One thing we forgot to mention - all testing was done using the retail version of the game, not a preview build. TIMESTAMPS 00:00 Start 00:48 Introduction and test setup 04:17 1080p Ultra benchmarks 05:48 Nvidia has a problem… 07:15 Back to the benchmarks 08:04 1440p Ultra benchmarks 09:27 2160p (4K) Ultra benchmarks 10:12 Preset scaling 11:04 Anti-Aliasing options 12:04 1080p Medium benchmarks 13:10 1440p Medium benchmarks 14:09 Closing thoughts
Just want to say the research and effort put into this video is outstanding - and still very relevant today, really hope CA and Nvidia is aware of this and the instability of performance as it makes huge sense.
Currently clocked just over 60 hours in game. 3900X - Nitro+ 6800 XT - 32gb ram - SSD install. At 4K Ultra I get around 35fps in Campaign and around 45fps in Battles. Only had a single crash in 60 hours. Using full screen mode (alt tab working). I think I'm just very lucky with the crashes. Performance could be better but not unplayable.
Okay. Looking at the video again. It kinda makes sense. I can see how the scaling really affects the performance. Honestly, I don't know why anybody would game at Ultra settings even if you could. You get so much performance back by just dropping to high. And some one like me, I like to optimize the settings. I drop Shadows, Grass Quality and Water to about medium. But I like Textures quality, AO, World Effects at the Max if possible. You get super high FPS and the game still looks great.
Yeah, just going in and enabling Ultra everything isn't necessarily the best move. But there is a ton of scalability on offer from the various settings, I think people just need to adjust their expectations!
Really as long as you get a decent 50 fps you can crank up the settings to max just for the looks, you don't need 200fps in this game, especially since the most demanding part is the TBS campaign map...
Disabling Control flowguard for the game helps with the hitching.(frametime spikes) in the campaign map. Also 3d portraits is another one that causes frametime spikes from what I observed.With it on 2d the frametime spikes drop by like 90%
I guess having a high end 12th gen Intel Mobo with the best 12th Gen CPU, 32 gb of ultra fast memory helps those scores out a lot 🤔. This game as every other TW game is CPU intensive far more than GPU. Hopefully someone does a more reasonable test with hardware that is more commonly used by actual players as opposed to rich YT tech channels backed by game developers/publishers.
@@mental7630 lawl, imagine buying a new comp in 2022 and only getting the sillicon mistake intel made called 11th gen.. RIP you. You should have done your reseach better before jumping on an 11th gen part.
I don t know if only me, but when i turn on any antialiasing, the game looks more blurry.. and if i turn off my antialiansing the game looks more sharp and chrisp...
I have a few of the cards listed between myself and my friends (3060, 1080ti, 6700xt, 2080ti, 3090) and we can’t replicate any of those results - all are consistently 10% under what you’re seeing, at least in 1080p. Newest cpu used is a 12600 on ddr4, but in most cases it’s GPU bound so I don’t think the ddr5 alder lake platform. Apologies, I don’t want to be one of those posters who just complains about your testing, I’m sure it’s fine, but it doesn’t seem to match elsewhere either and I’m trying to understand why. Is there any chance you tested this on the pre-release build? There have been reports of performance regressions between that and the release.
I am not an expert in games performance testings, but I believe RTS games are bound by single thread performance. 12600 is a great processor,but the 12900k is a better performer. I watched the HUB video on Age of Empire 4 and the 12900k was ahead.
@@zalankhan5743 it’s not single core dependent, especially in battle. I kept a close eye on gpu usage and it never dropped below 99% and we had a couple of other processors (5600x,5800x,10400) in the mix that seemed to make no difference Wh3 seems lighter on cpu in general - as a test I dug a 4690k - a 4 core i5 - and it didn’t significantly bottleneck a 3060
@@MarkTownshend yes Warhammer only bottlenecks on the campaign when pressing end turn where it only uses a single core for the ai turns, otherwise it's all the GPU.
Ive got a 7800x3d and 4090 setup. Sometimes I look at the FPS while im on the campaign map and its at like 70. I get more than that in big Cyberpunk battles with full path tracing on.
why is there no Nvidia rtx 2070? İm having significantly low fps compared to other systems. İ have a water cooled İntel İ7 8700K, 16 gb ram and a good airflow pc casing but it seems like game hits 72-75 degrees in the campaign map on high settings around 70 fps but gives more fps in battles with 68-70 degrees heat. İ have the latest drivers installed too.
Hello Disc SSD M.2 Adata XPG SX8200 Pro 512GB PCIE3.0 x4 Motherboard MSI B450-A Pro Max Ram G.Skill 16GB Ripjaws V DDR4 3200MHz CL16 DUAL KIT Procesor AMD Ryzen 7 2700X 8x4,3 GHz AM4 Obudowa SilentiumPC Signum SG1X TG RGB Zasilacz CoolerMaster SilentPro 1200W Gold Msi Rtx 3060 ti gaming trio I want to play on Ultra 2k Total War Warhammer 3. How many fps could I expect? And please tell me that this set is good, no bottleneck etc. Thank you!
Great stuff and a lot of work there but can you please change the title to 'GPU Performance Benchmark'? Because I hope that in part 2 you will cover RAM scaling (what are the benefits for going from 8 GB to 16 GB to 32 GB, maybe to 64 GB?), RAM speed scaling, CPU scaling (4 cores? 6 cores? 8 cores? 16 cores? How about 64 cores?), and more. Or is this a game that can run on a Celeron with 4 GB of DDR4-2400 and is entirely dependent upon the GPU?
AFAIK the game is really CPU demanding as well, especially in battles, in the Skaven TWW2 benchmark at 1440p I was CPU bound on a 3900X... so you won't get any good performance with less than 6 cores and 16GB RAM RAM speed isn't that much of an issue I think
I’m on the cusp of buying myself a Lenovo Legion. That’s with either Ryzen 7 5800 or an Intel i7 -11800H and RTX 3070. I’m leaning towards the I seven just because I know total war games are more CPU heavy. Anybody got any recommendations or advice? Before I pull the trigger…
You cant compare laptops to desktops, = it will preform a lot worse than most benchmarks you see on the web, esp from the big tech tubers who uses highend desktop parts. laptops are heat and energy deficrated, and cos of this a laptop RTX3070 is only around a desktop 3060/3060TI in pref, same with the cpus, undervolted so hard that they dont preform more than 60-85% of their desktop counterparts.
I ran it with a i59300h quad core 1650 4gig and 16gig ram laptop, ran alright had 1080p medium settings with unit size set to ultra ran at between 20-30 fps wasn't too good but definitely playable
Playing at Ultra on a 3080 (with 5900X and 32GB RAM), 3440x1440, it's gorgeous, 60ish fps battle benchmark lowest 45fps, no NVidia "frame drops" (using Afterburner graphs to track frametime, no big spikes) You also made the (excusable) "error" to include unit scale as pure "graphics settings", because it's in the same menu, but which it is NOT because number of figures in units also change drastically the gameplay itself... For example 1 mage vs 120 troops ain't the same as the same 1 mage vs 40 troops, as he'll be using the same spells that do the same damage whatever the scale ! This makes the game awfully unbalanced at lowest scales, as it's meant to be played on High+ unit size In order to do a "graphics only" benchmark you have to set the unit scale across all tests...
Time to debunk this myth that the game is CPU intensive. I play at ultra settings 1440p and my cpu is like 20% usage. this game is GPU intensive. my 3060 is heating up like a mad. this game has zero optimization.
Man, I don't care how "Demanding" a game supposedly is. If even a 1080ti can't get a locked 60 at 1080p. I call that Unoptimized Ubisoft Garbage. But then you can't even call it that. That's disrespecting Ubisoft a little too much. Lol.
This game is made by Creative Assembly and published by Sega. It has nothing to do with Ubisoft. And it makes me question KitGuru's competence by giving it a heart.
People still complains about the bad performance? I see none if you can still push frame by upgrading the GPU later on in the future... This game is well optimize for all it offer, way better than some games like City Skyline or even Crysis back in the day. Also 4090 might sound expensive but is what I consider the first 4K GPU of it class and people still complain about it. Back in the day GPU costing 600U$$ and those where 1080p class GPUs. I imagine more crying when they release the first 8K capable GPU and not matching a 4K GPU on price. My only wish is that they create new engines to push this RTS games into the multi-core performance, that gonna be something huge when it happen.
This Dogshit game Engine Doesn't support dlss or fsr . It needs an engine rework. I think its laziness and makers of one of the best games in all of history be incompetent..
CA refuses to create an engine that can actually run this game. I think it's the same recycled engine from Rome 2 (2013) and all the warhammer fans will just keep paying for outdated tech so they can live their warhammer battle fantasy. They just keep pumping out releases, yet the bad optimization remains exactly the same, and the fans just buy it anyway
One thing we forgot to mention - all testing was done using the retail version of the game, not a preview build.
TIMESTAMPS
00:00 Start
00:48 Introduction and test setup
04:17 1080p Ultra benchmarks
05:48 Nvidia has a problem…
07:15 Back to the benchmarks
08:04 1440p Ultra benchmarks
09:27 2160p (4K) Ultra benchmarks
10:12 Preset scaling
11:04 Anti-Aliasing options
12:04 1080p Medium benchmarks
13:10 1440p Medium benchmarks
14:09 Closing thoughts
The campaign map performance is abysmal, it needs to be looked at. Combat performance is "ok". Still about 30% worse than Warhammer 2 though
Just want to say the research and effort put into this video is outstanding - and still very relevant today, really hope CA and Nvidia is aware of this and the instability of performance as it makes huge sense.
Thank you for including the older cards. I'm sure I'm not the only one who is basically stuck with an old GPU due to the current market.
Currently clocked just over 60 hours in game. 3900X - Nitro+ 6800 XT - 32gb ram - SSD install. At 4K Ultra I get around 35fps in Campaign and around 45fps in Battles. Only had a single crash in 60 hours. Using full screen mode (alt tab working). I think I'm just very lucky with the crashes. Performance could be better but not unplayable.
Okay. Looking at the video again. It kinda makes sense. I can see how the scaling really affects the performance. Honestly, I don't know why anybody would game at Ultra settings even if you could. You get so much performance back by just dropping to high. And some one like me, I like to optimize the settings. I drop Shadows, Grass Quality and Water to about medium. But I like Textures quality, AO, World Effects at the Max if possible. You get super high FPS and the game still looks great.
Yeah, just going in and enabling Ultra everything isn't necessarily the best move. But there is a ton of scalability on offer from the various settings, I think people just need to adjust their expectations!
Really as long as you get a decent 50 fps you can crank up the settings to max just for the looks, you don't need 200fps in this game, especially since the most demanding part is the TBS campaign map...
i love the looks so much, crank up everything to max but play at 1080p coz thats my monitor native res
Disabling Control flowguard for the game helps with the hitching.(frametime spikes) in the campaign map.
Also 3d portraits is another one that causes frametime spikes from what I observed.With it on 2d the frametime spikes drop by like 90%
I guess having a high end 12th gen Intel Mobo with the best 12th Gen CPU, 32 gb of ultra fast memory helps those scores out a lot 🤔.
This game as every other TW game is CPU intensive far more than GPU.
Hopefully someone does a more reasonable test with hardware that is more commonly used by actual players as opposed to rich YT tech channels backed by game developers/publishers.
I just bought a new computer ran a benchmark for it and got 50 fps on ultra settings.
It's a
i7 11700 f
Rtx 3060 ventus 12g
Ddr4 16gig ram
@@mental7630 lawl, imagine buying a new comp in 2022 and only getting the sillicon mistake intel made called 11th gen.. RIP you. You should have done your reseach better before jumping on an 11th gen part.
@@AdaaDK works great thanks man :)
Dont forget to check out campaign map, that's where the horrible frames are at. I'm getting a lot higher frames in battles.
Same here, this is still the case
Allow me to add, shadow settings do affect performance a lot
I don t know if only me, but when i turn on any antialiasing, the game looks more blurry.. and if i turn off my antialiansing the game looks more sharp and chrisp...
Try fullscreen.
I have a few of the cards listed between myself and my friends (3060, 1080ti, 6700xt, 2080ti, 3090) and we can’t replicate any of those results - all are consistently 10% under what you’re seeing, at least in 1080p.
Newest cpu used is a 12600 on ddr4, but in most cases it’s GPU bound so I don’t think the ddr5 alder lake platform.
Apologies, I don’t want to be one of those posters who just complains about your testing, I’m sure it’s fine, but it doesn’t seem to match elsewhere either and I’m trying to understand why. Is there any chance you tested this on the pre-release build? There have been reports of performance regressions between that and the release.
I am not an expert in games performance testings, but I believe RTS games are bound by single thread performance. 12600 is a great processor,but the 12900k is a better performer.
I watched the HUB video on Age of Empire 4 and the 12900k was ahead.
@@zalankhan5743 it’s not single core dependent, especially in battle. I kept a close eye on gpu usage and it never dropped below 99% and we had a couple of other processors (5600x,5800x,10400) in the mix that seemed to make no difference
Wh3 seems lighter on cpu in general - as a test I dug a 4690k - a 4 core i5 - and it didn’t significantly bottleneck a 3060
@@MarkTownshend yes Warhammer only bottlenecks on the campaign when pressing end turn where it only uses a single core for the ai turns, otherwise it's all the GPU.
I got the exact same in my benchmark that he did in his. 113 to the dot
@@andyw.1164 interesting- what’s your cpu and memory?
Worth noting that all the recent CPU's and game optimization has made these numbers a little outdated
Would rather see High preset than Ultra preset - Ultra is always stupidly heavy in Total War.
Ive got a 7800x3d and 4090 setup. Sometimes I look at the FPS while im on the campaign map and its at like 70. I get more than that in big Cyberpunk battles with full path tracing on.
Demanding? no its not demanding its just Shit optimized. i get near the same fps on ultra as i do on lowest. with a 1080TI
What the hell is wrong with this game? Did they make it for rtx 9090ti?
Great video man this kinda helps.
Great idea, and really nice execution and presentation. Thanks
The launch day inclusion of Denuvo may be causing issues.
No cpu review?
Does the game support SLI/Crossfire? Two 3090s would solve it all 😏
epic video, much thanks!
why is there no Nvidia rtx 2070?
İm having significantly low fps compared to other systems. İ have a water cooled İntel İ7 8700K, 16 gb ram and a good airflow pc casing but it seems like game hits 72-75 degrees in the campaign map on high settings around 70 fps but gives more fps in battles with 68-70 degrees heat. İ have the latest drivers installed too.
love your videos
I'm so happy it looks like I can hit 60fps on 1080 medium when I get this game with extra anti aliasing. Most demanding game I'm gonna own.
Why are 5700xt and 6600xt so slow in this game..
Great video, thank you!
I somehow get 30 to 40 fps 4k ultra with my rtx 3080 ti. Am i doing something wrong?? It plays every game pretty good except total war warhammer 3
Why is there no gtx 1080 ?
Thank you!
Wow, really professionally done. Keep pushing forward, I see a bright future for you :)
Hello
Disc SSD M.2 Adata XPG SX8200 Pro 512GB PCIE3.0 x4
Motherboard MSI B450-A Pro Max
Ram G.Skill 16GB Ripjaws V DDR4 3200MHz CL16 DUAL KIT
Procesor AMD Ryzen 7 2700X 8x4,3 GHz AM4
Obudowa SilentiumPC Signum SG1X TG RGB
Zasilacz CoolerMaster SilentPro 1200W Gold
Msi Rtx 3060 ti gaming trio
I want to play on Ultra 2k Total War Warhammer 3. How many fps could I expect? And please tell me that this set is good, no bottleneck etc. Thank you!
Great stuff and a lot of work there but can you please change the title to 'GPU Performance Benchmark'? Because I hope that in part 2 you will cover RAM scaling (what are the benefits for going from 8 GB to 16 GB to 32 GB, maybe to 64 GB?), RAM speed scaling, CPU scaling (4 cores? 6 cores? 8 cores? 16 cores? How about 64 cores?), and more. Or is this a game that can run on a Celeron with 4 GB of DDR4-2400 and is entirely dependent upon the GPU?
AFAIK the game is really CPU demanding as well, especially in battles, in the Skaven TWW2 benchmark at 1440p I was CPU bound on a 3900X... so you won't get any good performance with less than 6 cores and 16GB RAM
RAM speed isn't that much of an issue I think
@@pascaldifolco4611 Then it would be very useful to have this properly tested.
Does this game come with DLSS or something?
No DLSS, though it would be a welcome addition. No ray tracing either.
Nope but there's render scale slider available, dunno what it looks like tho...
I’m on the cusp of buying myself a Lenovo Legion.
That’s with either Ryzen 7 5800 or an Intel i7 -11800H and RTX 3070.
I’m leaning towards the I seven just because I know total war games are more CPU heavy.
Anybody got any recommendations or advice? Before I pull the trigger…
You cant compare laptops to desktops, = it will preform a lot worse than most benchmarks you see on the web, esp from the big tech tubers who uses highend desktop parts. laptops are heat and energy deficrated, and cos of this a laptop RTX3070 is only around a desktop 3060/3060TI in pref, same with the cpus, undervolted so hard that they dont preform more than 60-85% of their desktop counterparts.
gtx 980ti?
Can I run it with a Ryzen 5 5600H, RTX 3050TI 4gb and 16 GB ram?
I ran it with a i59300h quad core 1650 4gig and 16gig ram laptop, ran alright had 1080p medium settings with unit size set to ultra ran at between 20-30 fps wasn't too good but definitely playable
@@mental7630 so I guess I'll run it fine
@@AD45 yeah pretty easily, I would say :)
The tradition of terrible performance continues. It's the same story on repeat every time Total War releases.
Thanks for the tests. Finally got a 3080 for a reasonable price. So looking at the upgrade from 2070super to 3080 will be fun.
Enjoy! Let us know what sort of performance boost you get, should be very tidy
@@KitGuruTech plus Warhammer is my go to game. ^^ The perfect test for me id say ^^
I got the 3090 and I’m getting and average of 145 fps 1080pm everything maxed out. Really smooth and pretty!
Where is rtx 3070ti bro ?
Getting average 45 fps with RX 590 with low setting (no AA, no shadow) but with ultra unit numbers. Ryzen 2600.
Too bad, Troy is so good optimized. The best optimized TW title I played.
Playing at Ultra on a 3080 (with 5900X and 32GB RAM), 3440x1440, it's gorgeous, 60ish fps battle benchmark lowest 45fps, no NVidia "frame drops" (using Afterburner graphs to track frametime, no big spikes)
You also made the (excusable) "error" to include unit scale as pure "graphics settings", because it's in the same menu, but which it is NOT because number of figures in units also change drastically the gameplay itself...
For example 1 mage vs 120 troops ain't the same as the same 1 mage vs 40 troops, as he'll be using the same spells that do the same damage whatever the scale ! This makes the game awfully unbalanced at lowest scales, as it's meant to be played on High+ unit size
In order to do a "graphics only" benchmark you have to set the unit scale across all tests...
don't play it at ultra then
5600x 16gb 3600 ram 6800 non xt 1tb nvme @ 1140p avg 107 fps. Well beyond acceptable
think all total war games are cpu intensive not gpu
Time to debunk this myth that the game is CPU intensive. I play at ultra settings 1440p and my cpu is like 20% usage. this game is GPU intensive. my 3060 is heating up like a mad. this game has zero optimization.
Man, I don't care how "Demanding" a game supposedly is. If even a 1080ti can't get a locked 60 at 1080p. I call that Unoptimized Ubisoft Garbage. But then you can't even call it that. That's disrespecting Ubisoft a little too much. Lol.
This game is made by Creative Assembly and published by Sega. It has nothing to do with Ubisoft.
And it makes me question KitGuru's competence by giving it a heart.
Well you are wearing safety trousers you are ok.
A secondhand 1080Ti for 400i-sh Euros it is then.
👍👍👍👍👍👍👍👍
These numbers are outdated i get higher fps overall now with my 3060ti and ryzen 5 3600 i get on same settings max fps 123 and average fps of 91.9
People still complains about the bad performance? I see none if you can still push frame by upgrading the GPU later on in the future...
This game is well optimize for all it offer, way better than some games like City Skyline or even Crysis back in the day.
Also 4090 might sound expensive but is what I consider the first 4K GPU of it class and people still complain about it.
Back in the day GPU costing 600U$$ and those where 1080p class GPUs.
I imagine more crying when they release the first 8K capable GPU and not matching a 4K GPU on price.
My only wish is that they create new engines to push this RTS games into the multi-core performance, that gonna be something huge when it happen.
Why is 2080 Ti much better than a 3060 Ti? 🤦♂
I'd say poorly optimized instead of just "demanding"
This Dogshit game Engine Doesn't support dlss or fsr . It needs an engine rework. I think its laziness and makers of one of the best games in all of history be incompetent..
Denuvo may be causing the issues.
AMD Ryzen 5 5600 and an EVGA RTX 3090 FTW3, I'm good lol.
Oddly enough I've had no issues
Show-off 🤣🤣🤣
❤
CA refuses to create an engine that can actually run this game. I think it's the same recycled engine from Rome 2 (2013) and all the warhammer fans will just keep paying for outdated tech so they can live their warhammer battle fantasy. They just keep pumping out releases, yet the bad optimization remains exactly the same, and the fans just buy it anyway
Game runs fine. Nearly 60 fps at 4k ultra all settings maxed out. Dropping to 1440p maintains well over 70 fps in campaign.
It's not a great game anyway.