Discussion related to the Engine functionality, implementations and API.
- Posts: 1112
- Joined: Mon Dec 04, 2017 10:08 pm
- Location: Germany
I'm currently compiling LuxCore on Windows.
I have 12 GB RAM, about 2.6 were taken by the system when starting the compilation, so about 9.4 GB free.
Up until now, it would only hit the RAM limit once for a short time (swapping a bit), but after the changes in the last weeks my system now spends about 15-20 minutes just swapping with 1% CPU usage.
I'm using the standard WindowsCompile cmake script.
So my questions are:
- Is this working as expected? Or is my system somehow misconfigured? How much RAM does the compilation need on your systems?
- Is there a way the compilation process can be improved so it takes less RAM (and maybe is faster)?
- Posts: 1089
- Joined: Mon Dec 04, 2017 8:36 pm
I think that since the Toby's cmake patch it runs more link phases in parallel instead of single one. But I have no problem to compile everything with 16GB of ram so there may be something wrong if you have all that problems with 12GB.
1) Run cmake scripts
2) Ctrl-C when it starts to compile .cpp (so the VisualC++ project has already been created)
3) Open the project WindowsCompile/Build_CMake/LuxCore/ALL_BUILD.vcxproj with VisualC++ 2013
4) Right click on pyluxcore (or any other target) and select "build"
For a very fast compilation, select also "Debug" build over "Release" (the link phase with LTCG is a LOT faster).