if you reduce the memory usage it will take longer to compile, right?
nope, depends on available ram. your main enemy is disk io. for example if a one core compilation stay on ram will be faster that 16 core full speed compilation which involve virtual memory.
If you don’t have enough ram it will attempt to use IO, but it uses the ram it requires. I don’t think reducing the memory usage and compute more will be faster than the IO penalty
in my case is faster. even memory compression is faster in my hw
Do you have ssd?
-MP1 didn't work?
that's for cl not the msbuild, with cl directly it needs a lot out of sln configs
yeah but I am sure you can pass commands to the compiler directly
found this, https://github.com/Microsoft/checkedc-clang/wiki/Parallel-builds-of-clang-on-Windows#msbuild hmm, it uses CL_MPCount
aw so this way you can pass commands to cl via msbuild, interesting, thanks
That's specific to -MP. Let us know how it works out for you :)
i will test it and tell you. thanks again.
i've tested inside the ide config before but did not work. i will test the commandline soon and tell you.
oh, if the ide config didn't work, then i doubt command line arguments will
maybe it's just a bug in the ide
it seems that ue4 source code in windows has a complex undocumented config system, their sln works internally with make files and some custom selfdocumented configs in c#. maybe that's why msbuild config does not work. still i did not find a unified make file and in their c# configs there is nothing about compilation threads/processes. if you know something about their config system please let me know.
Nope i dunno much about UE4 internals.
Обсуждают сегодня