4 ms·

What was your motivation for 128GB? VMs are the only reason I sometimes struggle with 32GB, but that’s not been an issue often enough for me to do anything. If

by bpye 3y ago

What was your motivation for 128GB? VMs are the only reason I sometimes struggle with 32GB, but that’s not been an issue often enough for me to do anything. If I really need some more memory it’s easy enough to spin up some cloud instance for a few hours instead.

cesarb 3y ago
> VMs are the only reason I sometimes struggle with 32GB, It's always annoyed me that the memory allocation for VMs is so static. If I allocate 16GB to a VM, it will fill these 16GB with its own filesystem cache, even when the host could make a better use of that memory (like using it for other VMs). There's virtio-balloon, but it has to be adjusted manually.
- ta988 3y ago
  And balloon wastes a lot of memory
fbdab103 3y ago
Not parent, but I do data analysis where more memory is always appreciated. Essentially all tooling now works with bigger than memory datasets, but you eat a performance cost as data gets paged in/out to disk. When you can load a 50GB dataframe directly into pandas, it lets you do things the dumb way rather than having to spend extra brain cycles figuring out how to compartmentalize the problem to stay within your RAM budget.
- saltcured 3y ago
  Same for me. Back around 2015, I had 64 GB in a workstation because I had to develop and test image analysis tools. These would be working with 3D and 4D dense arrays and rather than paging in/out from disk it was a matter of loading chunks in/out from RAM to the GPU's 12 GB of VRAM and assembling larger results to eventually serialize back to disk. When not working on these tools, it was all just buffer cache for Linux while the actual processes were easily living in 8 GB or so. Even today, I'd comfortably live in 16 GB on a laptop, except once in a while miss the option to just do sloppy large allocations for a single task instead of worrying about chunking, IO, and careful access order.
freedomben 3y ago
VMs were the (main) reason actually! I run Fedora on the host, but I write/maintain/package software for different distros and architectures, so having a VM for each target is really handy. I also like to have a VM dedicated to apps/tasks that I want to segregate from the main system, so I end up with quite a few VMs and several of which I want to have running concurrently. Another reason was to enable me to hack with performance optimizations like mounting freqent read/write files into memory (tmpfs). Putting Chrome's cache there for example can be a noticeable increase in performance. Another reason was just for the sheer badassness of having so much RAM
pixl97 3y ago
I use my setup as a 'pre-cloud' test environment where my costs are fixed. I slap $400 worth of memory in my workstation and I can load large datasets or piles of VMs without worrying that some mistake on my part is going to generate a huge bill on my part. It's pretty common for me to run 60-80GB worth of VMs during the working day, so not using could is a pretty massive savings for me.

by bpye 3y ago

cesarb 3y ago

> VMs are the only reason I sometimes struggle with 32GB, It's always annoyed me that the memory allocation for VMs is so static. If I allocate 16GB to a VM, it will fill these 16GB with its own filesystem cache, even when the host could make a better use of that memory (like using it for other VMs). There's virtio-balloon, but it has to be adjusted manually.

ta988 3y ago
And balloon wastes a lot of memory

fbdab103 3y ago

Not parent, but I do data analysis where more memory is always appreciated. Essentially all tooling now works with bigger than memory datasets, but you eat a performance cost as data gets paged in/out to disk. When you can load a 50GB dataframe directly into pandas, it lets you do things the dumb way rather than having to spend extra brain cycles figuring out how to compartmentalize the problem to stay within your RAM budget.

saltcured 3y ago
Same for me. Back around 2015, I had 64 GB in a workstation because I had to develop and test image analysis tools. These would be working with 3D and 4D dense arrays and rather than paging in/out from disk it was a matter of loading chunks in/out from RAM to the GPU's 12 GB of VRAM and assembling larger results to eventually serialize back to disk. When not working on these tools, it was all just buffer cache for Linux while the actual processes were easily living in 8 GB or so. Even today, I'd comfortably live in 16 GB on a laptop, except once in a while miss the option to just do sloppy large allocations for a single task instead of worrying about chunking, IO, and careful access order.

freedomben 3y ago

VMs were the (main) reason actually! I run Fedora on the host, but I write/maintain/package software for different distros and architectures, so having a VM for each target is really handy. I also like to have a VM dedicated to apps/tasks that I want to segregate from the main system, so I end up with quite a few VMs and several of which I want to have running concurrently. Another reason was to enable me to hack with performance optimizations like mounting freqent read/write files into memory (tmpfs). Putting Chrome's cache there for example can be a noticeable increase in performance. Another reason was just for the sheer badassness of having so much RAM

pixl97 3y ago

I use my setup as a 'pre-cloud' test environment where my costs are fixed. I slap $400 worth of memory in my workstation and I can load large datasets or piles of VMs without worrying that some mistake on my part is going to generate a huge bill on my part. It's pretty common for me to run 60-80GB worth of VMs during the working day, so not using could is a pretty massive savings for me.