1

I did a 4.2 dev build myself about a month ago since I'm running Ubuntu 18.04. Since I don't have a machine here suitable for building it, I spun up an instance in the Google Compute Cloud with 1 CPU, 6 GB of RAM, and a 100 GB disk. It took around 4-5 hours for the complete build including all prerequisites (LLVM etc.).

Since that build is getting a bit long in the tooth I was thinking of doing another build and I was wondering if I could speed it up by providing more CPUs. Are there any options I need to pass to the build system to tell it to make in parallel? Or is that even possible?

Also, how much should I bump the memory up by if I build in parallel? I realize there's no definitive answer, I'm just looking for insight from those who I'm sure have much more experience building it than I do.

Finally, any idea what the timeline for a final release of 4.2 and/or support for Ubuntu 18.04 is, or whether there have been significant changes to the 4.2 branch in the last month or so? I guess what I'm asking here is, is it really worthwhile for me to do another 4.2 build, or should I just keep using the 4.2 build I've already made until an official one comes out? Context: I'm just experimenting with the language for my own edification.

Thanks for your time. All input welcomed.

2

The build system will automatically detect the number of CPUs and use them. That being said, one will observe diminishing gains with each additional CPU. I'd budget about 300-500 MB of RAM per core and up to 70 GB of disk space on Linux for Debug+Assert results and about 4 GB of disk space for Release+Assert builds.

1 Like