Tech Behind ThingsHow the ordinary machinery actually works

Devices

More memory does not mean more apps stay open

Phones and laptops manage memory by predicting what you will need, and the prediction is tuned for battery life as much as for speed.

Close-up shot of a smartphone screen showing various app icons, indicating digital technology use.
Photograph by Amarnath Radhakrishnan via Pexels
Editorial note. Independent reporting and analysis. Nothing here is sponsored or paid for. How we work.

This looks at memory management from the practical end — what holds up once conditions stop being ideal.

What holds up in practice

  • Unused memory is wasted memory, so operating systems fill it deliberately.
  • Background apps are suspended or terminated by policy, not by capacity alone.
  • Memory compression and swap change what a shortage looks like.

Free memory is not the goal

An operating system that leaves memory empty is holding a resource that costs the same whether it is used or not. Caching recently used files and application state in that space makes the next access far faster than reading from storage. This is why memory usage graphs on a healthy machine look alarmingly full, and why they should.

The number worth watching is not how much is used but how often the system is forced to evict something it wanted to keep.

Background apps are governed by policy

Mobile operating systems suspend applications that are not visible, freezing them so they consume no processor time. A suspended app still occupies memory, and the system terminates it when it needs space or when policy says it has been idle too long. Those policies are tuned aggressively for battery life, because an app allowed to run in the background costs power continuously.

This is why an app reloads from scratch on a phone with plenty of free memory, which looks like a bug and is a decision.

Compression came before swapping

Rather than writing pages to storage, systems now compress inactive pages in memory, which is far faster than any storage device. A compressed page typically occupies a fraction of its original size, effectively increasing usable capacity at a small processor cost.

Under load, the visible symptom of memory pressure therefore shifts from disk activity to processor activity, which confuses old diagnostic habits. When compression is no longer enough, the system falls back to writing pages out, and that is when things become obviously slow.

Swapping to flash has a cost beyond speed

Writing pages to storage consumes endurance on the flash, which is one reason mobile systems avoid it and prefer to terminate apps. Desktop systems swap more readily because their storage is larger and more replaceable. A machine that is swapping constantly feels unresponsive in a distinctive way: brief total stalls rather than uniform slowness.

That symptom is the clearest signal that adding memory would actually help.

Why the numbers do not compare across platforms

Different systems account for shared libraries, caches and compressed pages differently, so a usage figure means different things on different platforms. Runtimes that manage memory automatically hold onto freed memory rather than returning it immediately, which inflates reported figures. Comparing a memory total between two operating systems is therefore close to meaningless without knowing the accounting.

Mechanically, the honest comparison is behavioural: how many applications stay resident and how often things reload.

Implementations differ, and vendors are not obliged to document the differences.

When adding memory actually helps

It helps when your working set genuinely exceeds capacity, which shows up as constant swapping or repeated reloading of things you are actively using. It does nothing when the machine is limited by processor, storage or network, which is most of the time on a modern device. On phones and most modern laptops memory is soldered, so the decision is made once at purchase.

Choosing capacity for the heaviest thing you routinely do, rather than for a headline figure, is the practical rule.

The takeaway

The system is not hoarding memory. It is caching, and then obeying a battery policy you did not set.

The constraint is almost always physical, and marketing rarely mentions which one.

Questions readers ask

Should I close background apps to speed up my phone?

Generally no. Closing a suspended app forces a full reload next time, which costs more processor and battery than leaving it. Close apps that are misbehaving, not apps that are merely open.

Is high memory usage a problem?

Only if the system is evicting things you are actively using. Full memory with no swapping is a system doing its job well.

Devicesmemoryramoperating systemsperformance
Farida Osei
Networks writer, Tech Behind Things

Farida writes about wireless standards and spent six years in network engineering before switching to explaining it.

Also by Farida Osei