InsideQC Forums

Color transition + dithering on the wall seems to actually generate texture detail.
I wonder if a threshold could be set to leave low-contrast neighbors undithered, but may look like blocky compression artifacts in-game.

Source code can be found here at this SVN repository: svn://mancubus.net/svn/hosted/fsfps/tnq/

->

svn: E000111: Unable to connect to a repository at URL 'svn://mancubus.net/svn/hosted/fsfps/tnq'

Try the http protocol.

I guess the server after the hardware failure regressed it, and I should update the page

Cheers!

I swear I tried that earlier without success. Worked now though.

Getting the stuff to work in that traitorous derailment of the series was (almost) a matter of copy and paste.

Does Engoo have the multi-core clock fix (Windows)?

[Un-clock fixed engines will run super speed and then super-slow, jerking up and down in speed and every machine has been multi-core for a few years now (so 100% chance of problem occurring). This is a Windows thing. Quakeworld and DarkPlaces have had this fixed forever.]

Baker: um... what does the multi-core clock fix involve? This is the first I've heard of it, so I suspect it's not in QF.

taniwha wrote:Baker: um... what does the multi-core clock fix involve? This is the first I've heard of it, so I suspect it's not in QF.

Seems likely that QF would have inherited it from Quakeworld.

MH tutorial:

http://forums.inside3d.com/viewtopic.php?p=36239

Thanks, Baker. While QF has indeed inherited the timeGetTime fix from quakeworld, MH's solution looks good. I've just now added it to QF and pushed the changes.

Baker wrote:Does Engoo have the multi-core clock fix (Windows)?

No it doesn't. Probably because the fix came when I wasn't working on engoo (the long two-year stretch of Dec 2010-Dec 2012) and i'm unsure what this will cause for performance for the low end (<300mhz).

However, it does have a different "fix" for Sys_FloatTime inherited from QIP.

I took a look at the QIP source and it doesn't address the issue that Quakeworld/DarkPlaces/MH post a tutorial on.

I admit I don't know the repercussions for DOS. But that the fix dated back to Quakeworld, seems like a lock that the MH fix is backwards compatible to Windows 95 and maybe even Windows NT (I don't claim to really know what Windows NT was, I'd guess some sort of in-between of Windows 3.1 and Windows 2000 or whatever that one before XP was called).

Both Quake 2 and Half-Life had the Quakeworld fix too, if that info helps.

QIP fix won't handle this. Pentium dual core needs MH fix. icore is unaffected. My conclusion is based on a testing pool of (2) computers. And one of those is hearsay.

How about AMD? I didn't really experience any fps rollercoaster in any part of Engoo development, and the only Pentiums i test on are all single cores.

Well, got an AMD Phenom triple core laptop here that never had a problem. But one user with a P6100 laptop experienced gradual time acceleration while playing, until the fix.

I've seen two completely different results with stock code on two i7s from different manufacturers; one was fine and the other was crazy fast. It's obviously a BIOS bug rather than a CPU/OS bug, so in theory you could hit it on any machine, even if another with the very same CPU/OS works fine.