Sun 27 Apr 2025 16:00 - 16:30 at 202 - Afternoon Session 2 Chair(s): Thanatad Songpetchmongkol, Oliver Krauss
Using the language independent genetic improvement tool MAGPIE (Machine Automated General Performance Improvement via Evolution of software) and logarithmic sampling, we measure the parameter fitness landscape when optimising the GNU glibc heap management of a million line C++ application, gem5. The malloc_info landscape is far smoother than is commonly assumed and savings of 11 percent with no loss of runtime speed are readily obtained by both Magpie and CMA-ES.
Sun 27 AprDisplayed time zone: Eastern Time (US & Canada) change
Sun 27 Apr
Displayed time zone: Eastern Time (US & Canada) change
16:00 - 17:30 | Afternoon Session 2GI at 202 Chair(s): Thanatad Songpetchmongkol University College London, Oliver Krauss University of Applied Sciences Upper Austria | ||
16:00 30mTalk | The gem5 C++ glibc Heap Fitness Landscape GI | ||
16:30 30mTalk | LLM-Assisted Crossover in Genetic Improvement of Software GI Dimitrios Stamatios Bouras Peking University, Justyna Petke University College London, Sergey Mechtaev Peking University | ||
17:00 25mMeeting | Discussion GI Aymeric Blot University of Rennes, IRISA / INRIA, Oliver Krauss University of Applied Sciences Upper Austria | ||
17:25 5mAwards | Awards and Closing GI |