In Windows operating systems, the error “Memory cannot be read” is perhaps one of the most common. Various gurus on the Internet give recommendations on how to eliminate it and restore the normal functioning of the system. It is not surprising that everyone is of the opinion "I am right, and the rest are not." Here are just some tips that are not always effective. However, there is no deception: the message “Memory cannot be read” may appear due to several reasons, therefore, there are also several ways to fix the problem.
First, a little theory. One of the benefits of the latest Microsoft operating systems is its redesigned and improved memory manager. If in Windows 9x various failures and reboots were commonplace, and even prolonged operation in normal mode could lead to problems, then in the line based on the Win NT kernel, this moment was paid the closest attention. As a result, the memory error message almost stopped appearing, because the mechanism of work was radically redone. By the way, this is why novice users who started exploring the world of digital technology through Win 7 may not even know what the blue screen of death (BSOD) is. It is worth paying tribute to Microsoft - the reliability of the system has increased significantly. Where, in this case, from the forums do users get questions about the error "Memory cannot be read"?
The mechanism of work is as follows: when starting any program, the system manager allocates a certain isolated piece of free memory to it. This solution allows you to optimally implement the work of applications in a multitasking system. If the running program tries to access the memory outside the allocated space, the message “Memory cannot be read” is displayed and the application closes. Thanks to this, it is possible to ensure the stability of the operating system. It is understandable that the problem most often lies in third-party applications that are not working correctly, and you should not expect help from Microsoft. Well, the theory is over, and we can move on to practical recommendations.
One of the most common causes is a hardware memory error that occurs in one of the installed modules. Immediately, we note that I happened to encounter this several times, so the recommendations are not unfounded. Although modern memory microcircuits are quite reliable, and if problems arise, then infrequently, when reporting error messages, you need to check the modules. Before conducting a software test, you need to open the case and remove the memory modules. Of course, provided that the case is not sealed with a warranty sticker. The comb of the copper contacts of the modules must be thoroughly cleaned with a clerical eraser: sometimes the oxides that appear worsen the electrical conductivity, and the eraser eliminates them by polishing the copper to a shine. After that, the memory is carefully inserted back into the connectors, not forgetting about the side latches. From this moment it should be observed whether the error has disappeared.
If everything repeats, then you have to resort to specialized programs. One of the most effective, giving almost 100% detection result, is Memtest86 + (not to be confused with TestMem). The peculiarity is that the test is intended to be launched in DOS, therefore it is necessary to download an image from the developer's website for recording on a disk (flash drive) and prepare the media in accordance with the instructions. It will remain to boot from this media and test all the memory. If errors are found, each module should be checked separately.
By the way, sometimes the cause of errors lies in inflated timings or overclocking of components. Therefore, it is imperative to remove overclocking (if any) and check the real and actual timings (you can find out in the CPU-Z program).
A little tip: if the problem is observed only with a specific application, then you need to update it or not use it at all, trying to replace it with an analog. And, of course, you should use the original operating systems, not their assembly.