• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

DOOM: The Dark Ages

Flogger23m said:
The path tracing in that video looks like ray tracing in other games. Will try to play it this weekend, so at least the update will be out but doubt a 5070ti will handle it well.
Click to expand...

Yeah they really went light and minimal on the default ray tracing. The path tracing enables way more reflections.
 
polonyc2 said:
I wish publications would not make blanket statements like this...that mainly applies to people playing at 4K...if you're playing at 1440p then I'm sure something like a 4080 Super will be fine
Click to expand...
They did a test with the 4080 super and the drops were pretty bad. Even if you enable frame gen you’re starting off from a base fps of 32 high and 24 low and most likely adding more latency.

Based off the findings in this article unless you’re on a 5080 or 5090 full path tracing isn’t with the huge performance hit.
 
Seems like frame generation could end up making this way more feasible (even at 4K) for other cards. There's also the option of telling it to dynamically target a specific FPS. Doom is such a fast game, you might not even notice the resolution scaling changes. I'll give it a shot on my 4090 when I get home.
 
Game is playable with PT at 1440p on a 5070 ti with DLSS quality and Ray Reconstruction on. Mostly in the 65-75 range, but some areas are in the 55-65 range on chapter 5. Some of the later levels are heavier so there is a chance I would have to drop to DLSS balanced on those or live with some areas playing in the low to mid 50s. Playing native with DLAA is not an option as that gives me 30-40 with occasional dips below 30. Playing the regular RT with DLAA was mostly in the 80-100 range through the game with some areas in the 70s.

PT adds a lot of cost to performance while making the graphics a little bit better. E.g. CP 2077 gives you something like 70% of the PT lighting quality with the regular RT and running PT is noticeably more natural light behavior. In Doom the Dark Ages it feels like the regular RT gives you 90% and you pay a heavy cost for the last 10%. On top of that you are so busy parrying, shooting and slaying demons that you rarely have time to notice a lot of the subtle upgrades. Games like Alan Wake 2 and CP 2077 are moving a lot slower which lets you really soak in the upgrades to the lighting, while I don't have time to do that in Doom and the upgrades of PT are much less in Doom as well.

I will most likely turn of the PT as I would rather have the smoother experience in a game like Doom than slightly better graphics. If you are on a 5090 and running 1440p then it is probably worth it, but not on anything lower IMO.
 
evhvis said:
Game is playable with PT at 1440p on a 5070 ti with DLSS quality and Ray Reconstruction on. Mostly in the 65-75 range, but some areas are in the 55-65 range on chapter 5. Some of the later levels are heavier so there is a chance I would have to drop to DLSS balanced on those or live with some areas playing in the low to mid 50s. Playing native with DLAA is not an option as that gives me 30-40 with occasional dips below 30. Playing the regular RT with DLAA was mostly in the 80-100 range through the game with some areas in the 70s.

PT adds a lot of cost to performance while making the graphics a little bit better. E.g. CP 2077 gives you something like 70% of the PT lighting quality with the regular RT and running PT is noticeably more natural light behavior. In Doom the Dark Ages it feels like the regular RT gives you 90% and you pay a heavy cost for the last 10%. On top of that you are so busy parrying, shooting and slaying demons that you rarely have time to notice a lot of the subtle upgrades. Games like Alan Wake 2 and CP 2077 are moving a lot slower which lets you really soak in the upgrades to the lighting, while I don't have time to do that in Doom and the upgrades of PT are much less in Doom as well.

I will most likely turn of the PT as I would rather have the smoother experience in a game like Doom than slightly better graphics. If you are on a 5090 and running 1440p then it is probably worth it, but not on anything lower IMO.
Click to expand...

I wouldn't play a game at 55-65 frame rates, especially not a Doom game. I think I'll use it without the PT options. I assume the menu is clear about which options are the PT settings?
 
Ray Reconstruction is broken, it's doing the opposite of what it should be doing, too much noise in areas of diffusion.


View: https://youtu.be/rpwRdejkt9Q

With DLSS Performance thanks to DLSS4 Preset K + Frame Gen it's over 100fps which is perfect as there's no annoying latency in the implementation here with a 4090 powering it for a 4K output as you can see.
 
Flogger23m said:
I wouldn't play a game at 55-65 frame rates, especially not a Doom game. I think I'll use it without the PT options. I assume the menu is clear about which options are the PT settings?
Click to expand...
The Doom the Dark Ages has so good frametime that FPS feels like it is 30-50% higher than it really is, however running at at 80+ vs running around 60+ is the difference between smooth and silky smooth. Basically 60 in the Dark ages feels like 80-90 in a lot of games but 80+ feels like 120+ in most games.

PT has it's own setting as most other games. Can't really remember a game where it was difficult to figure out how to turn on or off PT as the descriptions and labels are usually well written due to it being such a resource intensive feature.
 
Honestly, I didn't see much difference in the videos of PT versus non-PT. Maybe I'm blind, but it's not a compelling enough reason for me to cut my fps in half.
 
piratepress said:
Honestly, I didn't see much difference in the videos of PT versus non-PT. Maybe I'm blind, but it's not a compelling enough reason for me to cut my fps in half.
Click to expand...
The video or in person. The video's are so dramatically different I almost believe they intentionally make the RT scenes look worse.
 
criccio said:
The video or in person. The video's are so dramatically different I almost believe they intentionally make the RT scenes look worse.
Click to expand...

IDK how people can't tell the difference. Lighting always has been one of the biggest graphical enhancers since the advent of 3D graphics.

In Dark Ages specifically they were extremely stingy with regular ray tracing use to keep performance high.

In their default ray tracing only extremely shiny things reflect any light. IRL even the most matte surfaces you see have specular reflections.
The lack of specular reflections make objects seem disconnected from the world. Like they were rendered separately and copy pasted in.

With the path tracing option they made nearly everything have them and it looks much better.
 
RDNA 4 seem to handle PT nicely considering the challenge, if FSR 4 can arrive on vullkan
 
criccio said:
Also saying things like "50 series". So a 5060 is ok but not a 4090? Its just a dumb thing to say.
Click to expand...
I believe they're trying to highlight the architectural improvements: "The RTX 5000 series of NVIDIA graphics cards introduces features like DLSS 4 with Multi-Frame Generation and enhanced ray tracing and AI performance through new RT and Tensor Cores. These features are not available on previous generations like the 4000 series."
 
sharknice said:
IDK how people can't tell the difference. Lighting always has been one of the biggest graphical enhancers since the advent of 3D graphics.

In Dark Ages specifically they were extremely stingy with regular ray tracing use to keep performance high.

In their default ray tracing only extremely shiny things reflect any light. IRL even the most matte surfaces you see have specular reflections.
The lack of specular reflections make objects seem disconnected from the world. Like they were rendered separately and copy pasted in.

With the path tracing option they made nearly everything have them and it looks much better.
Click to expand...
It depends a lot on the level, the surfaces and how the lighting is done. Some areas have minor differences, others areas have quite noticeable differences. Most of the time it feels like going from high to very high graphics settings in most games, but at a massive frame rate cost. A lot of the times the non-PT version will look almost as good, just slightly worse lighting so it is not worth the FPS cost for me.
 
I think also there is a difference between having a lot of difference when side by side (which is often the case now, unlike before) and looking better in the artistic/subjective sense
 
Just ran the new built-in benchmark and, as expected, PT tanked my fps, cutting them in half. Also, the light sources only looked slightly better. However, I turned on 3X MFG (4X was slightly over my 16GB VRAM) and it restored my fps to pre-PT numbers. I suppose that makes it a viable option but with so little difference to me, I'm just going to finish the game with PT off.
 
  • Like
Reactions: noko
like this
This is a heavy ass game, makes my 5090 sweat more than path traced Cyberpunk.

Certainly loves CPU too. I was wondering how it scaled on the CPU side so I turned everything way down and saw it pegging 8 cores/16 threads at 100%.

edit: Not really sure what's going on with this game.

It seems to respect x3d cores and you can see all the work locked on them, but eventually throws its hands up and starts burning other threads as well. Can actually see it using 30+ (!!) threads on some of the benchmarks. Like my 9950x3d was at 95~ usage at some points on the forest benchmark. Crazy.
 
Last edited:
Ok so fixed the Ray Reconstruction being broken issue, turns out the game sometimes loads Preset E instead of the Preset K/Transformer model on load and this seems to be related to NVApp messing around with the preset mode in the game profile bit in the app. I use Nvidia Profile Inspector Revamped to manually assign the latest preset/dll injection but NVApp seems to change some settings randomly for some games.... So RR was not using the latest model which is why it was defaulting to Preset E.

I think a total shift over to OBS for game recording and binning NVApp is the way to go until this bug is completely resolved.

Once Transformer was active the same areas that were noisy are now perfect, and because it's K/Transformer, there is more detail too now.
 
piratepress said:

DOOM: The Dark Ages Path Tracing PC Performance Analysis​

https://wccftech.com/doom-the-dark-...ing-demons-never-looked-better-nvidia-dlss-4/
Click to expand...
I feel there is some error here, how RT DLAA latency (4.18ms) so much better than RT native (11.16ms), would have expected a bit worse specially with MFG on, same for PT DLAA vs native, is there some version of DLAA that still do under native res before applying its transformation that I do not understand ?

Maybe they call latency 1000 ms / rendered frame here, instead of meaning latency... or that when DLAA is not on; they do not enable Reflex...
 
Last edited:
LukeTbk said:
I feel there is some error here, how RT DLAA latency (4.18ms) so much better than RT native (11.16ms), would have expected a bit worse specially with MFG on, same for PT DLAA vs native, is there some version of DLAA that still to under native res before applying its transformation that I do not understand ?

Maybe they call latency 1000 ms / rendered frame here, instead of meaning latency... or that when DLAA is not on they do not enable Reflex...
Click to expand...

I think their numbers are just fucked because I can't replicate it.

I'm nowhere near what they're seeing latency wise. I don't see how they could possibly be getting low single digits.
 
LukeTbk said:
I feel there is some error here, how RT DLAA latency (4.18ms) so much better than RT native (11.16ms), would have expected a bit worse specially with MFG on, same for PT DLAA vs native, is there some version of DLAA that still to under native res before applying its transformation that I do not understand ?

Maybe they call latency 1000 ms / rendered frame here, instead of meaning latency... or that when DLAA is not on they do not enable Reflex...
Click to expand...
Divide 1000ms by their framerate and you get their "latency". If they consider latency that matters to gamers is time between each frame shown on a display rather than end-to-end latency then they must mostly be watching benchmarks rather than playing games.

Based on their native numbers the end-to-end latency for MFG with PT would most likely be somewhere in the 50-80ms range vs 40-70 range for native and RT would most likely be 30-60ms for mfg vs 15-40ms for native.

Based on them only using MFG vs Native I believe it is either them brown nosing Nvidia or they got paid to make the article. Either way it is a good reason for me to not use that site.
 
evhvis said:
Divide 1000ms by their framerate and you get their "latency". If they consider latency that matters to gamers is time between each frame shown on a display rather than end-to-end latency then they must mostly be watching benchmarks rather than playing games.

Based on their native numbers the end-to-end latency for MFG with PT would most likely be somewhere in the 50-80ms range vs 40-70 range for native and RT would most likely be 30-60ms for mfg vs 15-40ms for native.

Based on them only using MFG vs Native I believe it is either them brown nosing Nvidia or they got paid to make the article. Either way it is a good reason for me to not use that site.
Click to expand...

This is in the ballpark of what I'm seeing.

DLSS Performance on my 5090 gets the base framerate up enough that latency is generally around 40~ with MFG. Give or take some MS depending on what's happening. It spends a considerable chunk of time in the 35-38 range. Feels fine to me.

Performance looks pretty darn good with the transformer model. At 4k, I feel like I need to legitimately try to spot the difference in actual gameplay and it is well worth the in theory fuzzier image to shave that latency.
 
I finally picked this up now that the path tracing update is out. Seems smooth enough at a bit over 60fps with DLSS balanced on a 5090. I'll have to actually play it some rather than just kick the tires to see if I want to switch to performance or not, or maybe lower some settings. Probably try out quality as well. I'm more or less stuck at 2X framegen thanks to my monitor. It's a 4k OLED that tops out at 138Hz.

I also kicked the tires on my portable rig with an ARC B580 and ran into a bit of a problem. I start it up, the menus are super laggy and the steam overlay says I'm getting 2fps. Go into settings and look, Steam copied the graphics settings from my main rig. So path tracing is on. The B580 can actually execute it. I wouldn't call it running, but at least the built-in benchmark doesn't crash. Got about 15fps with XeSS performance at 1440p. This is going to be annoying switching between rigs. Anyone know how to turn this off short of disabling steam cloud saves?

Other than that the B580 does ok. ~65fps on medium + 4GB texture pool in 1080p native/TAA or 1440p with XeSS quality and feels better than 65fps usually does.
 
I'm a few hours in and... eh.

I dunno, I just want to kill shit without being beholden to some gameplay gimmick.

I don't know if this is gonna click. The beginning at least is honestly kind of boring.
 
socK said:
I'm a few hours in and... eh.

I dunno, I just want to kill shit without being beholden to some gameplay gimmick.

I don't know if this is gonna click. The beginning at least is honestly kind of boring.
Click to expand...

The first couple levels are pretty awful. It gets way better a few levels in, and the shield gimmick ends up being not so bad.
 
PT looks better but not astounding so. At 4K using Balance max with no motion blur, Chromatic Aberration crap, lowest on the benchmarks 69 and 85 for the less intense ones. Playing wise frame rate seems to be around 80 where I am at. Did not like the feel with MFG, just not enough base frame rate with a 120hz monitor. I also think I will not use PT, rather play with VSync running and game at a rather constant 120fps. Feels the best there and still looks great.
 
I played one level with path tracing. I had to bring it down from DLAA to DLSS Q. Hit about 120 fps with drops to 90 which was a bit jarring. Especially when a lot was going on. However, game did look stunning. The subtle softer lighting was very noticeable. I am on my Steam 100% run so taking it slow. Now that patch is out I can complete it for a second time. Won’t rush it this time. Going to be flying a dragon next.
 
Quite liking this, with some minor nitpicks. Performance without path tracing is fine, I am getting 100-120 frame rates with DLSS quality. With path tracing I get around 60 frame rates with DLSS quality. I don't find the game that good looking, even with path tracing. Looks like something from a number of years back. To only get 100-120 frame rates (without path tracing) with DLSS quality is a bit underwhelming given the visual quality, ray tracing or not. Has that odd watery/plastic look the other Doom games had. Though otherwise this is a big improvement over Indiana's engine build, or I just assume the engine handles this type of game better. Or maybe ID is just that much better at optimization, it runs great without stuttering or anything. I turned on frame gen and I don't notice much issues, so I can run it with path tracing on and get smooth enough gameplay.

My only nitpick is it focuses too much on collecting for upgrades. Just makes it a pain to need to go to off areas or look at every corner/hole in the wall to make sure you don't miss gold which is essential for upgrading. Not really sure what the obsession game developers have with junk collecting, it adds nothing to the game. Especially for a fast paced shooter. Aside from that, I find it to be fun. Plays nicely, and am enjoying it. There is a bit much of an emphasis on melee, probably around half your damage output will be from melee. Not really sure this can be counted as a pure FPS game. I still find it fun and while the preferred weapon switching is still present it isn't as bad as Eternal.

Max settings with path tracing:

DOOM The Dark Ages Screenshot 2025.06.21 - 12.13.59.61.png
DOOM The Dark Ages Screenshot 2025.06.21 - 01.38.33.72.png
DOOM The Dark Ages Screenshot 2025.06.21 - 01.44.38.05.png
DOOM The Dark Ages Screenshot 2025.06.21 - 11.10.42.03.png
 
I'm not sure this one is worth bothering with path tracing. It's too slow or needs too much DLSS since it's a fast paced game, and it's just not pretty. Maybe not the fastest paced, but it's still a shooter that's all about combat. Also it doesn't look nice. The graphics are sometimes impressive from a "technical" point of view, but it's just not pretty. Of course it's a Doom game. It's supposed to be ugly. Doing an excellent job rendering ugly doesn't really excite me.

I'll use path tracing in CyberPunk whenever I get around to playing Phantom Liberty and will use it in Indiana Jones and the Great Circle, but I'm not sure about this one. Performance in CP2077 seems to be a bit better, and IJ&tGC is a slow paced game so 60fps would be fine and I get a bit more than that, plus it looks great.
 
You must log in or register to reply here.
Back
Top