Hacker News Reader

Comment by jonplackett | Hacker News Reader

I think it depends what they've done with Big Sur.

Take a look at your activity monitor and see how many background tasks you laptop is running right now (I currently have 481 and I only have 4 apps actually running).

Compare that to iPad which can run 2 apps max and apple controls all the background activity.

View on HN

I use 27 Gb of RAM currently on my hackintosh and none of my work VMs are on right now. No way I can get an arm Mac until they're available with 64+.

This. I was hitting a wall on RAM instantly at 32Gb.

Hitting a wall how? The machine stops working properly? Or you see all the RAM allocated in Activity Monitor? Because MacOS aggressively allocates RAM even under a normal workload.

Well when you start getting memory pressure you start losing cycles to compressing and decompressing pages which makes everything run like complete shit.

That’s not how MacOS works. It “keeps the pressure on” so to speak to try and have as much in-memory before you need it. Have you experienced actual issues with specific software?

Out of curiosity, doing what??

IDE + VMs + office + slack + zoom etc

Blows my mind, I do most of the same on a 2013 MacBook Air with only 8GB RAM. I’ve never had a perf hiccup unless I try video editing or gaming. But I just don’t really have a want to do those things, hence why I’m still getting by on this thing. I was going to upgrade once for the hell of it but all the keyboard issues kept me seated.

I expected you to say you did something with graphics/media. I guess the VMs could be the difference. I don’t use. But I have 100 chrome tabs open at most times lol.

All these things work with 8GB RAM. macOS uses all available RAM aggressively, so if you have 32GB RAM it will appear as what you are doing requires 32GB RAM, which is not the case.

No they don't. It starts compressing pages when the memory pressure goes up and then it starts running like dog shit.

On my machine, if it goes more than 5G into swap it starts to become less responsive.

It really depends on your usage patterns. VM abusers and people who need to keep an eye on more than one project at the same time (as in multiple IDEs) can't do with 16 Gb.

And please don't tell me to start closing software, I'm willing to pay for more ram to have everything handy. Except... I can't.

Right now I have 6GB for VMs (for ~12 docker containers) and 8GB for various IDEs and editors. Probably typical for those on my development team. Surprisingly Chrome isn't even the top ten for memory footprint on my laptop.

So far 16GB has been fine on this 2017 MacBook. Wouldn't turn down 32 though :-)

You have reduced performance, you're just accustomed to it.

Could be true. But I get instant feedback from all apps, so what's the measure of "reduced performance" that would matter? I'm actually curious, if nothing lags, would I notice a difference by upgrading?

My view into your comment is that you think I'm used to the lag. That's not the case. It's not my only device just my only Apple laptop. My newest device, a iPhone 12 Pro, is an upgrade from my 6s that I surely noticed and knew I would. I had been frustrated by the 6s for a while but held out for 5G.

Yeah, I found running all my dev environments in Docker/VMs made everything too slow. So I've just installed everything natively and have my own little docker-compose style runner to simplifying running our in-house microservices. I would definitely appreciate decent Docker performance, but it was actually surprisingly painless to set this up.

I don't see this issue having IntelliJ, Xcode, VSCode, Slack, Discord, and more open simultaneously on a continuous issue...with 16 GB of RAM.

I'm somewhat impressed by some of these stories about RAM usage - what are people doing?

RStudio, interrogating a ~0.5 billion record dataset. A 192G Ram machine.

I did some benchmarking on that public taxi rides dump (~600m records) and could run through it all in under 15 seconds directly from SSD. No way in hell your CPU will keep up with anything close to memory bandwidth or beyond 3GB/s.

Using Clickhouse btw.

4 node Kubernetes cluster, vagrant build environment.

Debugging Ember or Spring applications.

The only thing could be some really aggressive swap disk situation offloading anything not needed that second. But that won't work for things where you need all the ram right now, video editing for example.

macOS use RAM aggressively, so it appear as your task requires more RAM than they actually do.

Reply on news.ycombinator.com