AnalogCoder: Analog Circuit Design via Training-Free Code Generation

The code implement for paper AnalogCoder: Analog Circuit Design via Training-Free Code Generation.

Introduction

Analog circuit design is a significant task in modern chip technology, focusing on selecting component types, connectivity, and parameters to ensure proper circuit functionality. Despite advances made by Large Language Models (LLMs) in digital circuit design, the complexity and scarcity of data in analog circuitry pose significant challenges. To mitigate these issues, we introduce AnalogCoder, the first training-free LLM agent for designing analog circuits that converts tasks into Python code generation.

Main advantages of AnalogCoder:

• AnalogCoder features a feedback-enhanced flow with crafted domain-specific prompts, enabling effective and automated design of analog circuits with a high success rate.
• It proposes a circuit skill library to archive successful designs as reusable modular sub-circuits, simplifying composite circuit creation.
• Extensive testing on a custom-designed benchmark of 24 analog circuit design tasks of varying difficulty shows that AnalogCoder successfully designed 20 circuits, outperforming existing methods.

In summary, AnalogCoder can significantly improve the labor-intensive chip design process, enabling non-experts to efficiently design analog circuits.

Evaluation of LLMs

LLM Model	Avg. Pass@1	Avg. Pass@5	# of Solved
Llama2-7B	0.0	0.0	0
Llama2-13B	0.0	0.0	0
Llama3-8B	0.1	0.7	1
CodeLlama-13B	0.6	2.5	2
Mistral-7B	3.3	7.7	2
Qwen-1.5-110B	0.3	1.4	2
Llama 2-70B	5.1	9.8	3
CodeLlama-7B	2.4	9.0	4
CodeLlama-34B	1.9	7.4	4
QwenCode-7B	1.1	5.6	4
DeepSeek-Coder-33B	4.0	10.2	4
Mixtral-8×7B	5.6	12.4	5
CodeLlama-70B	3.2	12.2	7
WizardCoder-33B	7.1	16.9	7
GPT-3.5 (w/o context)	8.1	18.5	7
GPT-3.5 (w/o CoT)	19.4	26.3	8
GPT-3.5 (w/o flow)	12.8	25.3	8
Codestral-22B	16.4	29.1	8
GPT-3.5 (SPICE)	13.9	26.9	9
GPT-3.5	21.4	35.0	10
GPT-3.5 (fine-tune)	28.1	39.6	10
Llama3-70B	28.8	36.4	11
DeepSeek-V2	38.6	44.3	13
GPT-4o (w/o tool)	54.2	58.9	15
AnalogCoder	66.1	75.9	20

Note:

1. All our results are reproducible.
2. The configuration of the environment does not require sudo privileges.

Installation

AnalogCoder requires Python ≥ 3.10, PySpice ≥ 1.5, and openai >= 1.16.1.

Python Install

git clone https://github.com/anonyanalog/AnalogCoder
conda env create -f environment.yml
conda activate analog

Environment Check

To ensure the current environment is functional, the following tests can be performed:

cd sample_design
python test_all_sample_design.py

When the program finishes running, if All tasks passed is displayed, it indicates that the environment is normal.

Otherwise, it will display Please check your environment and try again. It means you should check the configuration of the current Python environment, especially the settings related to PySpice.

Quick Start

You can directly run the following code for quick start.

python gpt_run.py --task_id=1 --api_key="[OPENAI_API]" --num_per_task=1

which will generate one circuit based on task 1.

Benchmark

• Task descriptions are in problem_set.tsv.
• Sample circuits are in directory sample_design.
• Test-benches are in directory problem_check.

Full Tutorial

1. Design basic circuits.

For circuits 1-15 are basic circuits which can be designed by AnalogCoder directly.

for task_id in {1..15}
do
   python run_gpt.py --task_id=$task_id --api_key="[OPENAI_API]" --num_per_task=15 --model=gpt-4o
done

All the generated circuits are saved in directory gpt4o/.

2. Build circuit tool library. (Optional)

We already provided a circuit tool library in directory subcircuit_lib. You can also build a new library based on the generated basic circuits.

python write_all_library.py

3. Design composite circuits. For circuits 16-24, they are composite circuits. They can be more easily designed with the circuit tool library.

for task_id in {16..24}
do
   python run_gpt.py --task_id=$task_id --api_key="[OPENAI_API]" --num_per_task=15 --model=gpt-4o --skill --retrieval
done

The argument skill means using the circuit tool library.

The argument retrieval means using LLM to retrieve tools from the library first.