Page 2 of 2

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 12:02 am
by Graf Zahl
There shouldn't be such a massive change between 3.2 and 3.3, though.
Can you do some 'bench' dumps from an affected scene? Maybe that gives some clues.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 10:51 am
by JDN
OK my computer is too old too play a 1993 game...
damn. :lol:


edit: "Can you do some 'bench' dumps from an affected scene? "
how can I ?

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 11:03 am
by Graf Zahl
open the console, type 'bind b bench' (or any other key if you do not want to use 'b'), then make a savegame at a place where this slowdown happens, load it into both engines and press 'b'. A file named benchmarks.txt will be saved. Post that here.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 11:17 am
by JDN
OK done:
one benchmark for my "working" version

and two with my "laggy" version
inside where it seems to work and outside where i tt laaa a aaag

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 11:32 am
by Rachael
JDN wrote:OK my computer is too old too play a 1993 game...
damn. :lol:
This remark is completely disingenuous and omits some very important things.

Your computer is too old to use a 2018 program that happens to play a 1993 game.

You can still use doom.exe to play a 1993 game on a 1993 computer.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 11:38 am
by JDN
Yes yes sorry, it was a joke... :oops:
I know GZDoom is a modern (and very well done program).
I didn't want to be impolite with people who create this masterpiece.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 12:02 pm
by Graf Zahl
The strange thing here is, that nothing in there suggests such a performance hit, the map basically contains nothing and even the increased vertex count due to the new approach is nowhere near problematic.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 1:33 pm
by JDN
maybe my computer was just at the limit it can bear with 3.2.5...
and maybe the updates added in new versions were the "liitle thing" that bring me some lags.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 1:46 pm
by Graf Zahl
Hard to believe. If this was a map that actually performed some real work that might be the case.
Just out of curiosity: How does this system handle some real maps? Although your card is relatively weak by today's standards it's not that bad for an old game like Doom. Something must have changed in GZDoom to affect this particular system this badly and I have no idea what it is or whether it may affect others too.

Re: JDN's performance issue [split]

Posted: Tue Jun 26, 2018 1:52 pm
by JDN
by real maps, can we say the Hunter's Moon maps ?
that is real heavy work for my CPU... but it works well when no monster run somewhere.
just a little lag.
Hunter's Moon map + Hunter's Moon weapons and monsters... and there, i'm reaching the limits of my CPU.

Re: JDN's performance issue [split]

Posted: Fri Jun 29, 2018 9:43 am
by JDN
I spend hours searching for settings that can help but nothing.
But I have fullspeed when using software rendering. :?

So now I have two folders, one for my 3.2.5 and other one for new versions.