Fast page outs won't help you when you are trying to run multiple VM's. The memory used by VM's is wired and cannot be swapped.
Even in such an extreme case, a 4 GB allocation for each VM would do the job. You won't play heavy games on both VM at once, right?
For just browser testing typically 4GB instances is fine. So you can keep 2 running comfortably.
Many docker containers don't have a huge memory footprint when running either because they are more io bound than CPU or memory bound.
The difference between 16 GB and 32 GB is 2 virtual machines if they need to be 8GB VMs.
The scenario where 32 GB is big enough where 16 GB isn't and not itself too small is narrow. If I need 4 8GB VMs to do testing then 32 GB is still too small.
If you need less than 8 4GB VMs but more than 3 at the same time it's useful.
Someone should put these 'limitations' vs. 'improvements' to the test. Have a huge video file or photoshop file and run the test. How hard does this laptop need to be pushed before the 16GB RAM limitation becomes something meaningful. I'd be interested in seeing this, just for nerds sake. Am I going to have to wait for an extra 1/10th of a second for Photoshop to run some type of exotic blur on my 30" x 20" high-res image? Or are frames going to be dropped while doing 4K video?
Like I said, I'd love to see how 32GB RAM would improve the extreme case. For me, I can't even push my iPad to the max, so I'm in Apple's sweet spot.
I was hoping we'd hit new price points (meaning more affordable). But we didn't. I just need a week to cry and absorb the enormous sum that's going to get sucked out of my credit card. Then I'm going to really enjoy that laptop. It's my 2nd laptop, the first being the last G4 whatever-it-was.
I'm not sure if we'll ever see these extreme-case-tests. If someone can offer up a guess, I'm all eyes. Is the 16GB limit going to add minutes to your workday? Seconds? Hours? In what case?
Here are some 16GB-related data points for tense multi-tasking situations by a techie: https://www.zdziarski.com/blog/?p=6355 We now await truthiness replies from the hi-mem sycophants.
Comments
Many docker containers don't have a huge memory footprint when running either because they are more io bound than CPU or memory bound.
The difference between 16 GB and 32 GB is 2 virtual machines if they need to be 8GB VMs.
The scenario where 32 GB is big enough where 16 GB isn't and not itself too small is narrow. If I need 4 8GB VMs to do testing then 32 GB is still too small.
If you need less than 8 4GB VMs but more than 3 at the same time it's useful.