1. What are the contributions mentioned in the paper "Statistical debugging for real-world performance problems" ?
In this paper, the authors first conduct an empirical study to understand how performance problems are observed and reported by real-world users.. Their study shows that statistical debugging is a natural fit for diagnosing performance problems, which are often observed through comparison-based approaches and reported together with both good and bad inputs.. The authors then thoroughly investigate different design points in statistical debugging, including three different predicates and two different types of statistical models, to understand which design point works the best for performance diagnosis.. Finally, the authors study how some unique nature of performance bugs allows sampling techniques to lower the over∗. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page.. Copyrights for components of this work owned by others than ACM must be honored.
read more
2. What future works have the authors mentioned in the paper "Statistical debugging for real-world performance problems" ?
Their study also points out directions for future work on fine-granularity performance diagnosis.
read more
3. What is the main purpose of statistical debugging?
statistical debugging collects program predicates, such as whether a branch is taken, during both success runs and failure runs, and then uses statistical models to automatically identify predicates that are most correlated with a failure, referred to as failure predictors.
read more
4. What is the common reason for users to file bug reports?
large performance difference under the same input and different configurations is also a common reason for users to file bug reports.
read more
![Table 5: Experimental results for in-house diagnosis (Xx(y): the x-th ranked failure predictor is highly related to the root cause, and is y lines of code away from the patch. (.): the failure predictor and the patch are more than 50 lines of code away from each other or are from different files. Xx[y]: a y-th level caller of the x-th ranked function in a profiler result is related to the root cause; x[0] means it is the function itself that is related to the root cause. -: none of the top five predictors are related to the root cause or no predicates reach the threshold of the statistical model.).](/figures/table-5-experimental-results-for-in-house-diagnosis-xx-y-the-3fikiq8i.webp)
