Star Citizen 7900 XTX vs RTX 4090 Benchmark | 4K 1440p 1080p
HTML-код
- Опубликовано: 30 июл 2024
- AMD's flagship GPU vs Nvidia's, which GPU is best for star citizen?
►Join the Game Room: www.game-day.xyz/
System Specs:
CPU: 13900KF All Cores at 5.8GHz
RAM: Viper 32GB 4000MHZ CL16 @ 4100MHZ CL16 [Tuned]
GPU: Nvidia RTX 4090
Motherboard: MSI MAG Z690 TOMAHAWK
CPU Cooler: EK Water Blocks EK-AIO 360 Basic
Power supply: Be Quiet straight power 11 1200w
OS M.2 NVMe: Crucial P2 1TB M.2
Game M.2 NVMe: WD_BLACK SN770 2TB
Case: Lian Li O11 Dynamic EVO
Monitor: LG OLED C2 42 Inch 120Hz
Resolution: 4K, 1440p, 1080p
Quality: Very High
Scattered Object Distance: High
Terrain Tessellation Distance: High
Planet Volumetric Clouds: Very High
Field of View: Max
Motion Blur: Full
Chromatic Aberration: Off
Film: no
CHAPTERS
---------------------------------------------------
0:00 Into
0:38 Test setup & Methodology
1:23 Lorville Benchmark 7900 XTX vs RTX 4090
2:52 Star Citizen CPU/RAM bottleneck - No GPU scaling
3:54 Area 18 Benchmark 7900 XTX vs RTX 4090
4:45 Orison Benchmark 7900 XTX vs RTX 4090
5:06 Expectations & Price vs Performance
5:49 Summary
Hope you enjoy the video. Comment with your thoughts about it. Don't forget to hit like, subscribe and share the video if you can.
#13900k #4090 #starcitizen #overclock #rtx4090 #7900xtx 
Игры
I think there needs to be a disclaimer here:
This is all because of clouds being a huge GPU hit on these cards. In reality, clouds at Medium and High in Star Citizen look near functionally identical. Meaning that the XTX is a great value when you consider that the 4090 is basically paying an extra $500 to flip one single setting switch from Medium to Max.
So, while I do have a 4090, I think that the XTX performs pretty admirably as long as you realize that currently all you're missing out on it is slightly less pretty clouds that most people would hardly notice the difference on.
No it's not. The 4090 is the only gpu that has earned it's MSRP. 60% more performance than the 3090 in 4K
@@pepe8196 No it does not.
The 4090 should not cost more than $1,200. All cards this generation are overpriced
I agree but I take it farther.
Every card this generation should have their price cut in half, but the 4090 should be $1000.
I have the XTX, and I leave clouds on Very High. What convinced me was looking out toward Stanton from the Lorville Habs. On High, the clouds were noisy af. On Very High, completely stable. I can't go back to the noise.
And the XTX handles the clouds much better than the 4080, its more direct competitor.
According to a Monthly Report, they're starting on optimizing the volumetric clouds to make them more performant.
@@Sader325 The 4090 is generally ~18% faster than the XTX in 4K. If the XTX was delivering about 82% of the performance of the 4090 for only 50% of the price as in your scenario, AMD would probably gain serious market share. 👀
But I assume AMD is currently more concerned about funding chiplet and stacking R&D and reaching feature parity to neutralize Nvidia's software advantages (in gaming at least).
if you want to use your cpu AIO for at least 2 years without any problems, put that radiator on the top pls. air bubbles are no good for the pump
Will replace with a custom loop soon
Doesn’t matter as long as the pump isn’t the highest point in the loop, so totally fine in this build
@@mortuusvenenum6599 you should take a look again because the pump is the highest point.
Although a 25% boost in Star Citizen is a little more important than most games given how low the framerates are we're talking about. ...I still just like the idea of turning the clouds down to Medium instead. :P
Stunning video. Love the analysis and breakdown of the differences between both. Looking forward to RAM frequency content to come!!! :)
Dude! Your comprehensive data plus the edit quality - you're killing it.
I have no doubt you will have more subscribers in no time!
Clouds are horribly unoptimized in this game and will be eventually replaced. the 4090 is literally the only card that can brute force it right now. But if you turn down the cloud settings you can save alot of money.
Nice comparison. $950 for the XTX now and $1700 for the 4090 as of today. No comparison at all now. 4080 on the other hand... $1100. Tenpoundfortytwo's video no longer relavent though. His was throttling IMO and overheating. He had one of the bugged coolers. I would like to see a new comparison on the current patch. I know 1440P VERY HIGH preset in SC on my 7900XT, I am getting 80 fps - 100 fps running around NB. I am looking forward to playing with it some more. Just reinstalled this morning.
Great video.
You need to post this on Twitter too. Will help you grow your channel faster.
thanks for your work bro!
so glad I went with the RTX4090 after hardware unbox's review for my new build waiting for the 7950x3d
Bra Video. keep up the good work.;) ses. cya.
So you reviewed a AMD reference design that AMD admitted had a problem with the cooler and told all the customers to send them back for a replacement cooler. And you compared the AMD reference card to a NVIDIA board partner card that likely has some improvements over the already impressive 4090 Founders Edition? Has the cooler been replaced on the AMD card? A much better Red Delicious Apple to Granny Smith Apple comparison would have been to not use the AMD reference card and instead use a AMD GPU produced by Sapphire, ASUS, MSI or another board partner. Excluding information on the state of the AMD cooler is a huge oversight and if this card still has one of the defective coolers it would heavily contribute to reduced performance.
I had no issues with the cooling if I did I would have mentioned it, I even reduced the fan curve.
@@gamedaytoday1 Thanks for the reply, but considering how much online drama there was about the whole issue you should have mentioned that up front. Almost no one is going to see your reply to my comment. I do appreciate the comparison though. It would be nice it was a partner card vs. partner card, though I doubt the gpu manufacturers are sending you cards to test, which means you spent money out of your own pocket for this testing.
From other reviews I have seen the AMD partner cards have much higher power limits and don't have the restrictive overclocking limits of the reference cards.
Exactly what I was looking for. Thank you for all this work.
Does the price to performance take in account power consumption/operating costs?
I’ve had problems with my 7900xtx system in SC, keeps on crashing in quantum drive.
Roll back your driver, it crashes at crusader. Not sure how the new driver is rn the one that just came out. Roll back to 05 driver I believe
I agree; it's a stunning piece of engineering! I wanted to be among the first to own a multi-chiplet gpu, i was originally going to get a 5800x3d but frankly i would have used it more for compilations than for video gaming and it was slower than 5800x (or 5900x or 5950x) so hesitated!
I have the 7950x RAM6000mhz and a 13900ks RAM7800mhz with a 4090 all water cooled and I can confirm that I get similar results. Something I want to add that I usually play with three OLED screens running at 12k horizontal resolution and the 4090 is the only GPU that can run Star citizen at that resolution at playable levels. In Orison with clouds maxed I get 30fps and in space I get 60+.
Where to find those GPUs for those prices? Currently 4090 sits close to 2500e, while 7900XTX is approx 1400e, which is basically the same...
Not in europe & the US, last time i checked the AMD card was around 1100€ & Nvidia 1600€
@@DerDrecksack87 well, I am in Europe, and those are the prices for these cards...
You used reference card for AMD and for 4090 you compared AIB model.
maybe if you took the XFX AIB the results would be slightly better for the AMD side.
great video BTW,
im looking for better drivers from AMD.
It's not just the GPU difference in your benchmark chart @3:20. The 13900k benchmark ALSO has tuned ram, while the 7700x benchmarks don't have it, or at least aren't labeled as such. You of all people have made it *very* clear that tuned ram can make a massive difference in SC performance, and I'm genuinely surprised you didn't realize that could be skewing the results.
While I understand where you're coming from regarding GPU under-utilization, when you have MULTIPLE unbalanced factors in a comparison chart, it really does hurt the chart's integrity.
If you want to get a truly fair comparison, you ought to run another set of benchmarks where all the CPUs are running with the same GPU, *and* all of them have well-tuned ram.
Will do just send me all the hardware and I can do the test.
On a serious note, the key take away is that a tuned system will always outperform a untuned system. I just got the 5800X3D and will show the results soon of how it perform stock and tuned. But I am not made of money so I can't get every platform and CPU, but the key take away remains the same: Tune up
@@gamedaytoday1 love the channel Game Day, I really appreciate the work you and ten pound do for star citizen!
@@donhearn2248 Thank you sir
@@gamedaytoday1 Lol, not suggesting you have buy more hardware XD I'm just saying in general, it's best to have only one variable that changes when doing comparison charts, which you usually do for most of your content! I just think that's why that particular chart received so much criticism from the wider public: each build had several unequal variables playing into the numbers, so you can't really draw anything conclusive from that chart on its own without more data.
Either way though, I appreciate the work you put into your content. SC is an especially difficult game to test, and not a lot of people are willing to put in the time and effort to do so. Looking forward to your future videos, especially that 5800X3D one! Will be interesting to see how that extra cache plays with tuned ram
Hello again game day, I just upgraded my pc again, now have a 7800x3d and a 4070 ti, both cpu and gpu utilization is low, with an fps to match in the cities. know any fix to this?
Cities and specially Lorville is demanding, what’s your ram speed?
@@gamedaytoday1 6400, cl32, i can also expand upon the problem, whenever im piloting a ship, my gpu and cpu utalization drops to 30-50% for some reason. Especially when in first person in the pilot seat.
Thank you
Did you have to buy these two gfx cards?.
Yes
Can you revisit this? Considering the updates that 3.23 brought. 🙏
I returned the 7900xtx could not afford to keep both.
i got 7700x on air with 6950xt and 32gb ddr5-6000 and im doing better numbers in lorville. im on 22.5.1 driver. i've heard 7000 series drivers are whack so hopefully with time those framerates wil improve with driver rollouts. but the only way for me to play without crashing every time i jump to crusader, 22.5.1 driver. my screen is 1080p60. i guess ryzen 7 with ddr5 isnt my bottle neck hahah
dude thats a sick ass build bro
why is not 7900xtx vs 4080?
The one thing that is consistently holding me back from buying any of the 40 series cards is the massive energy draw aka heat.
Both of my 3080s are already basically space heaters in my living room and office. Upgrading to a 40 series card just seems like insanity.
I guess it doesn't really matter anyways because there aren't any games that are out that can currently bury my system of such that I can't play max settings at a high frame rate. I was really hoping though that the 50 series would draw a little bit less power. I guess that's not going to happen though.
dude rad on the bottom like that is not a good idea aye.
Yupp. That's gonna destroy the pump within a year.
It’s crazy how the 4090 can be bottlenecked at 4k in some games even on the fastest CPU’s. I had a 5950x 3080ti build then put a 5800X3d then got a 4090 and just ordered a 7950X3d. It was a hard choice between the 13900k and 7950X3d.
Good Review. Those Clouds are in an unoptimised state and cripple performance. To include it and then brute force it with card that cost 55% will always make it look so much better. I Would like to see with full AMD build with tuned ram also with x3d old and new(in future). Also Include the power usage of the CPU in watts so that we can see how much running such a combination over many ours would look ilke, especially if users intend to be on star citizen for long periods. It will be intresting to find out what 3.18 will be like on the PU as it generally performs better, ive been testing for years.
The clouds are unoptimised yet it shows which GPU has more brute-force power if clouds are on a lower setting then the performance will pretty much will all be the same at all resolutions as you saw in Area 18.
Much cope, OP
@@gamedaytoday1 I'm confused what you mean. The difference was 75% with Very High clouds but only 19% in Area 18. What are you saying would be the same?
@@gamedaytoday1 Nobody plays with clouds on high my friend, this isnt a fair comparison really
Dude, your inverted case makes the AIO pump the highest part of the loop which means that over time air bubbles will settle in the pump, Consider switching to Air Coolers instead
Soon custom loop
@@gamedaytoday1 That works too
Careful. Utilization is a tricky metric, friend. It only says how busy the hardware is, not what it's busy doing. It's absolutely possible for software to waste GPU resources on inefficient operations. So to say that there's no more performance to be gained on the XTX literally just based on utilization is way too hasty. You have no real basis for concluding that Vulkan and other optimizations won't yield more performance on the XTX.
Try it on amd platform... it will be still worse, but the margine will be lower. I dont know why but S.A.M. impacts a lot the performance of my gpu :) in some games its 40fps difference which is madness at 3440x1440. AMD should to talk about it same as reviewers its prettty different expyrience kinda often. I am testing it third week and i was shocked how much impactufull it is.
What are the control panel settings for each card?
I would love to see this comparison running a 7800 x 3D
I think the 7900xtx isn't cpu bound rather it's thermal throttling .... with clouds turned on, it will hit Junction temps at 110 degrees Celsius.
No its GPU bound, the temperature never went above 90 degrees. The additional Phanteks t30 helped keeping it cool
Supposing a 20% performance increase from tuned Vulkan and a good FSR implementation, XTX at 4KUltra should hit, what, 90fps 1% lows? 🤔 High enough that FSR 3 artifacts would probably be minimal at 180 fps 1% lows? I'll take it. 🙋🏿♂️
not really surprising the 4090 is better I mean it does cost a lot more, I feel drivers aren't fully optimised for star citizen on the amd side as well
We would need to see AMD get a bigger market share before developers will optimise for AMD properly.
The fact that COD Modern Warfare can beat the 4090 on AMD hardware shows that most of this is probably a software issue.
I see on your channel - all you care about is Star Citizen 😂. Nice performance comparisons anyway 👍
With my 4080, I get 50-90 FPS in Lorville. When I go down the habitat building and stare to the outside from that first platform after the elevators, I get 57-61 ish at 4K (same settings as you have except clouds at medium).
Your cpu?
and if you look at the fps to price chart these new GPU are not that great. given 4k gaming has been a think for over 3 years now and we still don't have these $1300 cards that can reliably push over 100 fps on 4k gaming. its kinda sad
@@Ez0X_ 7900X
@@kiranplays2688 A lot of pixels to push at 4K… 60 fps in SC is actually very playable. It’s the stuttering that kills me…
Who ever playing 1080 in 2024? even my old LG phone got 1440p..
You can't make benchmarks in SC the FPS greatly depends on how fresh the server you're on.
Unless you can make dure you are on the same server every time this is eirthless
Yes you can, welcome to GameDay. The results are based on the average of 5 runs to take into account of server variance look at my other videos I forgot to mention it in this one. And at 4K the GPU is at 100% GPU utilisation so server variance does not even matter as you are GPU bound.
Why are you comparing 7900xtx and 4090 compare the 4080
Send me the 4080 and I will compare it thanks
No AMD driver support for SC.
I have an R9 7900X CPU and 7900 XTX GPU, why is my performance shit in Star Citizen? and it's even installed in an SSD! The most hilarous thing I noticed is the people that defend Star Citizen despite all the criticisms all fair I asked-Terrible optimization, stutters, glitches, and worst of all: restart and delete all progress every big update. For a game that has been poured $110 Million+ how is it still in ALPHA for 10+ Years!?!?!
What settings are you running, whats the resolution? what ram do you have. I just got the 7700x and tuned it and now it performance equal to a stock 7800X3D. So my bet here is that your memory is not tuned.
Get ready to pay $2500 on average for 4090. I'll stick with my 7900 xtx that cost me 1/3 of that instead.
Sold my 7900XTX & got a 4090 instead ...so much better
got to admit made me laugh. i own both these cards and they pretty7 much run the same.
How are you getting the same performance? Seems that your system could be cpu/ram bound?
My man’s looking for bottlenecks via fps… where’s the utilization??
and nivada is 100 percent more expensive
This is an absurdly biased "review". The author has named it "Star Citizen 7900 XTX vs RTX 4090 Benchmark | 4K 1440p 1080p" and then makes conclusions only according to the 4K tests. Recommendations, cost per frame, etc. - all based on the 4K results. Which is hilarious - if I have a 1080p monitor, which card should I take? Which card will be the cost per frame winner? What about a 1440p monitor? Would you recommend RTX 4090 again? The author needs to be more objective and fair when making reviews for the future. Of course RTX 4090 will be better at 4K - it costs $600 more. Let me illustrate to you what $600 are: if you have the same budget for a configuration with RX 7900 XTX and with RTX 4090 and after buying the cards you will have the money to pair the RX 7900 XTX with a Ryzen 9 7950x and 32 GB of DDR5 6000 MT/s with CL30, but the money you will have left for the RTX 4090 configuration will allow you to buy Intel® Pentium® Gold Processor 8505 and 8 GB RAM. Now go compare and draw conclusions! This is a fair comparison and in such a comparison you are allowed to "forget" that one of the cards costs +$600 more. Here you are allowed to feel disappointed of RX 7900 XTX.. but wold you be? Or RX 7900 XTX will destroy RTX 4090 not only at 1080p and 1440p (as per your tests, delivering the same results at 62,5% of the price), but also at 4K. And just to be clear - I know what CPU-limitation and GPU-limitation is. But I also know what an IQ-limitation and lack of objectivity is...
You sound like you haven't been laid in years. I bet not a single person in your life takes you seriously lol
4k is useless XD
I play in 4k... All the rest, lower resolutions are useless. I cant imagine myself playing again in lower resolutions. Imagine playing in 2k...ughh feels like playing on my old gayboy 20 years ago.
@@brakozonhs1 I play SC in 32:9 aspect ratio - cant image it in 16:9.
@@njasicko I had ultrawide monitor for like 2 years. Cant imagine myself going back to ultrawide at lower than 4k resolution or going back to ultrawide at all, since 99% of the content is made for 16:9. I was tired of trying to mod games/apps to support ultrawide.
@@brakozonhs1I use ultrawide for work and 4K OLED 48inch VRR for gaming.
@@zacthegamer6145 I use oled 48inch for both work and gaming, coupled with a vertical 24inch monitor.
this seems way off i get 50 fps on a 6900xt