Actually, I am on Ubuntu Mate, on an old MacBook Pro (mid-2012) 13", and looking at Conky.
Though, I have upgraded hardware parts to 480GB SSD on SATA port, and 8 GB RAM.
Requirements are
- Web development, including the backend
- Docker
- VirtualBox, if required, for e.g. Power BI, Microsoft Office
- Biomedical informatics course, which includes Data Science and GIS
I have noticed from Conky, though; that RAM doesn't even though the 4 GB bar, but CPU (core i5, 3rd gen, 2.50 GHz, 4 virtual cores) did sometimes fill all with 100%.
However, even when not all filled, while running sudo apt upgrade, the computer did sometimes lag.
I have also noticed that Swap, which I allocated at 4 GB, is only filled by 91.5 MB.
I might have things I haven't monitored, such as
- GPU
- Current HD read / write
Currently, the most important question is, will by MacBook Pro will lag when running macOS Catalina or Window 10?
macOS is just like a time bomb. It gets slower every upgrade. That's why my sis stopped using this Mac...
Latest comments (5)
Having low swap utilization is a good thing, it means you actually do have enough usable memory for what you're doing (though 8GB is a bit low IMO for both containers and virtualization, especially if you're doing web development).
For the record, you actually have a third generation Core i5, not second (the mid-2012 MacBook Pros all had Ivy Bridge chips, not Sandy Bridge). Given that it's a 13" model, it should be a Core i5-3210M. However, a third generation Core i5 is also about equivalent in terms of practical performance to a current generation Pentium (in fact, you're just above the performance envelope for a Pentium Gold 6504U, which is only really used in netbook and chromebook type systems, except it's got a lot lower power consumption and supports much faster memory). That by itself is almost certainly 95% of the issue, as it both severely limits the memory speed (DDR3-1600 hasn't been top of the line for quite some time, and a new system will almost always have at least DDR4-2133, or more often DDR4-2400) and just generally limits your processing power (both CPU and GPU, keep in mind that the second generation Core iX chips predate Intel actually providing an iGPU that was good for anything beyond basic web browsing, but the hybrid GPU setup).
As far as how things will behave under Windows or macOS, expect lag given that it's a 7 year old computer. macOS does not handle old hardware very well, and Windows can usually be assumed to perform worse than Linux for almost anything except gaming (there's a reason supercomputers don't run Windows, and it's not just cost). On the flip side though, you do just barely meet the requirements for macOS Catalina or Windows 10, so you could run them, they'd just be really slow.
Swapping only occurs if your RAM is full. Most install guides suggest deactivating swap altogether these days.
When working with heavy node tasks, a high cpu clock speed is important, because node doesn't really do multithreading. More cores come into play when working with docker.
Your GPU is mostly not being hit. Apart from training AI models, development isn't gpu-intensive.
The most important piece of hardware for me is a fast SSD. That actually speeds up the times needed for file operations - which is 90% of what I do day to day.
So yeah, your virtualbox hits hard on your CPU, but you're fine with everything else.
I disagree. RAM makes a big difference in a virtualized OS (especially Win10 in Virtualbox - easily testable by increasing allocated guest RAM).
You can gain performance at the cost of some features being harder to configure (shared folders, clipboard, though both can still be accomplished) by using KVM (Kernel Virtual Machine, one of my favorite Linux features) instead of VBox, as it is closer to bare metal.
I normally allocate 4GB to Windows on VirtualBox, on another MacAir (around 2018, I think).
And when running PowerBI, it still lags.
Yes, I also use Node.js and Docker.
Thanks for explanation, though.
I don't yet know much about AI models, but I am currently learning about Decision Tree Analysis, Clustering, and Association Mining. But indeed, I heard about GPU requirements from Deep Learning, on YouTube.
I also tempted to try DaVinci Resolve (VDO editing).