Unveiling 3.23 - Is It Better Now With Vulkan?
HTML-код
- Опубликовано: 30 июл 2024
- Star Citizen 3.23 CPU and GPU performance review. 1080p and 4K, how does it perform? And which CPU is better 7800X3D or 13900K?
►Join the Game Room: www.game-day.xyz/
► / gameday_youtube
14900K - amzn.to/3xj7ckd
7800X3D - amzn.to/4ae9Bvd
7200Mhz (a-die) - amzn.to/3wVbLRE
Intel System Specs:
CPU: 13900KF All Cores at 5.8GHz [Overclocked]
RAM: TEAMGROUP T-Force Delta 32GB 7200MHz @ 8200mhz [Tuned]
GPU: Nvidia RTX 4090
Motherboard: Asus Apex Z790
Custom Liquid Cooling
@tenpoundfortytwo System
CPU: 7800X3D
GPU: RTX 4080
RAM: 32GB 6000MHZ CL30
CHAPTERS
---------------------------------------------------
0:00 Intro
0:20 Lorville Vulkan vs DX11
2:03 Lorville 3.23 vs 3.22 vs 3.20 vs 3.19 vs 3.18
2:47 Area 18 Vulkan vs DX11
3:01 Area 18 3.23 vs 3.22 vs 3.20 vs 3.19 vs 3.1
3:28 Orison Vulkan vs DX11
3:38 Orison 3.23 vs 3.22 vs 3.20 vs 3.19 vs 3.1
3:57 Orison Clouds test methadology
4:24 Orison 4K Clouds + DLSS performance
5:13 MircoTech Clouds Visual Analysis
6:23 7800X3D vs 13900k
6:56 Why X3D Struggles
7:42 Patch Summary
#13900k #7800x3d #starcitizen 
Игры
Thanks GameDay, will you do Tuned 7800X3D vs tuned 13900K as well? Want to see how it performs in Area 18 and Orison now.
Sure if your comment gets 150 Likes I will do a dedicated Tuned 7800X3D vs Tuned 13900K 3.23 video.
@@gamedaytoday1 can you suggest settings for 4x16gb bdie 6000 cl36. Aslo Would love to see results of an all AMD Machine, as all results here are with nvidia gpu.
@@kaisersolo76 yes the best settings for bdie is to buy a-die.
Yes please
@@gamedaytoday1 if trends continue this way, your video needs to get to 300 likes for that to happen. 🤞 You wouldn't, uh, pin the comment, would you? 😛
DLSS accidentally improving cloud's quality while running faster 😅
Pretty Much!
Remember when Star Citizen was actually playable on a 13900k and 3080 OC V2 12GB at 4K with 45-130 fps.. now it gets like 35 and TONS of stuttering with the same 10gb shader cache and all but 1 e-core turned off for SC with process lasso.
THANK YOU for being the first YTer (I've seen) to explain clearly what the AMDip actually is.
is there any videos of you or settings how you tuned your i9-13900kf for star citizen? kinda would like to copycat those settings cause what ive seen in your videos was more than impressive!
Great video. Did you try the ecores yet? It’s finally working. But I didn’t do any performance test.
It’s now working
another great video how did you get CapX to work with vulcan mine does not seem to want to display but with dx11 it works fine is this a known issue? or am i missing something in the settings ?
Keep the vids coming. also your watercooed rig looks COOL lol
I had to use DX11, CapX on vulkan does not work
On the 7900xtx I gained about 10% on Vulkan in the room you wake up in in loreville. Frametime was smoother, GPU utilization seemed to be a bit better. Was a CPU bottleneck scenario on a 7700x.
Thanks for video. Which is best RAM for Z690 Rog Maximus Hero? And what is the way to manually tune RAM?
These 7200Mhz (a-die) - amzn.to/3wVbLRE
I am writing a DDR5 tuning guide for my discord subscribers, will share it there soon
Your channel is amazing, I've learned so much!
5800x3d 7900xtx 32gigs cl14 3400mhz Vulkan is running alot better for me.
Awesome video bro, yeah I can definitely feel the difference in latency in my run. I think the performance will be better with the upgrade for sure. Do you offer any custom build recommendations for subscribers?? Really trying to upgrade my setup performance this year! 😊
Yes I do, based on your budget and performance objectives.
Excellent video!
Thanks for the video. DLSS makes 4k gaming viable with SC, so at least that's a good news. I see I'll have to wait for next gen of cpus to play Star Citizen, hopefully 4.0 will be out by then.
Yes sir and we need DLSS for Ray Tracing as well, hope they also implement path tracing.
well that's obvious that new cpus will boost performance. however when the multi-threading rendering arrives we will see another boost in fps and render quality. dlss is decent even today.
When you are comparing frame time graphs to Tenpound's results, you are comparing 1 run from your data to 10 runs from Tenpound. It would be easier to visually compare 10 runs to 10 runs or 5 runs to 5 runs.
Interesting, did not know you can combine multiple results into one in CapX. This will save a lot of time. Thanks sir will share data
Do all RGB teamgroup DDR5 RAM sticks use A dye? Building a 7800x3D GIGABYTE X670E AORUS PRO X build..not intending to overclock but maybe someday lol...was considering: T-Force Delta RGB DDR5 Ram 64GB (2x32GB) 6000MHz CL38 Intel XMP 3.0 & AMD Expo Compatible 2 sticks but I heard CL30 was the sweet spot...
Wait for the next gen Ryzen will be announced next week and released in July. Also get the ASUS Gene motherboard instead. I don’t know about those Ram sticks I have not tested them.
Did you notice Vulkan VRAM leaks? Many people are talking about.
Should be fixed now with Patch 3.23.1
ddr5 8200@13600k is not that easy ...
anyways 1080p doesnt really matter for most of us, cause 2k and 4k are more used.
dlss is really good in getting rid od shimmering and softening the distortion. kinda jealous of my 4060 ...
CPU test = 1080p , the game gets CPU bound at 4K DLSS anyways with the 4090 so the 1080p results are valid.
@@gamedaytoday1sure thing, for my case 2k and 7800x3d + GPU iam eyeing for.
1080p is 2K, just sayin' ...
my rtx2060 and ryzen7 2700 usually run at around 50-95% (depends on the situation) but usually usage% are matched.. anyway when i enabled vulkan it really shit on my fps, like gpu 100% running and dlss was mandatory to get close to the fps i was getting without vulkan. also when using dlls my gpu usage drops around 20% though no noticable difference in cities
Tenpoundfortytwo is using a 4080 not a 4090 so the comparison is just wrong.
Not when you are CPU bound
I have a 4080 super and it’s at 90-100% gpu load most of the time with 7800x3d. So unless you are comparing the exact same area in the same server. There’s no way to know that. Thats a super sketchy comparison. You’re also using 8200mhz ram which most people don’t have. You should at least clock that down to a reasonable ram speed that most people buy and offer it as a third comparison.
@@GROGU123 test it for yourself. Set your settings at 1080p put your clouds at High. Any run from the habs in Lorville to the hammerhead near new deal all the way down overlooking the city, same run as tenpound. Do it 5 times to remove server variations then report your average FPS or just share the GPU utilisation.
Or if you actually saw tenpounds video he did test in 1080p and 1440p and got the same frame rate. He always says he is cpu bound with his 4080 when he test at 1080p, So please use common sense here. If you are getting high GPU utilization it’s probably because you are not running with high cloud settings.
Why do I get so many dumb comments all the time?
@@gamedaytoday1 I don’t know why so many youtubers are too dumb to do comparisons. That’s not how things work, i guarantee if you slap a 4090 in his system, you would get more average frames as well. Gives an apples or oranges comparison and pawns it off as legitimate. In the past 3 days alone I had a 20fps variance in Area18 in the same spots moving to belvic convention center depending on when I logged in with zero changes to my settings. Theres too many variables to use data from another person’s system the way you do. Did you go to college and learn anything about testing methodology? Your data is highly tainted.
@@GROGU123 I don't know why its hard to do it yourself if all these youtubers are soo dumb.
Or just show me a benchmark where adding a 4090 or a faster GPU in a CPU bound game increases the average. If you are right, there should be many examples to choose from.
Or explain to me why he got the same FPS in 1080p and 1440p.
All I get is "I had 20FPS different variance in Area18". You just prove that you need to take the average of multiple runs, which I do for all my star citizen benchmarks, I take the average of 5, he does 10.
Sorry for calling your comment dumb, it can't be easy to be in the middle of the IQ curve.
All Video Game Devs should simply ditch DirectX in favor of OpenGL or Vulkan. DirectX always ties users to a specific version of Windows, like an anchor stuck under a rock. The chain needs to be cut so that the ship can sail freely wherever it needs to go.
AMDip at its finest...
there is currently no GPU that is going to make the dated star citizen look as good as the new unreal engine 5 content. They need to remake the visual better, remodel, retexture, and get rid of the FOG AND DIRT EVERYWHERE. FPS can't make star citizen better, correct lighting, practical fog and collision that makes sense will. SC is a hot mess right now. and has been the same for several years. Lets hope server meshing gets fully realised along with 4.0 so they can make a visual pass over the old game content and make SC less of meme then it is right now. Because if a demo on youtube some non game dev can beat out star citizens visuals with ease, I ain't going to put more time and money into it. A fancy ship doesn't make the game, the whole package does. WITH LESS FOG IN SPACE!
while this all sounds great, Vulkan is terribly unstable and nobody cares to render a deadlocked game faster. I've tried Vulkan and it just fails every time.
tldr: no, it still sucks
Pseudoscience and clickbait, nice.