English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Compiler bugs pose a significant threat to safety-critical applications, andpromptly and effectively isolating these bugs is crucial for assuring thequality of compilers. However, the limited availability of debugginginformation on reported bugs complicates the compiler bug isolation task.Existing compiler bug isolation approaches typically convert the problem into atest program mutation problem, but they are still limited by ineffectivemutation strategies or high human effort requirements. Drawing inspiration fromthe recent progress of pre-trained Large Language Models (LLMs), such asChatGPT, in code generation, we propose a new approach named LLM4CBI to tameLLMs to generate effective test programs for compiler bug isolation. However,using LLMs directly for test program mutation may not yield the desired resultsdue to the challenges associated with formulating precise prompts and selectingspecialized prompts. To overcome the challenges, three new components aredesigned in LLM4CBI. (1) LLM4CBI utilizes a program complexity-guided promptproduction component, which leverages data and control flow analysis toidentify the most valuable variables and locations in programs for mutation.(2) LLM4CBI employs a memorized prompt selection component, which adoptsreinforcement learning to select specialized prompts for mutating test programscontinuously. (3) A test program validation component is proposed to selectspecialized feedback prompts to avoid repeating the same mistakes during themutation process. Compared with the state-of-the-art approaches (DiWi andRecBi), our evaluation demonstrates the advantages of LLM4CBI: It isolates morebugs, ranging from 13.6% to 90.9% in various settings, than the otherapproaches. Additionally, we demonstrate that LLM4CBI is extensible, allowingfor easy integration with other LLMs.

LLM4CBI: Taming LLMs to Generate Effective Test Programs for Compiler Bug Isolation