscas85_c17_stats.txt

# View the Verilog mapping
type Results\magic\iscas85_c17_magic.v

Batch Processing with Custom Output (Windows)

# Create output directory
mkdir my_results

# Process all iscas benchmarks
.\delphi bench C:\path\to\BENCH\netlist -p iscas -o my_results

Performance Comparison (Windows)

# Detailed benchmark with 5 iterations
.\delphi benchmark C:\path\to\BENCH\netlist\iscas85_c7552.txt -i 5

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

For more information or issues, please open an issue on the GitHub repository.