1

I have a decent spec MacBook Pro:

The MBP has 64GB of RAM.

enter image description here

This is the output of htop command after 6 days uptime:

As you can see, the Swp usage is quite high.

enter image description here

Note: I do have Docker and JetBrains PyCharm running often.

Is the Swp usage impacting performance, and how do I fix this?

After restart, it's fine for a while, but then reaches a critical point where even Safari lags a lot.

Update

As requested, here are some pics of Activity Monitor:

enter image description here enter image description here enter image description here

6
  • I'd do better if I could see pics of Activity Monitor, showing All Processes, the CPU & Memory tabs, sorted appropriately [I've no clue what htop shows, except that it appears to be only looking at User processes]
    – Tetsujin
    May 14, 2021 at 14:37
  • @Tetsujin I just added the requested pics - anything jumping out at you? (apologies for the delay - I had to wait until my system became laggy again)
    – iamyojimbo
    May 28, 2021 at 7:56
  • The new pictures would seem to confirm jksoegaard's answer - it's not swap that's the issue. Memory management is all in the green & CPU usage is low.
    – Tetsujin
    May 28, 2021 at 8:04
  • There is nothing in the Activity Monitor screenshots which indicates laggyness.
    – nohillside
    May 28, 2021 at 8:16
  • well, unless you run everything within Docker...
    – nohillside
    May 28, 2021 at 8:18

1 Answer 1

1

Swap usage in itself does not mean that performance is impacted.

You seem to be judging swap usage as "high" based on the fact that 8.19 out of 9 GB are used - however the swap size is automatically expanded as needed. So there's no real "9 GB limit" or anything like that - so you cannot use this as a measure for saying that swap usage is either low or high.

Your performance is impacted when things are actively being swapped in or out while you're waiting for the computer to do something. Data being swapped in/out while you're not using the computer won't hurt the performance you see, just as data being swapped out and staying there won't hurt the perceived performance.

3
  • I see, makes sense - so what is the correct way to objectively monitor the causes of any system-wide performance impact?
    – iamyojimbo
    May 16, 2021 at 13:02
  • There’s no single, correct way of doing that. There are various tools that each can give valid information for determining some causes. So there’s a choice between multiple tools that give the same type of information - and they can all be good to use - as well as a choice of tools that gives different types of information, where you need to choose the relevant ones. If you have no idea what causes the slow-down, you’ll need to check with multiple tools before hitting the “jackpot”. In some cases slow-downs have multiple, unrelated causes.
    – jksoegaard
    May 16, 2021 at 13:07
  • @iamyojimbo The memory pressure graph would get orange or red if your memory management (swap) was becoming difficult for the computer to manage
    – Ezekiel
    Oct 25, 2021 at 16:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .