What is the quickest way to diagnose whether a slowdown is hardware or software?

[Quanta]

Trying to avoid wasting hours on fixes that make no real difference.

Most of the advice I have found online is just recycled from the same few sources.

Happy to provide more detail if it helps narrow things down.

Curious what the consensus is

[Quanta]

That is pretty much what I found too. That is the sensible starting point

[Lucy05]

Solid advice that. I have found that the biggest savings come from the boring stuff nobody wants to do.

Might save you more than you think

[RedKnight]

Cannot disagree with that. Form matters but so does the matchup and that gets ignored a lot.

Cannot wait for the game to settle it.

Background processes and startup items cause more problems than hardware failures in my experience

[Quanta]

Cannot disagree with that. Form matters but so does the matchup and that gets ignored a lot. Cannot wait for the game to settle it. Backgrou

Worth checking that assumption before committing to it. The fastest fix is often just checking what is running in the background and killing half of it.

Happy to help further if you get stuck

[Ellie22]

Okay that makes more sense than what I had in my head. I appreciate people explaining the detail rather than just the headline.

Cheers for the explanation

[Beth3.0]

That is the approach I always take now. Good luck with it

[SGHolly]

Okay that makes more sense than what I had in my head. Appreciate the detail.

[Aaron]

Fair enough. Thanks for that

[RedKnight]

Okay that makes more sense than what I had in my head. Appreciate the detail.

Still think the same, yeah. People always overreact to the latest result, which is half the problem with these debates.

We will see how it plays out.

The event viewer usually tells you exactly what is happening if you know where to look

[Red Builder]

Feels like the right read on it. I find the financial angle of any big story is usually the most underreported part.

More to come on this I suspect

[QuantumLeap]

Same thing happened to me. Would recommend giving it a go

[Aaron]

Couldn't agree more. Appreciate it.

[Owen84]

Pretty decent summary of it. Might go back to it.

[Steady Dylan]

Yeah I can see that now. That is genuinely useful

[MickFoley00]

That is pretty much it. Proper useful that.

A lot of Windows issues sort themselves with a fresh install of drivers

[Nina24]

Counterpoint to the live USB advice: it's useful but not as clean a signal as people think. A live OS boot often runs entirely in RAM and doesn't stress the storage subsystem the same way your installed OS does. So you can boot a live session, think everything is fine, reinstall Windows, and still have a slow machine because your HDD is quietly dying. The live boot test is a decent starting point but don't treat a positive result as a hardware all-clear.

The more reliable hardware check for storage specifically is CrystalDiskInfo on Windows, which reads the drive's SMART data. Look for reallocated sectors, pending sectors, or uncorrectable errors. Any of those in anything other than zero is a yellow or red flag depending on the number. For RAM, as mentioned, MemTest86 is the standard. For CPU and general thermal issues, check whether your CPU is running at significantly below its base clock speed under load, which indicates thermal throttling, and that's usually a heatsink or paste problem rather than a dying chip

[Connor82]

Can we also just acknowledge that the answer is sometimes "both" and that's the most annoying outcome to diagnose? Dying hardware often causes software symptoms. A drive that's starting to fail might read slowly, which makes the OS take forever to load files, which looks like the OS is broken. Flaky RAM causes crashes and corruption that look exactly like software instability or even malware. The hardware fault is upstream of everything so it produces software-looking symptoms all the way down.

This is why the standard advice of "just reinstall the OS" as a first response to slowdowns drives me insane. If there's a hardware fault causing the problem, you've just wasted several hours reinstalling onto a broken system and the problem comes back within days. Diagnose first, fix second. The live boot and SMART check take maybe twenty minutes total and will save you hours of going in circles

[ProperJobs]

Temperatures are massively underrated as a first diagnostic step and I never see them mentioned first. Download HWMonitor or HWiNFO, run a heavy task, and watch what your CPU and GPU temps do. A CPU thermal throttling down to half its rated speed because the heatsink is clogged with dust will make a machine feel ancient. A GPU doing the same makes games feel like they're running on a potato. This takes about two minutes to check and can instantly point you at a physical cause that costs nothing to fix except a can of compressed air.

I've seen machines that looked like they needed a full OS reinstall turn out to have a heatsink so packed with dust that the CPU was thermal throttling from the moment any real load hit it. Clean the machine, repaste the CPU if it's been a few years, and suddenly it performs like new. Always worth ruling out before you go anywhere near software troubleshooting

[CollapseState87]

The fastest first step is a live OS boot. Grab a Linux distro on a USB stick, boot from it, and see how the machine performs. You're now running an entirely different software stack with no installed drivers, no startup programs, no registry rot, nothing from your existing install. If the machine still crawls, you've got a strong signal pointing at hardware. If it's suddenly snappy, your hardware is probably fine and your installed OS environment is the problem. This single test eliminates half the decision tree in about ten minutes.

From there, if you suspect hardware, run a RAM test like MemTest86 overnight and pull up a tool like HWiNFO or CrystalDiskInfo to check drive health and temperatures while you're at it. Failing RAM and dying drives both masquerade as general sluggishness in ways that look suspiciously like software problems. A drive sitting at 100% activity in Task Manager is a classic one that fools people into reinstalling Windows when what they actually needed was a new SSD

[RandyOrton26]

Slightly left-field suggestion that has saved me personally more than once: check the power plan on Windows. Somehow, on a surprising number of machines, the power plan gets set to "Power saver" either by a Windows update, a driver install, or just someone clicking something they shouldn't have. Power saver mode deliberately throttles the CPU and can make a perfectly healthy machine feel genuinely terrible. It takes ten seconds to check in Control Panel and change to "High performance" or "Balanced."

Similarly on laptops, check whether you're running on battery when you think you're plugged in, or whether the battery is so degraded that the system is throttling to compensate for it not being able to deliver peak power. Battery health issues causing CPU throttling on laptops is a real thing and not widely known. Both of these are software-adjacent causes that have nothing to do with your OS being broken or your hardware actually failing, they're just configuration or battery condition issues that are trivially fixable once you know to look

[Coastal Current]

Software side of the diagnostic tree: open Task Manager or Resource Monitor and just watch it for a few minutes while the machine is slow. You're looking for which resource is pegged. CPU at 100%? Something is eating it, find the process. RAM maxed out? You're swapping to disk and you either have a memory leak, insufficient RAM for your workload, or both. Disk at 100%? Classic sign of either a failing or very slow drive, or something hammering it with I/O. Network maxed out unexpectedly? Something is uploading or downloading in the background.

The 100% disk usage problem in particular became incredibly common after Windows 10 launched and a lot of people blamed hardware when it was actually a combination of Superfetch, Windows Update, and Defender all deciding to do their thing simultaneously on machines with spinning hard drives. If you see disk at 100% but CrystalDiskInfo shows the drive is healthy, you're almost certainly looking at a software issue. Disabling Superfetch or upgrading to an SSD fixed it for millions of people

[Cass]

Here's the quick and dirty version for people who don't want to do a deep dive: does the machine slow down immediately on boot before you've opened anything, or does it start okay and degrade over time during a session? Slow from the first moment usually points at startup software, malware, or a hardware fault that's constant. Starts okay then degrades usually points at a memory leak, thermal throttling kicking in as things heat up, or something accumulating in the background over time.

That single observation narrows the field enormously. Starts slow from boot: run a startup item audit and a malware scan before touching anything else. Degrades over time: watch your temps and your RAM usage over a thirty minute session and see which one climbs. It won't tell you the exact cause but it points you at the right category of problem in under a minute of thinking

[Craig]

One thing I haven't seen mentioned yet: check whether the slowdown is specific to one application or system-wide. If it's one app that's slow and everything else is fine, that's almost certainly not a hardware problem. It's either a bloated or corrupted install of that specific application, a settings issue, or an update that made it worse. Reinstalling or resetting that app is the first move.

If everything is slow uniformly, you're looking at something lower level, which is where the hardware vs OS environment question becomes relevant. And if the slowdown is specifically tied to a particular kind of task, like anything involving large files is slow but small operations are fine, that's a very specific pointer at storage throughput or RAM capacity. Symptoms that are selective are actually helpful, they're narrowing the possibilities for you if you pay attention to the pattern