Intel Inspector XE for Windows OS

Software Intel Inspector XE Is a simple tool for checking memory errors and multithreading for applications with serial and parallel code based on Windows and Linux. Code validation tool Intel Inspector XE Finds not only memory errors (access to uninitialized memory, leaks, etc.), but also problems caused by the interaction of threads (data races, deadlock, etc.), both existing and potential. Intel Inspector XE Performs a dynamic analysis of executable processes (not source code), checking what happens in the application, how it allocates and frees up memory, generates threads, uses synchronization objects, and so on. After that, the administrator gets a list of the problems found. Intel Inspector XE is available for Windows operating systems (freely integrated into Microsoft Visual Studio) and Linux. Intel Inspector XE A unique approach is used to analyze all instructions for reading / writing memory and their addresses at the binary code level. The analysis tool is based on the Pin Dynamic Binary Instrumentation Tool, which is embedded in the analyzed process just before the start and allows you to monitor the execution of almost any instructions, provides access to the contents of the registers, the context of program execution, symbolic and debugging information. Depending on the goals and objectives of the analysis, you can generate several types of Pintool-tools, which are configured to collect a certain type of data in the executable program. The main features of Intel Inspector XE:

• Search for memory errors and multithreading.
• No need for specialized structures, access to two interfaces: graphical user and command line.
• Detect hidden errors in complex parallel programs.
• Support for checking C code, C ++, C #, F # and Fortran.
• Selection of three levels of analysis, depending on what speed and what quality of verification are required.
• Integration into debuggers Visual Studio, GDB and IDB for instant correction of the found problems.
• Heap Growth analysis to identify the reasons for the continuous increase in the load on the memory while the application is running.
• Accelerate the analysis procedures by narrowing their scope, starting monitoring only during the execution of the alleged problem.
• Isolate false positives when searching for errors so that they are not displayed in the list.
• Filtering lists of errors by priority, type, state, module, etc.
• Collective work - the lists generated by the program can be accessed by all team members.
• Setting up what memory should be excluded from the analysis.
• The ability to use in the analysis of software for Intel Xeon Phi products.
• Analysis of hybrid MPI and OpenMP applications for memory errors.