Sometimes it becomes necessary to do a computer stability test. For example, stability is usually checked after building or purchasing a new computer. It is also advisable to carry out a stability test after replacing components or after servicing the computer. This allows us to identify possible problems at an early stage and take the necessary actions.
Computer stability testing is usually carried out in stages. First, they check the stability of the processor, then the stability of the video card, and so on. Creating an accentuated load on individual components allows you to quickly determine the source of problems if the computer does not work stably.
To create the maximum load on computer components, you need specialized programs designed specifically for testing. After all, even the most demanding professional programs or computer games are not created required loads. For example, you can use the LinX program to test a Furmark video card, and to test an S&M power supply. There is also universal programs, which include different tests. One of the most popular programs of this kind is the OCCT program. This program allows you to conduct a stability test for all major computer components (processor, video card, power supply). At the same time, OCCT has its own built-in system monitoring, which allows you to monitor load levels, temperatures, voltages and much more.
In this article we will use OCCT, since it is more convenient and saves time on installing programs. You can download OCCT on the official website. Also, when testing, you may need a program, more about this at the end of the article.
It should be noted that a computer stability test can lead to computer failure, for example due to overheating. Although this is very unlikely, it is a possibility. Therefore, everything you do is done at your own responsibility.
Checking (diagnostics) of a computer or laptop– a mandatory operation when overclocking their components. Residents who do not use the overclocking option may need to test their computer when purchasing new system. This is especially true when the hardware is taken “by hand”, i.e. without any special guarantee. In this case, it is necessary to carry out diagnostics processor, video card, memory etc. to make sure stability of a computer or laptop under load.
Computer stability test usually performed by long-term load on its components. For each of the nodes being checked, they usually use their own diagnostic program. It would seem that you can launch one resource-demanding game and the processor, video card and memory will be automatically checked, but this is not so. Games usually put a lot of strain on graphics system, A CPU in this case it may not be involved full power. As a result, the CPU was left without testing under maximum load. That is why there are independent diagnostic programs.
Please note that testing (of any component) should take as long as possible: a couple of minutes is usually not enough even to simply warm up to the maximum temperature of the diagnosed hardware. You can be more or less sure of stability only after 10-15 minutes of operation of the unit at the maximum temperature. Those. after launched diagnostic program, it is necessary to monitor the temperature of the component being tested, and after its temperature has stopped rising, at least 10-15 minutes of its operation without failures are necessary. It usually takes up to 5 minutes to warm up.
How to check the processor.
Previously, the site already posted information about, which creates the maximum load on the CPU and warms it up well. This is one of best utilities of a kind. In principle, there are many similar programs that, in addition to a stability test, are also utilities from the “benchmark” section for assessing CPU performance.
Quite suitable for checks and modern archiver WinRAR, which has a built-in function for assessing the performance of a computer’s computing capabilities and no errors at the same time (“Tools”-“Benchmark” or “Alt+B”). In addition, it also loads the memory well, so if its operation is unstable, the error counter will be different from zero. Now many archivers have a similar benchmark function - including a popular and quite good one archiver "7-Zip".
CPU temperature monitoring under maximum load can reveal a problem such as insufficient cooling. In rare cases, the maximum permissible temperature reaches 95 degrees. Usually somewhat lower. So if you see readings from the corresponding sensors close to these values, you should replace the cooling system CPU (in the best case, replacing the thermal paste under the heatsink is enough).
Checking the video card.
Checking the video card– the most spectacular stage computer diagnostics or laptop. Unlike boring tables in CPU test utilities, at this stage the monitor will please the eye of amateurs computer games all sorts of fantastic characters, landscapes or simply beautiful abstract figures that will require all the power of the video card.
One of the most popular programs for checking a video card - “FurMark”. In addition to the fact that the utility copes with its tasks perfectly, it takes up very little disk space. Those who are unable to remember the name of this program can always find it in search engine on request " hairy donut”, since this is exactly what the picture on the monitor looks like during the diagnostic process. By the way, to monitor the temperature of the graphics chip, memory frequencies and the GPU itself, there is no need for third-party utilities - “” displays this data on the screen on top of the “donut”. This is exactly what software product we used in the assessment temperature regime of the video card "" from Asus. An undoubted advantage is that FurMark is a free program, which can be downloaded from the official website.
« 3DMark" is another extremely popular program whose main function is. The history of this utility goes back to the distant past, what is happening on the monitor is breathtaking, and you can’t take your eyes off the magnificent graphics of the highest quality in latest versions this software for the latest and fastest graphics adapters. Some versions include not only GPU checking, but also CPU stress testing, so if there is a categorical lack of time for testing, you can limit yourself to a complete run of all the tests from the proposed "3DMark" utility set (bypassing separate CPU testing [ but at least 3-5 minutes in LinX is still desirable]). The disadvantages of 3DMark include the impressive amount of disk space it takes up and the need to pay some amount to forgive additional tests, and not just basic ones. Except free "3DMark Basic Edition" the developer offers a “3DMark Advanced Edition” for $25, and a “3DMark Professional Edition” for a whopping $1000.
How to check RAM.
If there are malfunctions in the operation of modules random access memory problems usually appear when testing other components, but there are also specialized verification programs. We have repeatedly mentioned in articles about “”, which includes suitable utilities. One of them is "". If you select this item in the BootMenu, then RAM check will start automatically in a few seconds. Immediately after downloading, the program will prompt you to select safe mode(F1) or multi-threaded (F2).
During the process, the upper right area of the screen will display progress, the number of complete passes and a count of errors found. If bad blocks are found, a list of them will be displayed at the bottom of the screen. After successful completion of a full pass through the entire available memory area, the test will automatically start over, and the message “ Pass complete, no errors found, press Esc to exit" Pressing "Esc" will restart the computer.
Can carry out diagnostics and operating system tools. Windows 7 has a built-in " Checker Windows memory " Find it through the built-in search menu " Start"or manually: "Start" - "Control Panel" - "System and Security" - "Administrative Tools" - "Windows Memory Checker". When you select this item, the system will prompt you to reboot. After a reboot, the built-in testing mechanism will automatically start. After the test is completed, the computer will automatically restart into Windows, and you will see a message in the notification area with diagnostic results.
How to check your hard drive.
Unlike other components of your computer Storage System may have a cost several times higher than the rest of the hardware. More precisely, it is not the drives themselves that are very expensive, but the data that is stored on them. In view of this, it is recommended to check not only the drive being installed in the system, but also regularly monitor the condition hard drive , at least looking through the values from S.M.A.R.T. The number that says the most about a disk's reliability is "reassigned" sectors (« Reallocation Sector Count") - This bad sectors, who were replaced by spares from the reserve area. View SMART You can use most disk checking programs.
Hello friends... Computer stability - how to check? Today we will talk about free program with the Soviet name OCCT Perestroika - which is designed to test a computer for stability. The question often arises of checking your computer for the cause of various errors.
Whether it's overheating of the computer, spontaneous reboots, freezes, or worse than that blue screens of death, where all these problems come from. This utility conducts testing in pairs, that is, memory, and also does a separate check and .
The program shows the load on the frequency, temperature and voltage of the CPU/GPU, displays information about some elements of the power system in real time, and can also check each component separately. After launch, the program records and reports the minimum and maximum values of the sensors.
The user can change the CPU clock speeds at any time and graphics adapter, and then immediately run the benchmark, which creates an artificial load on the components described above. When testing the system, the temperature and voltage program sensors should not exceed the values stated by the hardware manufacturer.
If, when testing the system, an increase in temperature is observed on the sensors, then the frequency should be lowered, as well as the cooling and power system should be upgraded. OCCT Perestroika displays visual graphs of load, voltage and temperature on the monitor screen. The program also has the ability to obtain data in the form text file or export as a table.
The program has a special 64-bit test mode for better work on the appropriate operating system. You can download the utility from developer site by pressing one of the buttons.
The installation is extremely simple, I will not dwell on it. The main program window looks something like this.
The window located on the right may be slightly different; it is configured with the corresponding button in the left window with the image of a gear. When you click on this button, a window with settings will open where you can tick the box that will be displayed in the Real-time Monitoring window.
CPU:OCCT
Testing the processor, memory and motherboard for load, temperature, power supply under stressful conditions. The testing type must be set to “Auto”; we do not touch the other points.
Test version- I installed 64 - Bit, according to my operating system. To find out the bitness of your system, go to “Start”, select “Computer” by clicking right click mouse and select “Properties”. In the “System Type” section you will see the bitness of your operating system.
Test mode— in the settings, select a small, medium or large data set from the drop-down menu. When selecting a small set, only the processor is tested. The middle set tests the processor and . By selecting a large set, the program will test the processor, memory and motherboard chipset. For the test, I chose a large data set.
Number of threads— the number of threads, here you need to check the box next to the “Auto” item in order to use everything possible. For testing example, let's take Intel processor core i3 2125 it is dual core. But thanks to Hyper-threading technology, each physical core can pull two threads at once. That is, there are 4 logical cores.
Before starting testing, close all running programs and programs hanging in the tray. Then press the “ON” button. Testing will take approximately an hour. Upon completion of the test, go to the folder along the path C:\Users\Anton\Documents\OCCT\, there you will see parameter graphs where everything is clearly shown.
CPU:LINPACK
Testing exclusively the processor and only the processor. Not recommended test for laptops, netbooks, since the cooling system in them is very weak. This test is very warms up and loads the processor. Recommended only in necessary cases, and also with powerful cooling system. If you are not sure, use the first test.
Test type — as always, I tore out “Auto”. The items “Duration” and “Period of inactivity” are left unchanged.
Memory— we leave it as is, but at the same time it is necessary to close everything running programs, and programs in the tray, otherwise reduce the value to 70-80 percent. When choosing a 64-bit system, again according to your bitness.
AVX - Linkpack compatible
If your computer has a processor compatible with this extension, check the box. If you don’t know what we’re talking about, it’s better to leave this item blank.
We tap the “ON” button, after which we do not touch the computer for an hour.
Testing video card stability - GPU:3D
Testing takes place exclusively for the video card, or rather GPU and memory.
Type testing
DirectX version- if version 11 is installed on the computer, leave it as is, if version 9 is installed, then mark nine accordingly.
Permission
Full screen mode, enable error checking- check the boxes.
Shader complexity- in essence, this operation is performed by the video card in one pass; in most cases, the maximum value is selected.
Limiter— leave as default.
We press the “ON” button, wait for the testing to start, since it will not start immediately, we leave the computer for an hour and nervously smoke on the balcony. The test looks something like this.
Testing the POWER SUPPLY power supply
This test is designed to maximum load all batteries, thereby diagnosing the power supply.
Type testing- select “Auto”. “Duration” and “Periods of inactivity” are left as they are.
DirectX version- if version 11 is installed on the computer, leave it, if version 9 is installed, then we mark nine accordingly.
Permission— choose according to the resolution of your monitor.
64 - Bit- if supported by the system.
Use all logical cores
- Be sure to check the box. On some systems it may not be available if there are no kernels, or there is simply no access to them.
We press the “ON” button and go drink coffee, about an hour.
What to do if the stability test ends with an error
The simplest thing you can do is clean your computer from dust, read. Replacing thermal paste on the processor or graphics chip. To eliminate suspicion, connect a more powerful power supply during the test.
If there is an error in the test or it simply does not pass (does not start), it is necessary to remove overclocking, if any. If the video card or processor has not been overclocked, and testing has not started, then cover it under warranty (if there is one left).
Try lowering the clock speeds, in some cases it helps to run testing (special utilities for the motherboard will help you with this). Otherwise, think about replacing your computer or upgrading. This is where I will end my article, bye everyone...
Hello, friends! In this article, we will diagnose components using the most powerful computer stability test - OCCT. The OCCT test stands for OverClock Checking Tool. This special utility capable of maximally loading the components of your computer by subjecting them to all possible tests to identify errors. In other words, using OCCT you can perform computer stress test for stability.
OCCT notifies the user of any errors found. If any are found, then something is really wrong. In everyday work, errors may not be noticeable, since you are not subjecting your computer to such loads. But it is highly likely to appear in the future sooner or later. It is possible that this will be in the form of a blue screen of death. To avoid such surprises, you can and should test your new or updated computer.
Article structure
- Downloading and setting up the OCCT computer test
- CPU, memory and motherboard test - CPU:OCCT
- CPU test - CPU:LINPACK
- Graphics adapter test - GPU:3D
- Power supply test - POWER SUPPLY
According to the developer of OCCT, a 30-minute test will be enough for most users. But for greater reliability, it is advisable to run tests lasting 1 hour.
1. Download and set up the OCCT computer test
Go to the Download tab and at the very bottom there will be download links
I like Zip Version since it does not require installation.
Launch OCCT.exe
You can see the appearance of the program in the figure below.
The monitoring window on the right may vary slightly. This window is configurable. To do this, click on the orange button in the left window
In the options that open, in the last column you can configure what will be displayed in the Monitoring window
You can see my settings in the picture above
After these settings, the Monitoring window takes on the following form:
2. CPU, memory and motherboard test - CPU:OCCT
Test type install Auto.
We do not touch durations and periods
Test version I installed it automatically correctly - 64 bit. Corresponds to the bitness of your operating system. (In order to go to Start, click Computer, right-click and select Properties. In the window that opens, in the System Type section, you will see the bitness of your Windows)
Test mode. You can select Small, Medium, or Large dataset from the drop-down list. According to unverified but reliable sources, when selecting Small amount of data, only the processor is tested for errors. When selecting Average data volume, the processor and RAM are tested. When selecting Large data volume, the processor, memory and motherboard chipset are tested.
Selecting Large Data Set.
Number of threads— number of threads. Check the box - Auto in order to use everything possible. Processor under test Intel core The i3 2125 is dual-core, but thanks to Hyper-threading technology, each physical core can pull two threads at once. That is, there are 4 logical cores.
Before starting the test, it is advisable to close all running programs and exit programs that hang in the notification area.
At the end of the test, Explorer will open at C:\Users\Anton\Documents\OCCT\
In the folder with the current date there will be graphs of various parameters depending on the processor load. Everything is clearly shown there.
If errors are found during the test, you will see a warning. Read what to do in this case in the Conclusion.
3. CPU test - CPU:LINPACK
This test only heavily loads the processor. It warms it up better than CPU:OCCT
Test type choose Auto. Leave the duration and periods of inactivity as is.
Memory we leave it the same way as it is.
If you have a 64-bit system, check the appropriate box.
If your processor supports the AVX command system extension, check the appropriate box.
Here is an excerpt from Wikipedia
I have a Sandy Bridge processor that supports AVX, so I check the box.
We also install if the Use all logical cores is not checked.
We run the test and do not touch the computer for one hour.
At the end of the test, we look at the temperature graphs. If no errors were found during the test and the temperatures are normal, then everything is in order. Otherwise, look at the Conclusion.
4. Graphics adapter test - GPU:3D
Test Type: Auto. We do not touch the durations and periods of inactivity.
Version DirectX - 11. For older video cards it is better to install DirectX9.
Then check the boxes to enable full screen mode and enable error checking.
Shader complexity. When you hover over this field with the mouse, a hint is shown below in the Help section
That is, for AMD video cards we select 7, for NVIDIA - 3. Since I have integrated graphics from Intel, I leave it at default.
Memory usage. Memory limit for testing. It seems to me that it is advisable not to check the box. Let him use as much as he needs.
Frame limiter I leave it the same way default.
Good day, dear friends, acquaintances, readers, admirers and other individuals. Today we will test the computer via OCCT.
It is often necessary to find out the cause of the problem, or simply any problem, from reboots/freezes to computer shutdowns.
In “field” (i.e., in normal working) conditions, this is not always possible, because some problems have a rather peculiar, so to speak, floating character and are not so easy to diagnose. And usually it’s not enough to just find out what is to blame, and not the software part, but you also need to understand what exactly is plotting, or rather what specific piece of hardware is faulty. In such situations, it comes to our aid specialized software to check stability.
After launching, we will see in front of us a red USSR-shaped program window (see screenshot above) in which, in theory, the Russian language should be immediately set. Before this, a window with a donation button may appear, for now you can close it (well, or immediately support the developer, it’s up to you).
If this is not the case, then click on the gear icon on the right, and then set the desired parameter. Or use it as is.
How to test a computer in general
The program has a set of tabs:
- CPU:OCCT and CPU:LINKPACK, - testing processor stability under stressful conditions (load, power supply, temperature, etc.);
- GPU:3D, - testing the stability of the video card;
- POWER SUPPLY, - testing the stability of batteries (motherboard, power supply, circuits, etc., in general, load tests).
Let's try each of them, because each has its own parameters.
ATTENTION ! Use OCCT with caution on laptops due to the high load and heat generated. On laptops with a weak/damaged cooling system (and other elements), this can lead to unpredictable consequences. It's probably wise to use AIDA64 on them.
Before the test, go to the above-mentioned settings (where you set the language) and set the processor temperature limiter (usually 85 is too high a value) and other (if necessary) components.
This is done as follows. We exhibit:
- Test type: auto;
- Test duration: 1 hour 0 minutes;
- Test mode: large data set.
Comments on the items raised:
- Works specified time , i.e. an hour or more (or until an error is detected), allows you not to waste extra time on diagnostics;
- Test time is the test time;
- Data set - determines the load level and the heat generated, as well as the number of elements being tested. If the data set is small, then only the processor is tested; if it is medium, then the processor + memory; if it is large, then the processor + memory + chipset. In a large set, the heating is stronger, but you can find more errors; in a small set, there is less heating, but you can miss something important;
Other parameters:
- Inaction at the beginning and at the end - leave it as is, allows you to reduce the load before/after launch and read the necessary data;
- Test version - select the one that matches installed version operating system;
- The number of threads - as a rule, the "Auto" checkbox is enough, but if it is determined incorrectly (less than the number of physical and logical processor cores), then you can set it manually by unchecking it.
The second tab, namely CPU: LINPACK, is another test, but only the processor, and not many elements at once (see the description of the first tab above).
Testing Warning
You should be careful with this test, because it loads and heats up the processor extremely (including the core power supply, if supported by the motherboard) and is an extremely extreme test. Recommended for use only if available most powerful system cooling and the urgent need to diagnose the processor. In other cases, it is better to use the first test.
For those who have decided (usually required if the first test did not reveal problems, but “visually” they are preserved):
- Test type: auto;
- Test duration: 1 hour 0 minutes;
- Test mode: 90% memory (I recommend closing all possible programs, etc., or reduce this value to 70-80%);
Then all that remains is to press the ON button and wait an hour (or less, if an error is found, the computer will freeze, turn off, or show some other signs of overheating and failure) while the system is scanned. The analysis of the results is described at the end of the article.
Crossfire and SLI are supported, checking and identifying many errors during strong heating during loads, and also, using a special system, artifacts (image distortion) are determined. You can do testing with different numbers of shaders, FPS and everything else.
Here, in fact, we present the following:
- Test type: auto
- Test duration: 1 hour 0 minutes
- Type: full screen (checkmark);
- Checking for errors: for the first test there is usually no need to check it, for the second (if the problem visually persists, but no errors are found) it makes sense to check the box;
- Shader complexity: in general, this parameter is responsible for the number of operations performed by the video card in one pass (most often, the maximum available value is selected, or, if you need to specifically test for an application, then select the value that the application uses);
- Limiter: 0 (zero), or 60 (if you use vertical synchronization and need to test the work for it).
Then all that remains is to press the ON button and wait an hour (or less, if an error is found, the computer will freeze, turn off, or show some other signs of overheating and failure) while the system is scanned. The analysis of the results is described at the end of the article.
The test itself looks usually like in the screenshot above. It does not start immediately (see periods of inactivity), it can change the type of picture (image). Significant visual distortions (it is difficult to confuse them with something else) are artifacts and indicate problems with the video card, its memory and something else.
ATTENTION ! Quite complex to analyze, it is recommended to use it only if the first tests did not reveal anything in any way, but the problems persist. Dangerous and absolutely not suitable for cheap (noname) and low-quality power supplies. Use at your own risk.
Similar to the previous test, the following is set here:
- Test type: auto
- Test duration: 1 hour 0 minutes
- DirectX version: if available, then 11, if not, then 9, if you need to specifically test for some application, then select the value that the application uses;
- Resolution: current, or, if you need to specifically test for some application, then select the value that the application uses;
- Type: full screen mode (checkmark);
- 64 bit: if the system and processor support;
- AVX-compatible Linkpack - it makes sense to avoid if you don’t know what we’re talking about;
- Use all logical cores, - must be checked if the checkbox is available (may not be available if they are not present or there is no access to them).
Then all that remains is to press the ON button and wait an hour (or less, if an error is found, the computer will freeze, turn off, or show some other signs of overheating and failure) while the system is scanned. The analysis of the results is described below.
Analysis of OCCT test results
As a result of the tests, you can get the following result:
- Graphs - most often, in the absence of a severe physical failure (shutdown, reboot, freeze, etc.), are the result of any test, contain temperatures, voltages and other data for analysis;
- Error (in the program), - usually this is a kernel error or something else that stops the test (but the computer is working), most often its number is indicated or at least short description(kernel failure of such and such);
- Blue screen of death, - that it makes sense to read;
- Physical failure (or protection triggered) - shutdown, reboot, freeze and similar horrors of life.
How to take off with this;
- For analysis temperature charts read (pay special attention to the maximum permissible values); if in doubt, see the documentation for the overheating component (paper, or on the manufacturer’s website) to analyze the maximum permissible temperatures;
- For analysis nutrition-related charts, it is worth understanding that minor discrepancies are acceptable (by tenths, hundredths, or less, orders of magnitude), excluding certain values(for example, the processor power supply can change quite a lot due to energy saving technologies, frequency regulation, overclocking, etc.). If it’s difficult to understand, then refer to the documentation;
- For analysis blue screens read;
- Physical failures are often associated with overheating and activation of protection systems. In its basic form, it is usually solved, not complicatedly, by adding (or changing the position for blowing/exhausting) fans in, small ones. Less often, the entire cooling system (how to choose) is changed to a more powerful one, or they (failures) are the result of a complete failure of the component. The latter are the most difficult to diagnose; most often, power supply failures are immediately clear (not complete shutdown computer or does not turn on immediately) and/or video card (image artifacts).
If there are complex problems which you need to figure out by looking at graphs and so on, then contact, for example,.
Afterword
I repeat that this is one of the most powerful stability tests that, in principle, can be found. It is quite often used by overclockers (those who overclock computer components) to check stability, which says a lot.
As always, if you have any (reasonable) thoughts, questions, thanks or additions, then, as usual, write them in the comments to this article (or on the forum mentioned above).
Loading...
