Your favorite free games to test hardware and OS on?

[ChickenTacos]

Greetings.

What are your favorite, free Steam games that you use and run and game and operate to test PC hardware, or your VR hardware, or your device hardware (of any form, of any place), or, perhaps even your operating system, or operating systems? Tell all the games, as well as tell all the reasons why you use SAID games to test the hardware and OS(s). Thank you, all!

[JacquesPatat]

free game to test system, i take it that part is a joke, don't think there is many resource hungry free games (can be mistaken so please don't fight with me ).

can't remember the last time i used a game to test my hardware/OS but if i have to test my hardware then i prefer to use the industry standard and most probably everybody favourite piece of software 3dmark to test the system out and it gives you some technical info regarding the test what it ran

[Illusion of Progress]

None in particular. I use whatever games I decide to want to try/may be playing at the time to round out any testing.

But games would be a rather poor choice to end testing stability at. They're great to include; awful to start and end at.

Initial, but not conclusive, testing is often done with stress testing applications. These push power consumption, heat generation, and utilization far better than games do, which really only push the GPU and somewhat the CPU (but usually not if you have a higher tier CPU). Games simultaneously are and aren't good for stability testing because they push hardware too variably, so this makes them good to use because they do catch things that static or synthetic benchmarks and stress test applications may miss (though many of these are including "variable" types of testing), but not good on their own.

OCCT has a suite of tests it can do. CPU, RAM, GPU, etc., etc. I'd say it's pretty good all around.

MemTest86 is good for RAM but I usually don't test that unless I have a reason to believe it may be unstable.

Prime95 (namely with small FTTs) is good for the CPU cooling capability and stability.

There's other good ones, but here's some I wouldn't rely on.

CPU-Z's stress test is weak. Its benchmarking is also known to be poor[chipsandcheese.com]. Use it for information gathering only.

Passmark is a reliable source in general, but one good example of an exception to this is that its own single core test poorly measures cache just like the above. Yet cache has been one of the most recent breakthroughs in performance for IPC, even if it's limited in where it applies. So how good is an IPS measurement when it's poorly measuring something that can heavily impact IPC, which itself is half of what makes up IPS? Simple, it's not. It gives a measurement as though the cache doesn't exist. The lesson is synthetics are just that and need an awareness of what/how exactly it is measuring it. And that's often not disclosed or well understood.

AIDA64 is rather middle of the road at best for testing cooling capability and stability. I think the FPU test is fair but the rest are weak/easy to pass (at least for the CPU).

Userbenchmark shouldn't even be given traffic. It's neither a good stress tester nor a good information getter.

Games are good for the GPU mostly, but there's also OCCT's suite, Furmark, or your choice of benchmark programs (like 3DMark/TimeSpy/whatever it is people use these days).

Games should round this out.

If it passes some key stress tests and a fair enough collection of games, it's probably rather stable. Until it's not. The old saying is "a system is stable until it's proven it's not". Sometimes one game comes along and pushes "seemed to be stable before" systems into unstable territory. Doom3 was infamous for this to CPU overclockers back then. Unreal Engine shader compilation is doing this to some Intel 13th/14th generation Core i7/i9 CPUs now. Unfortunately, there is no one all conclusive "test this one thing and it's 100% stable for certain if it passes" method. And if there was, games certainly wouldn't be it.