Approximate Oracles and Synergy in Software Energy Search Spaces

Abstract

Abstract-Reducing the energy consumption of software systems though optimisations techniques such as genetic improvement is gaining interest. However, efficient and effective improvement of software systems requires a better understanding of the code-change search space. One important choice practitioners have is whether to preserve the system's original output or permit approximation with each scenario having its own search space characteristics. When output preservation is a hard constraint, we report that the maximum energy reduction achievable by the modification operators is 2.69% (0.76% on average). By contrast, this figure increases dramatically to 95.60% (33.90% on average) when approximation is permitted, indicating the critical importance of approximate output quality assessment for code optimisation. We investigate synergy, a phenomenon that occurs when simultaneously applied source code modifications produce an effect greater than their individual sum. Our results reveal that 12.0% of all joint code modifications produced such a synergistic effect though 38.5% produce an antagonistic interaction in which simultaneously applied modifications are less effective than when applied individually. This highlights the need for more advanced search-based approaches.

A survey of genetic improvement search spaces

“…However, an effective mutation (i.e., an individual of better fitness than the original code) in automated program repair, for instance, only needs to pass all tests, while for improvement of runtime such a mutation would not be considered effective if it led to a slowdown with respect to the original code. Moreover, certain studies in non-functional property improvement consider approximate results [5], thus allowing for test failures.…”

“…Furthermore, restrictions were put on the number of lines considered for mutation imposed by the specialised BNF grammar and initial profiling to determine the most time-consuming parts of code. Bruce et al [5] investigated the GI search space for energy consumption optimisation. They were the first to consider synergistic effects between mutations and the first to consider approximate solutions in this context.…”

“…Authors also provided two extensions to SPR and Prophet, that is, condition synthesis extension (which considers certain binary comparison operators) and value replacement extension (which introduces certain variable and constant replacements). 5 These additional transformations further increased the search spaces considered. It is, however, unclear whether all transformations allowed for both systems were considered as the search space.…”

Impact of Test Suite Coverage on Overfitting in Genetic Improvement of Software

“…GI navigates the search space of mutated program variants in order to find one that improves the desired property. This technique has been successfully used to fix bugs [1,14], add an additional feature [3,20], improve runtime [12], energy [7], and reduce memory consumption [5,24].…”

2019 Index IEEE Transactions on Software Engineering Vol. 45

“…Nuseibeh, B., see Peters, F., TSE June 2019 615-631 Nygard, J.F., see Lu, H., TSE Feb. 2019 148-170 O Oentaryo, R.J., see Hoang, T., TSE Oct. 20191002-1023Oertel, N., see Mottola, L., TSE June 2019576-596 Okoli, C., see Mesgari, M., TSE Oct. 2019 Oliveto, R., see Palomba, F., TSE Feb. 2019 194-218 Oliveto, R., see Scalabrino, S., TSE Jan. 2019 68-86 Oliveto, R., see Ponzanelli, L., TSE May 2019 464-488 Oppermann, F.J., see Mottola, L., TSE June 2019 576-596 P Palma, F., Moha, N., and Gueheneuc, Y., UniDoSA: The Unified Specification and Detection of Service Antipatterns;TSE Oct. 20191024-1053TSE Feb. 2019 194-218 Pan, Y., White, J., Sun, Y., and Gray, J., Gray Computing: A Framework for Computing with Background JavaScript Tasks; TSE Feb. 2019 171-193 Panichella, A., see Beller, M., TSE March 2019261-284 Panichella, A., see Jan, S., TSE April 2019335-362 Penta, M.D., see Scalabrino, S., TSE Jan. 2019 Penta, M.D., see Ponzanelli, L., TSE May 2019 464-488 Peters, F., Tun, T.T., Yu, Y., and Nuseibeh, B., Text Filtering and Ranking for Security Bug Report Prediction; TSE June 2019 615-631 Petke, J., see Bruce, B.R., TSE Nov. 20191150-1169 1106-1124 Riccobene, E., see Arcaini, P., TSE May 2019 507-520 Romer, K., see Mottola, L., TSE June 2019 576-596 Romero, J.R., see Ramirez, A., TSE Aug. 2019760-781 Roychoudhury, A., see Bohme, M., TSE May 2019489-506 Ruhe, G., see Nayebi, M., TSE Sept. 2019839-857 Russo, B., see Scalabrino, S., TSE Jan. 2019 Russo, B., see Ponzanelli, L., TSE May 2019464-488 S Santone, A., see Canfora, G., TSE Dec. 20191230-1252 to the Crowd for the Release Planning of Mobile Apps; TSE Jan. 2019 68-86 Scanniello, G., see 363-390 Shepperd, M., see Song, Q., TSE Dec. 2019…”