Metro 2033 and other PC games which can only be maxed out with hardware from the distant future.
- Thread starter ph0b0s123
- Start date
Recommended Videos
Was trying to keep this as a thread about over the PC games. Metro 2033 is just the latest example.
Now I wanted to keep this on topic of overpowered / under optimized PC games. Which as I said is not so command at the moments since PC graphic envelope pushing has been on hold due to directx 9 consoles. So I probably have not picture the best time to do this as there currently are not many examples.
I understand what you are saying but publishers doing that is risky as sometimes to make things faster for the next GPU gen instead of using raw processing power, some new way of doing the graphics is done. This happened a few year ago when hardware transform lighting came in. Games that supported T&L were faster on later GPU's. Games that were already slow and did not support T&L did not get any faster with new GPU's. This is always the concern when developers put out a games that on max settings is not playable on current hardware. As I said before if the developers have been given early video cards that will be coming out after the game is released to equate max settings to then OK. But assuming that later generations of GPU will be able to play a game you are making is just wrong. Also how can you properly bug max settings in a game when the hardware does not exist to play it properly. Effectively you are running beta / untested settings.Nutcase said:
Now I wanted to keep this on topic of overpowered / under optimized PC games. Which as I said is not so command at the moments since PC graphic envelope pushing has been on hold due to directx 9 consoles. So I probably have not picture the best time to do this as there currently are not many examples.
Now I wanted to keep this on topic of overpowered / under optimized PC games. But since there have been so many responses like the above of, your system is not fast enough, my low powered system can play the game just fine. I thought I would respond to those. First thanks for the responses.Woodsey said:
First is there anyone out there running the game at 1920 x 1200, DX 11 tessellation and DOF enabled, with 'very high' settings and AAA. If so are you getting frame rates that always stay above 35 FPS? To me anything under this is not playable. As posted before
Saying your machine is running the game does not help much as my machine can run the game as long as I turn down the settings or game resolution. My problem has always has been that the games does not run smoothly at the above settings the whole way through the game. On most levels it sits at 60FPS (have VSYNC enabled). But during certain levels the game drops under 35FPS. The firing range at the beginning being one example.
This is what I mean by being disappointed in how the game runs. That on max settings it is still slow on hardware that is arguably 2 generations above what was available when the game was released.
But anyway after turning off DX 11 DOF, and as I progress through the game, scenes that cause low frame rates are becoming rarer. So I am mostly happy.
But please no more I can run that game fine unless you are running it at the same settings I am. Otherwise your comments are pointless.
I think you're looking at this the wrong way, plus it's hardly a new trend in PC gaming. Hell, I remember the original Unreal had extremely advanced settings far beyond what most computers at the time could run meanwhile even non 3D games sometimes boasted these sorts of features with Baldur's Gate 2 supporting some ridiculously high settings for their time and the like.
Besides, who cares if you can't run the game at its absolutely maximum on release? Personally I see it as thinking ahead; Doom 3's ultra setting was designed to run on technology which barely existed at the time of the games release but because the game supported such technology it gave the game some extremely long legs. Play Doom 3 on the Xbox and then play it on a modern PC, and the PC version looks like a sequel it's such an improvement. Gives players a reason to keep returning to older favourites and also leaves the door open for future mods and game support which their console counterparts wouldn't be capable of receiving.
To be completely blunt, I think you're being a bit silly about this. In the end it makes no difference whatsoever and if the game is truly designed with future systems in mind then even the medium settings will likely excel beyond what you could expect from consoles while still offering the game presumably solid gameplay. Only the PC version has a longer-lasting appeal as an added bonus, nevermind mods and the like as well!
And as a side-note 30fps is the point at which the human eye is effectively tricked, much higher than that is simply overkill. That's why many console games are limited to 30 fps.
Besides, who cares if you can't run the game at its absolutely maximum on release? Personally I see it as thinking ahead; Doom 3's ultra setting was designed to run on technology which barely existed at the time of the games release but because the game supported such technology it gave the game some extremely long legs. Play Doom 3 on the Xbox and then play it on a modern PC, and the PC version looks like a sequel it's such an improvement. Gives players a reason to keep returning to older favourites and also leaves the door open for future mods and game support which their console counterparts wouldn't be capable of receiving.
To be completely blunt, I think you're being a bit silly about this. In the end it makes no difference whatsoever and if the game is truly designed with future systems in mind then even the medium settings will likely excel beyond what you could expect from consoles while still offering the game presumably solid gameplay. Only the PC version has a longer-lasting appeal as an added bonus, nevermind mods and the like as well!
And as a side-note 30fps is the point at which the human eye is effectively tricked, much higher than that is simply overkill. That's why many console games are limited to 30 fps.
Video games take 1-5 years to make. The developers always have to design the game to run on hardware that is unavailable and in the future. Occasionally, they overestimate the speed at which the industry will improve graphics and processing and the game requirements are astronomical. The only time I can remember a developer supposedly doing this on purpose was Crysis.
Things are generally better than they were in the 90's. Remember Ultima 9? It only ran on 1 brand of card for the longest time because (apparently) the dev team only had 1 test box. And saying that it ran on that system is a lie. It might have walked.
Things are generally better than they were in the 90's. Remember Ultima 9? It only ran on 1 brand of card for the longest time because (apparently) the dev team only had 1 test box. And saying that it ran on that system is a lie. It might have walked.
I have an I5 750 3.75Ghz on air, 4GB DDR3 at 2Ghz and a 5970. Run 2033 on full settings at 1080 with a minimum FPS of 50-60.
If the OP has an OC'd I7 EE, 2 x 580s and 6GB DDR3 then something is seriously wrong with your setup. Could be bad drivers due to a corrupt windows installation or your PSU is beginning to fail and throttling power to the components because the 12V rail can't handle the load.
For the record, what is the OP's power supply? With an OC'd I7 and two 580's i'm going to guess at least 1050W.
If the OP has an OC'd I7 EE, 2 x 580s and 6GB DDR3 then something is seriously wrong with your setup. Could be bad drivers due to a corrupt windows installation or your PSU is beginning to fail and throttling power to the components because the 12V rail can't handle the load.
For the record, what is the OP's power supply? With an OC'd I7 and two 580's i'm going to guess at least 1050W.
Horizon Based Ambient Occlusion most likely. Tessellation shouldn't be too hard since all DX11 GPU's have tessellation units built into them to specifically deal with it.Cowabungaa said:
Explain why as a result of the motherboard? Failing motherboards tend to intermittently cause BDoD's/failure to boot, not necessarily a slow-down in FPS. I do agree with inefficient cooling as processing units will throttle back to keep the thermal load under control but this can more than likely be remedied by re-applying the heatsink (CPU) properly with after market thermal paste or increasing the airflow to the affected component/a good dusting.starfox444 said:
An actual bottleneck resulting in lower FPS is almost always GPU, CPU & RAM related.
Because if they made the game only to play on hardware generally available and owned by the public, that would be the highest setting. They can make it look better for people who go out of their way to spend more money on processors, RAM, GPU's, etc, but that's just an option if you choose that route. If your computer can't handle the game on the highest settings, then that means the highest settings weren't made for you. If they were, the game would max out at your specific settings, and anyone with better hardware would have to settle for what you have. They test limits of hardware and software so that people who are willing to shell out the extra bucks can enjoy that, but it still works, and it's still the same gameplay if you have to run it at a lower resolution and with anti-aliasing turned town. That's the great thing about PC games to me, they're scalable. If you have a low end machine, most games will run and you can have fun, but you sacrifice some visual quality, if you have a high end machine you can run it at full graphics, and you can do anywhere in between. Fine tuning is enjoyable, and why I still like playing games on my PC versus a console.
I ran the damn thing from start to finish on an old 9800gtx+ in a 2+ year old economy machine (core parts replaced since) at full settings 1600x900 with a second monitor extending the desktop and it was smooth as a baby's behind...though slightly choppy before I forced aero off when it was open. And it was running from STEAM. You are clearly doing it wrong.
Awesome game btw. I juxtaposedly recommend it.
Juxtaposedly is now a word.
Awesome game btw. I juxtaposedly recommend it.
Juxtaposedly is now a word.
I got a single GTX 480 @ 850MHz, 8GB ram @ 2GHz and a i5 750 @ 4,5GHz. Played through Metro a couple of times with High, DX11 and everything on and it never dropped under 40-ish.
Also didn't notice any difference between high and very high other than raped fps.
Since I got older/cheaper hardware and about the same preformance there must be something seriously wrong with some of your stuff.
Also didn't notice any difference between high and very high other than raped fps.
Since I got older/cheaper hardware and about the same preformance there must be something seriously wrong with some of your stuff.
This is actually why I mostly converted to console gaming. Graphics can be better on a PC, but if you don't have the cash to rebuild your box then you just can't play some stuff. On the other hand, if you don't mind the graphical hit, consoles provide a reliable medium.Vern said:
Sometimes a game just has terrible optimization, either with nVidia or AMD, SLI vs. single-card, more than two cores (Fallout 3, I'm looking squarely at you), or just piss-poor programming (Oblivion---still gets choppy even on hardware that is two Moore's Law generations removed from the game's original release.)
ph0b0s123, when it comes down to it, your biggest problem is going to be (strangely enough) the age of the cards. These graphics cards are so new; there really hasn't been time for the driver to mature. Even in single card configuration, there are still plenty of bugs to be worked out by the nVidia driver team to get the best out of your GPU's.
Furthermore, this issue is compounded by your SLi setup, which with really, really new drivers can actually cost more frames then its worth. SLi works in a couple of ways generally, depending on the game. The most common method actually splits the frames between each card, one frame to one, and the second to the other and so on. However if this procedure isn't coded efficiently then you will find that this will cause a bottleneck. The second method is to split the screen in two and have one card render the top half, and the other renders the lower half.
Because the driver is now even more involved in the rendering process (layman's terms) if there are any minor efficiency issues in the driver, you will feel them even more.
Give the driver a month or so to mature, and keep checking the nVidia site, and the SLizone forums for beta drivers and 3rd party drivers. Oh, and it also goes without saying that you should keep checking for chipset drivers too, as they play a big part in SLi frame rates too.
Hope this helps.
Furthermore, this issue is compounded by your SLi setup, which with really, really new drivers can actually cost more frames then its worth. SLi works in a couple of ways generally, depending on the game. The most common method actually splits the frames between each card, one frame to one, and the second to the other and so on. However if this procedure isn't coded efficiently then you will find that this will cause a bottleneck. The second method is to split the screen in two and have one card render the top half, and the other renders the lower half.
Because the driver is now even more involved in the rendering process (layman's terms) if there are any minor efficiency issues in the driver, you will feel them even more.
Give the driver a month or so to mature, and keep checking the nVidia site, and the SLizone forums for beta drivers and 3rd party drivers. Oh, and it also goes without saying that you should keep checking for chipset drivers too, as they play a big part in SLi frame rates too.
Hope this helps.
Erm, right.psivamp said:
The interface between your GPU and CPU is the PCI-E controller. A x16 PCI-E 2.0 slot allows for up to 500GB/s bandwidth. If a motherboard has a x16 2.0 slot conforming to that specification, then that bandwidth will always be the same for a single GPU. If a motherboard has two PCI-E x16 2.0 slots but only 16 lanes of total bandwidth then a dual GPU setup will go x8x8. While you may think SLI/Crossfire setups like this are pointless as the bandwidth per GPU is halved, the real world performance difference between a x8 slot and a x16 slot is about 4%.
Any given motherboard (even mediocre ones) will more than likely have two separate PCI-E controllers, one specifically for the x16 2.0 lanes (integrated onto the CPU itself if you're using the latest nehalem/sandy bridge intel architectures), the other to handle any other x1, x4, & x8 slots for other PCI-E based peripherals. So long as they conform to the PCI-E spec, the bandwidth will be the same between any motherboard. The only PCI-E differences will be that higher end motherboards may have better PCI-E controllers to allow for more lanes of bandwidth (for instance, 24 PCI-E x16 2.0 lanes for an x8x8x8 tri-SLI setup). However when you get to tri-SLI you go into the realms of stupidity.
The bottleneck is either the CPU (not processing the calculations to feed to the GPU quick enough), the GPU (can't render the calculations quick enough) or the RAM over the FSB/HTB (can't transmit data for calculations quick enough to the CPU). The power supply could over-arch all this if it's not giving enough power to the components and so artificially throttling performance.
If the OP's hardware is so high end as he described, it doesn't seem like a bottleneck to me provided his PSU is good enough. It could be a bad installation of drivers or software somewhere.
Yeah, I don't know how much or how little that bandwidth cut matters. I would assume that it wouldn't have mattered much (which is basically what you said, if I understand you properly). I haven't built an SLI machine and probably never will. And, looking back, I think I've only built one tower since PCI-E came out. Fucking technology.Griffolion said: