One. OCCT's licensing mechanism is pretty simple. As long as your license is valid, you're free to do as you please.

There's no limit on the number of concurrent installation, running instances... the only limit is the license duration. Using your license on a friend or customer's computer (for instance, in an after-sale scenario) is also fine as long as you're the one operating OCCT, or someone from your household or working for your company.

Make sure you remove it after use, as distributing and copying your license is strictly forbidden !

Power supply
CPUSmall data setHeavy---MediumStays inside the CPU Cache
CPUMedium data setHeavyMedium--MediumMiddle-ground test
CPULarge data setHeavyHeavy--MediumHeavily uses memory
Power SupplyHeavyHeavyHeavy-Heavy

OCCT can't tell you where they come from. On the software point of view, it is as if it asked the for 2 + 2 and got 5 as an answer.

  • Is it because the CPU miscalculated and answered that 2 + 2 = 5 ?
  • Is it because the memory stored a 3 instead of a 2 ?
  • Is it because when transferring the 2 from the memory to the CPU, the 2 got transformed to a 3 ?
  • As there's no sure way of reproducing the error that occurred, there's no way of knowing what is going on. To further pinpoint errors, you have to dig yourself : lower the CPU frequency a notch and see if things are going better, up a voltage very carefully and see if it holds better... and pinpoint where the errors comes from. If you have access to spare parts, it is even easier to diagnose : replace and see what's going on.

To summarize, I am investing more and more money into OCCT, into licensing costs mostly ( HwInfo is the prime example ), and while i'm only starting now (after 18 years !) to make a small amount of money of OCCT, it isn't enough to make a living (far from it). Also, on a more personal side, I'm losing my main job due to COVID-related circumstances, so i have to admit monetizing OCCT has become more important, especially since i'm the sole income for my family. To give you an idea, i'm spending at least 30 hours per week on the program. Please support me in developing OCCT :)

Only the PRO version of OCCT will accept to launch on a computer joined to a windows domain. This is meant at detecting uses in professional and commercial environment, as they heavily rely on Windows domain. In the past, as there are enthousiasts that have a Windows domain at home, we offered a way of disabling this check as they were indeed not in a commercial environment, but this saw too much abuse. About 90% of the requests were from people trying to avoid the PRO license entirely. We thus removed that possibility, and you indeed need a PRO license now to run OCCT on a computer joined to a Windows domain.

There can be two causes to that : The first cause is usually an Asus (AI Suite...) or Corsair (iCue,...) program running. They aren't using industry-wide safety measures over monitoring, and their engines can collide with HwInfo, causing a lockup. A fix is to stop them ( beware, they are also running services in the background - go to services.msc and stop those). There's nothing me or HwInfo can do about it - report that to them, so that they start following the industry standard for monitoring. Don't be afraid to send my email to them so i can help !

The second cause may be a buggy driver that's hanging when asking for monitoring information. Go into OCCT's setting (the wrench icon on top), and disable flags there - you'll get less monitoring values, but finding the flag where the buggy driver resides might help. As a rule of thumb, disable "Drive scan" first.

If you're still stuck, try downloading this fail-safe config file, unzipping it in OCCT's folder and overwriting the previous one if it exists. This is a "fail safe" configuration, that disables all flags. This can be useful in cases where OCCT hangs right after starting. If all of the above doesn't work, contact us through reddit or email !

No, the Free version has always been designed to be usable "as is". OCCT isn't withholding information from you (I would call that stealing).


Send an email to and please give us at least a day for sending you your license back.


CPU:OCCT, Large data set, Extreme Mode, with Auto number of threads is, hands-down, what will detect CPU/Memory/Motherboard stability issues faster. Memory has been reported to be very good at finding Memory errors, despite being quite young. Use GPU:3D in fullscreen mode for GPU testing, VRAM for testing your Graphic card's Memory, and Power supply for testing your PSU. Most errors are detected within the first 5 minutes. However, i recommend doing at least a 1 hour long test just to be sure.

CPU:OCCT, Small data set, Extreme Mode, with Auto number of threads. Highest AVX-level available.


The memory test, using default settings, is the best option here. It doesn't test every cell in your memory, but requests so many data transfers your chips will get very hot !


3D with error checking. 3D cards are not designed to run at 100% power, and if you want to simulate a game load, make use of the limiter !

Use 3D without error checking, and make sure to disable the limiter.

NO ! A Power virus is a stress test that you do not have control upon. This term was applied to OCCT's GPU test when we launched its first version, 10 years ago, by the marketing team of a certain brand of graphic card's manufacturer, when their shiny new graphic card could not withstand the test at all - it blackscreened instantly and required a reboot. This was due to a design failure of their reference design board. I don't know why the term stuck, despite being wrong. GPU:3D is a stability test just like the others.

This is intended - Getting rid of the display window enables the test to run on graphic adapters that aren't tied to a display, and later on, perhaps test several GPUs at once, even if they are of different brands. Stay tuned !


My VRAM test requires OpenCL 2.0, and my OpenCL version routine isn't checking against the OpenCL version supported by the hardware properly. This will be addressed very soon. If you are running this test against an old GPU ( > 6 years old), check if it is OpenCL 2.0 capable. Overall, i need to improve OCCT by handling those cases better, which is scheduled in a future version.


OCCT's Power test is the most intense stress test you'll find out there. It aims at diagnosing power supply issues by heavily solliciting the most power-hungry components, CPU and GPU.

This test is likely to warm your computer case a lot, which can cause micro-failures in components supporting your CPU or GPU, and generate periodic errors.

Failing this test after a long time is something that happen relatively often. Power supplies, motherboards, and in general components that make up your computer aren't designed to withstand this load for a long time - your components would cost more if they were designed to withstand this, and this would mean a higher price.

While I personally think this is not a good thing, it is common in the hardware industry (for instance, your GPU will never reach the peak performance figures it shows on its box, it'll throttle much earlier than that).

Is it really important ? It's a tough question. It all depends on the load or conditions you want your computer to withstand. You're very unlikely to run into similar loads, so withstanding one or two hours of this test is a good indication of a solid setup all-around.

As this is pretty much opinion-based, I won't give you a definitive answer here.

Just know that I have a lot of examples of computers (including mine) that can withstand 24h+ Power tests, so you can rule out the test as the cause of your errors.

Stability testing since 2003

OCBASE/OCCT - Adrien Mercier

80 rue de l'abbé de l'épée

59500 Douai