Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Discussions Rules and Guidelines
Report this post
You're right. At first I thought the cities must be a cpu limitation, but after more testing, I'm pretty sure its actually more gpu taxing.
BUT cpu is still relevant. For example, if I assign only pcores to the game, I get stuttering. That is very odd for games. Usually they don't care about ecores at all. But apparently this one does, meaning MT performance is important, possibly even more than single core performance? As my pcores are downclocked from 5.5/5.6 to 5.1 atm and I see no difference in performance. But take away the 12 ecores and its night an day difference. Very, very unusual.
Yes that very thing has been bothering me a lot. It works great for me. But the reports I'm reading are all over the map. People with 3060 tis saying it runs great, people with 3090s saying it runs like crap. Gah. I just don't get it.
Could be like star ocean divine force where you had to rebuild shaders a few times before it got it right.
Ask yourself this,
How well do you know about PC hardware?
Do you know how to optimize your own PC?
While there's nothing wrong with expecting optimization from a game, but I personally won't solely rely on that because even with good optimization, it still won't run great if my PC is a mess.
Learn about PC optimization, otherwise this won't be the first or last game you have trouble running, the next one and the next one and the next one will be the same.
In mine I'm not GPU bound even at the 4K situation, but CPU bound. However, I'm running a RTX 4090 which most are not. You're probably running a weaker GPU so you would hit 99% usage in both situations so before you reduced GPU demand your CPU wouldn't go further due to bottleneck by GPU but as you reduced GPU load (still pushing to 99%) you found your CPU had room to spare.
Yes, there can be a few factors I can name off and probably more I have not realized.
There is the odd interaction with E-cores Iameatingjam mentioned as one as ecores are always
DirectStorage complicates things greatly and the impact can vary based on GPU / GPU load context, elements of optimization regarding the new technology DirectStorage using GPU decompression in only this and one other game (Ratchet and Clank and it didn't even work right there) unknown to us here, what kind of drive is being used, PCIe constraints, RAM performance, and more all sorts of complexity involved.
Issue with drivers and windows updates in some cases confirmed. DDU is often the best way to fix this and many have cited success doing a proper clear with DDU and reinstalling drivers or trying other drivers in general (DDU should be used in both cases if you can't find a fix tho, but normally DDU isn't necessary most of the time).
Issues with shader compilation not compiling correctly and clearing that can fix the issue.
Verify integrity of files in case of an issue there.
There were known to be some upscaling bugs, notably with DLSS but not for everyone... that got some patches in this game.
Some minor optimizations for certain hardware configurations that Square Enix patched for the game (details unknown).
Bug fixes patched so far (details unknown for some).
Resizable BAR is something many try to use because of its supposed performance gains but it almost always actually hurts performance and causes instability / stuttering... frankly, a total failure of a technology in its current state.
Sometimes GPU hardware scheduling can cause issues (unfortunately, in order to use frame generation it must be enabled).
Some refuse to use any upscaling and only native.
Some pair weak CPUs with notably beefier GPUs that outclass their specs, which is frequently fine but once you come across a more all around or CPU intensive game it falters hard.
Sometimes a simple reboot fixes things.
Chroma app in Razer was causing issues for some, but not all of us that have it. Not sure if they fixed it yet.
Some people don't really understand the quality settings and what to properly set and are exceeding their specs trying to run max settings not understanding this isn't worth it nor should people expect to run max settings in some newer games because they're often essentially future proof / best of class settings at the top end, see my prior post about the slider example comparison for more details and to get a better idea of what settings you can drop while maintaining image quality.
If so this is a common issue in games where there are sometimes a single-threaded bottleneck. This game does use other threads quite a bit (up to a point where it can no longer threaded efficiently or literally at all to additional threads). Some games are really very bad at this, criminally so, like Elden Ring where they're blatantly single-thread bottlenecked (a primary worker thread is utilized far more and basically permanently capped preventing full GPU utilization if GPU powerful enough compared to CPU while other threads are far lower utilization generally) while others less so because the disparity isn't so overwhelming (there may be some bias where one thread is slightly heavier load but the difference isn't so staggering, still technically single-thread limited depending on the scenario but wont usually be cited as such because close enough that it isn't super blatant and general workloads slightly fluctuate across the threads so not to mistake a brief spike in activity as a bottleneck, and most importantly because core 0 tends to have system processes in the background most commonly assigned to it resulting in a slightly higher inherent utilization).
If it is something else tho... regarding Core 0 weirdness, I'd be curious what you mean.
EDIT: Is it related to stutter/hitching? I found a reddit thread, trying to figure out what you meant, that took issue with core 0 and they found core parking was causing issues. https://www.reddit.com/r/FFXVI/comments/1fsy6zd/psa_disabling_core_0_for_ff16_pc_eliminated_my/
Well I wont deny that since there seems to be so little consistency and I can only report what I see on my end. I also have a 4090, and a 14700k. I run at 4k60 with dlss on balanced ( quality works fine too but since I'm also using dldsr, I can't really tell the difference so I'll take the lower power draw) neither component is 100%ing. Even when it comes to utilization, its not so simple. There are different ways to measure it, and it doesn't always mean what we think it means.
But I have been artificially limiting my components to try and figure out where the bottleneck is. The game really doesn't seem to mind it when I downclock the cpu, even if its by a lot, it does seem to mind when I downclock the gpu or start taking away more than a few cpu cores. I should try running it off a sata ssd and see if that makes much of a difference.
Anyway, as you said pretty well in the second part of your post, it aint easy, since there are so many variables in computers. Nevertheless, usually game performance doesn't care about that much more than GPU and CPU plus or minus a few %, and also to a lesser extent RAM and SSDs, the latter mostly for loading times but can sometimes affect lows.
But something about this game is affecting a lot of people and not affecting a lot of others. Is it dirctstorage? Maybe it has to do with the pre-compiled shaders. Maybe its driver related, or maybe it doesn't play nice with some kind of common piece of hardware or software that we aren't even thinking about. Maybe its some random unrelated bios or windows setting. Who knows?
Perhaps some reports are a result of ignorance but I surely wouldn't dismiss them all as, well those people are just stupid and don't know how to use PCs... like some people seem to be implying ( not you). Thats a great way to drive people away from PCs, and by extension, reduce the amount of games we get. Not to mention, it doesn't get us any closer to the answer.
Though I do find it annoying when people open threads to complain about performance but don't seem to want to go through the troubleshooting steps.
You should check out SpecialK discord and see about the DirectStorage tweaks for this game. It could have some rather substantial improvements based on what the creator found with preliminary testing. I haven't bothered because my performance is more than good enough for me not to dig into that, but since you're already enjoying doing a lot of tweaking and testing it may be worth a shot.
That's a lot more than 40.
Anyway defend this game's optimization if you want. It ran notoriously bad on PS5 and $1,000 GPUs get 70fps at 1440p on PC, without any RT. Seems bad to me, but I'm glad you're enjoying the experience. No real need for more debate.
Dude, I've repeatedly given you the benefit of the doubt that maybe you just didn't understand tech and programming well enough and were frustrated but after going through lengths to answer your questions and having to habitually correct your claims, often with you giving outright false information and fake source claims to pitch your perspective as you expect no one to fact check it just like you did with this comment I'm quoting immediately above, I can only conclude you're intentionally being maliciously misleading at this point beyond questionable doubt.
Someone else has already been caught making this lie before in this very thread, and I responded to their post and yours at the same time. Remember when they brought up Digital Foundry and claimed they showed evidence that the PS5 version "ran just that bad"? I do... Because I linked the article they wrote showing they said the opposite of it and praised it as one of the most amazing console releases in a long time, only taking minor issue with performance mode and recommending quality mode over it, besides the FSR1 instead of using FSR2 point.
In fact, I'll present that again for context for anyone else to see.
Well dang, that doesn't seem to match up with what you just said about the PS5's version. This is the total absolute opposite of "ran notoriously bad on the PS5" that you just claimed.
As for your claim about the performance using expensive PC parts and no ray tracing I hard debunked that in post #38 (my post) of this thread with concrete data and further explained about this game's scaling, too. You're definitely intentionally trolling at this point. That much is very clear.
Like I said I don't have any issues. The load is distributed amongst all the cores quite evenly and none of them are at 100%, nor is the gpu fully loaded (because of the 60 fps cap). The frametime graph is smooth. I was doing the investigating for the sake of others.
Digital Foundry saying "30fps for a game like this is fine" is one reason of many I don't really care about Digital Foundry's opinions. The performance mode... which 75% of people use according to the recent PS5 Pro announcement... looked and ran like crap.
I pick out certain things to respond to because doing multiquotes is annoying and I try not to spend too much time on forums. It's not an insult. We heavily disagree on this game's technical merits, but it's all good.