I alredy nodraw some stuff and set hammer running priority to real time what else i can do?
Well thats pretty unclear
a nice video on func_visclusters https://www.youtube.com/watch?v=Ds5E7HBQ-nU
Identify which parts of your map is making VVIS throw a fit and find a workaround by either simplifying the geometry or compiling a fitting model.
If you can’t pinpoint the location of your map that is giving you issues , try moving a cordon around your map until the compiling time spikes.
Also func_detail is your friend in any and all circumstances.
If your compile is slowing/getting stuck on the Portal Flow, you gotta optimize the map, by turning into func_detail complex shapes (with more than 6 faces). And stuff that create useless visleafs (barriers for example)
Your map without entities, props, func_detail, displacement and water should have very basic shapes. You can use the Auto Visgroup window, to disable/enable things you want.
Before compile, dont forget to re-enable everything.
func_detail is one of the most important things.
Secondly, yes, func_viscluster, for open spaces. Third, you can get into advanced hint brushing, but I recommend you get into the first two first.
Or just get a better CPU. ;0
I got really good pc
“a really good PC” is just as helpful as saying as you got PC. If you don’t have a strong single threaded processor, you don’t have a “really good processor” for the compiling tools. So unless you backup your numbers with processor/memory capacity then its really not going to go anywhere.
If you do have a really strong processor. Then your map is poorly optimized for the tools. No computer is going to help you out if you’re going to be sloppy. Good computer or not, optimization should always be done.
Compiling tools are multi-threaded (although they are 32-bit), but that’s besides the point. One of the reasons I got an i7-4790k is so I can compile what used to take me an hour or two in a couple of minutes.
Also, question, how long are your compiles even taking you?
This all depends on what you’re compiling. Throw a map out there that’ll allow us to benchmark compile times across a variety of hardware. I got quite a few processors ranging from 2x8t or 16t xeons to desktop processors to provide you some results.
One thing I’ve considered is grabbing a 32 core vps from vultr for like an hour.
The compile tools only support up to 16 threads.
Also for price, grab spot instances on the cheap for a couple hours in AWS.
By cheap I mean, ~$0.8/Hr (Plus drive space, bandwidth, etc) for EC2, assuming a Spot instance in US-West-2.