Programmers often use a core dump to examine the problem with the use of a debugger. A core dump can include all the system memory or a part of the program that failed. There are several reasons why a computer or program can crash:

• Corrupted data
• A severe user error
• Virus-infected files
• Problems accessing data files
• An outdated operating system
• A segmentation fault or bus error
• A poorly ventilated or dusty computer tower
• A system-detected fault in the software or hardware
• Computer overheating caused by a faulty heat sink or fan

Generally, a core dump file includes the random access memory (RAM) contents of a certain process or part of an address space of the process and values of processor registers. The core dump files can be used to analyze the cause of the dump, viewed as text or printed.

Because a contemporary OS process address space may share breaks and pages with other files and processes, a more intricate image is used. In Unix-like systems, core dumps typically use the standard executable image format:

• Mach-O in Mac OS X
• a.out in older versions of Unix
• Executable and linkable format (ELF) in modern Linux, Solaris, Unix System V and Berkeley software distribution (BSD) schemes

Originally, a core dump transferred the contents of memory precisely in order to record the state of the computer. The core dumps were actual printouts of around a hundred pages or more that consisted of octal or hexadecimal numbers. The pages were studied by programmers to research the cause of the crash or abnormally terminated program. Eventually, the introduction of debuggers eliminated the need for massive stacks of printouts.