Quake II RTX

Soldato
Joined
6 May 2009
Posts
18,097
Use this for Undervolt stability testing.
Yup, really hammers the GPU. More so than any other game or benchmark
@Rroff do you know if it’s possible to loop the demo demo1 or demo demo2 scene?
A loop with timedemo1 should really stress the GPU if looped for 30 mins
 
Associate
Joined
2 May 2020
Posts
259
Yup, really hammers the GPU. More so than any other game or benchmark
@Rroff do you know if it’s possible to loop the demo demo1 or demo demo2 scene?
A loop with timedemo1 should really stress the GPU if looped for 30 mins
Would also love to know this
 
Man of Honour
Joined
13 Oct 2006
Posts
81,822
EDIT: Stick this in a file called something like loop.cfg in the baseq2 folder then \exec loop in the console:

timedemo 1
alias l1 "demo demo1 ; set nextserver l2"
alias l2 "demo demo1 ; set nextserver l1"
l1

It will stop briefly between demos to reload the file though - so ideally you'd want to record a longer demo and/or not set timedemo.
 
Last edited:
Soldato
Joined
6 May 2009
Posts
18,097
EDIT: Stick this in a file called something like loop.cfg in the baseq2 folder then \exec loop in the console:
It will stop briefly between demos to reload the file though - so ideally you'd want to record a longer demo and/or not set timedemo.

Works great, even with the small stop. Thanks

I suppose another alternative is to use the photo mode (or automate something in photomode if possible) Photomode also really hits the GPU hard
Quake 2 RTX is a much better GPU stress test than Timespy, Heaven or any other benchmark. Ideal for GPU only stress testing
I highly expect someones overclock could be perfectly stable in Timespy for hours but run a Quake 2 RTX loop with high global illumination in high res and it'll fall over after 15 minutes.

I know, run Prime95 small FFTs in the background to kick some CPU heat and TestMem5 [email protected] config for RAM.

Don't think there could be a heavier system stress test than these three combined
 
Soldato
Joined
6 May 2009
Posts
18,097
I tried recreating a similar screenshot I had taken on my 2080Ti. Not exact but some good gains. Benchmarking is around 50% more FPS. Not sure what this works out at, probably about the same


Both cards overclocked to their limit

2080Ti
HTTB4Pn.jpg


3080
8yB5mch.jpg
 
Associate
Joined
23 Aug 2005
Posts
1,273
I've been playing a fair amount of this recently. I played a load of Q2CTF back in the day. Doesn't look like mods like zoid's ctf are supported which is ashame. I have a ton of demos from back then, those can be played in Q2 RTX, just the 'not modded' dm and duel demos play though. It's a great tech demo of path tracing anyway, I'm getting around 110fps on 3070 at 1440. That denoiser is magic.
 
Soldato
Joined
6 May 2009
Posts
18,097
It's a great tech demo of path tracing anyway, I'm getting around 110fps on 3070 at 1440. That denoiser is magic.
Do you have global illumination on high and resolution scaling turned off?

What do you get running timedemo 1, demo demo1?

Never seen 100fps+ even on my 3080 (more like 80fps @ 1440p)
 
Associate
Joined
23 Aug 2005
Posts
1,273
fov 105
115fps global illumination low
102fps global illumination medium
93fps global illumination high

scaling erm EV -1.0 I don't think I've messed with that

Whats the difference? more rays/pixel of light? I prefer the faster frames :)

undervolting 875mv, 1830MHz, +700Mhz mem clock, power limit 109%/90C, GPU temp was 77C but fans were getting loud at 65% coz its a FE card, its on BIOS fan control so the zero fan mode actually works but are louder when gaming.

Quake games feel different at different frame rates, Quake 3 prefered 125fps coz you could jump higher, Q2 is the same, as the framerate dances all over the place the feel of running around changes. But its a tech demo so I guess its moot :)
 
Last edited:
Associate
Joined
23 Aug 2005
Posts
1,273
Hah you're right, I found the scaling option kinda hidden away and a bit cryptic, it was on at 80% ! so all those stats are off by 20% I'm guessing. Now its around 80fps I don't like it lol. I've gone with 50% scaling for the retro 800x600 look ;) and now gone to 200fps :) Also I tried antialiasing with 50% scaling and its horrible. This game needs the DLSS 2.0 magic.
 
Soldato
Joined
6 May 2009
Posts
18,097
Hah you're right, I found the scaling option kinda hidden away and a bit cryptic, it was on at 80% ! so all those stats are off by 20% I'm guessing. Now its around 80fps I don't like it lol. I've gone with 50% scaling for the retro 800x600 look ;) and now gone to 200fps :) Also I tried antialiasing with 50% scaling and its horrible. This game needs the DLSS 2.0 magic.
:D The scaling does work really well when playing but if you stop and look you can tell when it scales to 80% or less.
I just set target FPS at 60 with minimum scale 85%

Go for retro look and feel, 800 x 600 at 20-30fps max!
Gaming in the mid - late 90s was never over 30fps. Not with cutting edge games like Quake II anyway. Only 3dfx managed to break the 30fps barrier. I remember installing a Voodoo 2 and firing up Quake II in 1024 x 768 and experiencing smooth (30fps+) Revolutionary
 
Associate
Joined
23 Aug 2005
Posts
1,273
Yeah with 50% scaling its kinda like software mode before I even owned a graphics card. Even with a Voodoo 2 for a time I set the fps to 35 so I could do the crouch/grapple exit out of railgun room on q2ctf1 map :) later discovering the world of configs so crouch was 35fps and standing was max fps :) Overall though Q2 RTX is amazing, does a great job of showing path tracing, its also a great game to jump into for some GPU stability and temp testing.

I'm a fan of those graphics, they used them to get effect in Devil Daggers and a few retro shooters like Post Void.
 
Man of Honour
Joined
13 Oct 2006
Posts
81,822
I think I'm trying a bit too hard LOL for something you'd barely even notice anyway

9ullRcZ.png
 
Soldato
Joined
1 Apr 2014
Posts
14,750
Location
Aberdeen
I think I'm trying a bit too hard LOL for something you'd barely even notice anyway

Shell casings casting shadows? Collectively the little things like that add so much to the ambience. They help the game look right. There's one web game I play with a pseudo-3d background and the shadows cast by the trees are ever so slightly off and inconsistent. It's just a background, but once you've noticed it, it's not right and it sticks out like a sore thumb.
 
Soldato
Joined
30 Aug 2014
Posts
4,128
Do you have global illumination on high and resolution scaling turned off?

What do you get running timedemo 1, demo demo1?

Never seen 100fps+ even on my 3080 (more like 80fps @ 1440p)
I'm getting around 70 FPS on average on my stock RTX 3080 FE at 1440P.
 
Man of Honour
Joined
13 Oct 2006
Posts
81,822
Shell casings casting shadows? Collectively the little things like that add so much to the ambience. They help the game look right. There's one web game I play with a pseudo-3d background and the shadows cast by the trees are ever so slightly off and inconsistent. It's just a background, but once you've noticed it, it's not right and it sticks out like a sore thumb.

I created the shell casings from scratch - didn't really need that level of detail but I was like why not. The nice thing with ray tracing is even the incidental stuff gets shadows, etc.
 
Man of Honour
Joined
13 Oct 2006
Posts
81,822
Not that anyone would notice while playing but struggling to nail the look I'm going for with the shell cases (even if they do match the ammo case pickup ones in style)

SxH6BnM.jpg


Between my limited artistic ability and the fairly basic physically based material implementation they just aren't looking as good as I'd like.
 
Top