Identifying Compiler and Optimization Level in Binary Code From Multiple Architectures

Abstract

While compiling a native application, different compiler flags or optimization levels can be configured. This choice depends on the different requirements. For example, if the application binary is intended for final release, the flags and optimization settings should be set for execution speed and efficiency. Alternatively, if the application is to be used for debugging purposes, debug flags should be configured accordingly, usually involving minor or no code optimization. However, this information cannot be easily extracted from a compiled binary. Nonetheless, ensuring the same compiler and compilation flags is particularly important when comparing different binary files, to avoid inaccurate or unreliable analyses. Unfortunately, to understand which flags and optimizations have been used, a deep knowledge of the target architecture and the compiler used is required. In this study, we present two deep learning models used to detect both compiler and optimization level in a compiled binary. The optimization levels we study are O0, O1, O2, O3, and Os in the x86_64, AArch64, RISC-V, SPARC, PowerPC, MIPS, and ARM architectures. In addition, for the x86_64 and AArch64 architectures, we also determine whether the compiler is GCC or Clang. We created a dataset of more than 76000 binaries and used it for training. Our experiments showed over 99.99% accuracy in detecting the compiler flags and between 92% to 98%, depending on the architecture, in detecting the optimization level. Furthermore, we analyzed the change in accuracy when the amount of data was extremely limited. Our study shows that it is possible to accurately detect both compiler flag settings and optimization levels with function-level granularity.

Compiler Provenance Recovery for Multi-CPU Architectures Using a Centrifuge Mechanism

“…Pizzolotto and Inoue proposed the Optimization Detector [30], which uses CNN or LSTM [8], [9] to identify the compiler's optimization level. They checked the classification performance of binaries with two different compilers and five different optimization levels for each of seven different CPU architectures.…”

C2TACO: Lifting Tensor Code to TACO

“…Other tasks range from code style detection [50], generating accurate variable names [41], correcting syntax errors and bugs [32,57] code completion [36] and program synthesis [14] to API recommendation [35], and specification synthesis [43]. While powerful, such approaches are inaccurate and are not mature enough for precise lifting.…”

BinCC: Scalable Function Similarity Detection in Multiple Cross-Architectural Binaries

“…This was originally highlighted by Saebjørnsen et al [22] and we can confirm that not only the compiled binary is semantically different at various optimization levels but also structurally different, and we expect every CFG-based approach to suffer from this limitation. In order to overcome this problem, however, a previous work of us was focused on determining automatically the compiler and optimization level used in a binary, in order to mitigate failed analyses due to mismatched compiler configurations [50].…”