LLMPerf: GPU Performance Modeling meets Large Language Models

📅 2024-10-21
🏛️ IEEE/ACM International Symposium on Modeling, Analysis, and Simulation On Computer and Telecommunication Systems
📈 Citations: 0
Influential: 0
📄 PDF

career value

245K/year
🤖 AI Summary
To address the bottleneck in GPGPU performance modeling—its heavy reliance on manual feature engineering and hardware simulation—this paper proposes, for the first time, an end-to-end LLM-based approach for predicting OpenCL program execution time. The method takes raw OpenCL source code as input and integrates program-structure-aware representation with a regression-based output design, eliminating the need for handcrafted features or low-level hardware simulation. The model is fine-tuned on a large-scale OpenCL source code dataset, achieving a mean absolute percentage error (MAPE) of 24.25% on a self-constructed validation set and 46.1% on a public OpenCL benchmark suite. Its core contribution lies in pioneering the application of large language models to GPGPU performance modeling, enabling fully automated, feature-free prediction from source code to runtime. This establishes a novel paradigm for compiler optimization and heterogeneous programming.

Technology Category

Application Category

📝 Abstract
Performance modeling, a pivotal domain in program cost analysis, currently relies on manually crafted models constrained by various program and hardware limitations, especially in the intricate landscape of GPGPU. Meanwhile, Large Language Models (LLMs) have demonstrated their effectiveness in addressing diverse programming challenges. Our work establishes a connection between LLMs and performance modeling, employing the LLM as a performance estimator. Through experimental exploration with carefully designed large-scale OpenCL datasets, we highlight the potential capability as well as the main difficulties of using LLMs in handling performance modeling tasks for OpenCL device source programs. As the first study for this line of work, our LLM-based performance model achieves a mean absolute percentage error of 24.25% for a large-scale generated validation set. On a set of publicly available OpenCL programs, our model achieves a mean absolute percentage error of 46.1%.
Problem

Research questions and friction points this paper is trying to address.

LLMs used for GPU performance modeling
Challenges in OpenCL program performance estimation
Achieving low error rates with LLM-based models
Innovation

Methods, ideas, or system contributions that make the work stand out.

LLMs used for GPU performance estimation
Large-scale OpenCL datasets for experimental validation
Achieves 24.25% error in performance modeling
🔎 Similar Papers
No similar papers found.
M
Minh-Khoi Nguyen-Nhat
FPT Software AI Center, Hanoi, Vietnam
H
Hoang Duy Nguyen Do
FPT Software AI Center, Hanoi, Vietnam
H
Huyen Thao Le
FPT Software AI Center, Hanoi, Vietnam
T
Thanh Tuan Dao
FPT University, Hanoi, Vietnam