Can also save tons of space with `cp --reflink` if your filesystem has support (XFS and Btrfs but not ext4). It's essentially fork(...) for files. It uses block-based copy-on-write at the destination. Can copy an identical file over another to deduplicate without sharing writes.
Conversation
I moved to XFS from ext4 for better parallel NVMe performance and started using reflinks a lot. XFS doesn't have transparent compression support yet but I found it hurt build performance too much and most of my space is used by already compressed Git object/pack data anyways.
1
1
4
1
2
actually reducing the underlying disk space used by git wasn't implemented but would be possible by lazily fetching git history and otherwise maintaining a shallow clone. it would be a very large project but it would be fun to work on
1
1
XFS delayed allocation + delayed logging combined with 128GB of memory and a fast NVMe SSD means that barely anything ends up being I/O bound. End up with huge sequential writes as build outputs as flushed out. Reads mostly hit page cache. SSD stats have 10:1 write:read ratio.
1
2
Generating delta updates for GrapheneOS has parts that are extremely I/O bound unless it's done in tmpfs and 128GB memory is not nearly enough without using fewer jobs than CPU cores. It'd already go OOM without tmpfs with 1 job per thread (32) instead of 1 job per core (16).
2
3
this is also where the use case for big corps vs individual devs comes into play; at twitter we offloaded stuff onto remote process executions via the build tool but that's not available to most people + it requires rewriting the entire build for a remote enabled build tool
1
2
I need to do official builds, signing and delta generation locally for security and trust reasons so that's inherently bottlenecked by my local compute resources.
I wanted to make a 64 core Zen 3 Threadripper machine to help with this but Zen 3 Threadripper was never released.
2
3
Currently have Ryzen 5950X (16 core Zen 3) with PBO2 enabled and i7-6950X (10 core Broadwell-E). That Intel CPU from 2016 takes 2x longer so the 9 official builds are split 6:3 (previously 9:4 before Android 13).
Part availability is the real blocker rather than it being cost.
1
2
Canadians aren't allowed to buy nice things. It's really hard to get computer parts without a huge price premium especially with the chip shortage. Couldn't get Zen 2 Threadripper parts, then wanted to wait for Zen 3, which got delayed a year then never launched to retail at all.