Random Chitchat 2012-2016

[Lapayo]

True, I'm running in Debug mode which has its penalties, but I don't think they'd be so big.

Well, On my system the difference is huge.
In Debugmode it takes me about 30 sec to be able to join. (30Sec to be up)
In Releasemode the server is up within a few secounds O.o

[NiLSPACE]

he's right on release mode it takes just one second thanks Lapayo

[xoft]

I'd consider one order of magnitude to be the normal difference between the debug and release versions. But with the new trees, it's more like three or four orders of magnitude - instead of 20 chunks per second, I get one chunk in a minute?
Oh, and I'm comparing old debug to new debug, not old release to new debug, so the difference is even worse.

STR_Warrior, Release version is built using VS, menu Build -> Configuration manager -> select "Release" as the Active solution configuration at the top combo; then compile like normal. The resulting exe will be Release. Don't forget to switch back to Debug if you want breakpoints and watches to work

[NiLSPACE]

i now have it on release mode but there is a bug (also with debug) when it generates terrain it uses 30% of my cpu sometimes even more

[xoft]

Well, what would you expect? Terrain generation is the most difficult part. Especially with your trees
The Debug doesn't use that much CPU, because the debug runtime has a serializing memory allocator - each memory allocation locks all other threads from allocating memory until it finishes, causing all threads to serialize into at most one core, if they are allocating memory too often.

[NiLSPACE]

ooh so that is why the generation is about 30% slower if you have marblecaves on?

[xoft]

No, marblecaves are computationally intensive in their nature - for each and every stone block in the chunk, several math functions are calculated to find out if the block should be dug out or not.

[NiLSPACE]

and does that also happen with the upcomming ravines??

[xoft]

Unknown yet. Probably not.

[NiLSPACE]

and that is good right?