Sorry your browser is not supported!

hello i just bought a new pc and my graphics card is a nvidia 560 i was just wondering if that is a good graphics card or not?

Well it's better than mine, so im guessing it will be fine. What are your other computer specs? You can see mine in my signature to compare, i can get like 50+ FPS on most levels/games. My specs are okay, i wish i could get a better CPU but im kind of low on money...

Hey sic1ne.

If your graphics card is an nVidia Geforce GTX 560 Ti, then you're in great shape. The GTX 560 is the best "bang for your buck" graphics card from nVidia, and PC enthusiasts such as myself are very fond of it. It is in nVidia's latest graphics card series (the 500-series), so you'll have no problem running FPSC, graphically speaking, with your GTX 560. However, keep in mind that things such as AI and threading can cause FPS drops if your CPU isn't up to snuff!

Cheers.

@ exu corporations

@ dillionare


my pc specs are! sorry for the crappy link i thought it would b easier to link you to my specs! sorry guys lol i wasnt shure what specs to post lol so now u can see my whole setup!

copy and paste the link below its best buy link to my pc

http://www.bestbuy.com/site/Asus+-+Essentio+Desktop+/+Intel%26%23174%3B+Core%26%23153%3B+i7+Processor+/+8GB+Memory+/+1TB+Hard+Drive/2634096.p?id=1218341074379&skuId=2634096

You actually have a GT 530 from nVidia, which is considerably slower than the 560 Ti, but should still get the job done. As far as everything else is concerned, you have an excellent computer there, my friend!

@ Dillionaire thanks man hope to see a future project from you!!!

@Dillionaire- You're running GTX580s in triple-SLI, right? How much GPU does FPSC actually use for you? And do you get sizable FPS drops with water?

I ask because I'm running a single GTX570 (i7-2600K CPU) and I've never seen FPSC use more than 10% GPU load. Usually it's using about 6-7% (depends on other apps I have open but I'd say never more than 5% greater than windows and other apps already open). In a demo level Ive been building, I've been finding that water can drop my FPS into the mid-30s if there's several entities being reflected. Without water it's pretty consistently 58-60FPS (this is a built game running at 1920x1080 with 16xAF and 32xCSAA). I also am running with postprocessing mode 1 which doesn't seem to have any detrimental effect with this card. Ive noticed huge gains in FPSC performance going from my old notebook to this disktop rig but I don't know if its the GPU or mostly the CPU upgrade that's given me the boost.

Sorry to hijack your thread, sic1ne! Good looking rig you got there. Hope serves you well.

I'd go with GPU EAI, also I thought SLI / Crossfire caused problems with DX 9 apps?! Which is one of the reasons I never bothered to crossfire mine!

However, on the single HD4670 with full AA and AS I get 48-60fps on average for a full gridded outdoor map with texsize:256 and lightmap:32!

For those who are not convinced their GPU's are capable, I managed (with a bit of simple tweaking) to get (ON THE SAME MAP AS ABOVE) 30fps on an integrated 4500 on a dell 1100 netbook! So even with the crappy cards, a good enough FPS is possible with the right tweaks!

@sic1ne

Thanks, bud. Backatcha, put that GT 530 to good use!

@Errant AI

When I'm running FPSC with my gaming rig, I get a GPU load of about 5% when FPSC is running, of course with nVidia's forced 8x AA and 16x AF. However, when I'm using FPSC or simple windows applications, I disable my other two GTX 580's to save on power consumption. But, being the devilish enthusiast I am, I've tried running FPSC in 2-way SLI, and each GPU load goes down to about 2% or 3%, which indicated that FPSC scales well. Unfortunately, FPSC does not support 3-way SLI, as a test game fails to get past the DarkBasic application launcher. Oh, and I should probably mention that I've overclocked each of my cards.

Anyways, I can explain all that to you! The simple answer is...both! Firstly, let's look at the impact of your CPU. You have the current top-of-the-line Sandy Bridge CPU, the Core i7 2600k. It's a beast of a processor, as it's four cores has about the same performance, if not more, as my hexa-core i7 980x. The CPU handles all things not graphically related, such as AI handling and level building. I bet you've noticed your test levels compiling much faster, right? With your much more powerful CPU, your graphics card is less likely to be bottlenecked. Now, for your GPU... Many processes consume your graphics card's VRAM, or video memory. You have 1.25 GB at your disposal. Most shader programs in DBPro and FPSC, post processing, and all rendering are handled by the GPU. All polygons and textures are held as files in your VRAM. Also, every pixel generated by your GPU and sent to your display is also entirely held in that VRAM. Now, AA and AF edit each pixel refreshed to your monitor, which means that the information normally required to store all those pixels can double, triple, or even quadruple in size, especially with the high levels you and I use. So, with complex shaders such as FPSC's water, there are many pixels that have to undergo a lot of AA and AF postprocessing, which takes up a lot of space in your VRAM. Once your VRAM maxes out, your FPS will sag as a result because the GPU must catch up with itself. That's where an extra powerful CPU comes in. All the pixels are first generated in the CPU as code, then fed to the GPU for post processing. The faster your CPU can do this, the more quickly the pixels can be moved out of the VRAM to make room for other information. So, you can see that the upgrade from an integrated 256 MB GPU to a 1256 MB one and a shiny, new monster processor makes a huge difference!

Oh, and one question...is it possible to get test builds to run at 1920x1080, or do you have to compile them into a full .exe? You mentioned the test level ran at 1080p, so I'm just curious. Anyways, I hope I helped you out!

@Sting

DX9 apps are fine with SLI/Crossfire, as Call of Duty's most recent games are all based off a modified Quake3 Engine, running in DX9. All of the games from COD4 to Black Ops gain substantial FPS increases from SLI. However, Tri-SLi in FPSC does have issues, so you're definitely correct there! I think that DBPro wasnt coded with such advanced configs in mind.

And, you're totally correct, with a little tweaking, most Integrated Graphics Processors, or IGPs, and low end discrete graphics cards can handle FPSC quite well. I overclocked the onboard IGP for a friend's Sandy Bridge CPU and got a constant 45 FPS on a test level!

@all

Sorry for the wall of text! I tend to get very explanatory...

Cheers.

@ Errant AI and everyone who commented thanks for the help guys its much appreciated!

No, you need to make a buildgame for that. However, I've been finding that to be better than doing testgames for certian tasks such as AI script tuning because you can still edit the scripts and the entity info debug display is much easier to read on buildgame and the built game launches faster than a testgame compiles (although, yes, my buildtimes seem at least 10x faster going from old Dual-Core to the 2600K).

Thanks for that rundown. It's helped me to understand better how the VRAM works and the framerate sagging. I'll keep a better eye on my VRAM usage monitor. Sorry to ask another question but there are few of the graphics settings which I don't really understand. In particular there is one for "maximum pre-rendered frames". Right now I have it set to 3. Can I lover the value to help aleviate some of the VRAM usage?



I appreciate your willingness to explain things in a detailed yet easy to understand manner and hope you don't mind another question...

I'm usung a Gigabyte GA-P67A-UD4-B3 which runs PCIE @ 16x/0x or 8x/8x. For future upgrading, would it be better to add a 2nd GTX570 or to go to a dual GPU card like the GTX590? I'm currently using the Asus ENGTX570 DCII so airflow might be a concern if I were to go SLI... I don't think I need the extra GPU horsepower because games don't strain the GPU as-is but from your explanation the extra VRAM would surely be nice.

@Errant AI

Certainly, I'm always willing to help anyone with their questions, especially a forum member as revered as yourself. Never hesitate to ask!

To answer both of your questions, I have to cover just a little background information. So, to begin...whether you're running a single card, SLI, or Tri-SLI, your computer will only have access to the VRAM of your primary graphics card. So, for example, although I have three 1.5 GB GTX 580's, I only have 1.5 GB of usable VRAM instead of the the total 4.5 GB. That said, I can answer your other questions.

The "maximum pre-rendered frames" option is a feature that acts as a trade off between smoother framerates with more VRAM usage and choppier framerates with less VRAM usage. The selected value in your nVidia Control Panel settings is the number of pre-rendered frames that will be held in your graphics card's VRAM. Now, pre-rendered frames are frames held in the video card's framebuffer to simulate "motion blurring" between frames to provide perceived motion smoothness. This means extra frames are held in your VRAM, and are only outputted to the display monitor once the same number of frames have been generated by the GPU and are ready to replace the first couple frames. This creates an optical illusion onscreen that gives the false appearance of faster frame rates. However, all this waiting to output frames results in input lag, and more VRAM usage. The more pre-rendered frames allocated, the more pronounced the input lag, as well as a sag in FPS due to VRAM overconsumption.. So, in contrast, if you reduce the number of pre-rendered frames, your mouse and keyboard inputs will be noticeably more responsive, and your frame buffer will consume less VRAM, but 60 FPS may not appear as smooth as usual. I'd recommend reducing the number of pre-rendered frames from the default 3 to just 2, or if you're really bold, even 1! This should increase VRAM performance about 4% - 7%. You just have to find the right balance your eyes can withstand for the sake of better performance.

Now, about future upgrades! As I said before, adding a second GTX 570 won't increase your usable VRAM, so the only performance increase you would see is that each GPU would take turns rendering one frame. In practice, one GPU would render only 30 frames every second, which makes their jobs a lot easier. I should just let you know that since each card would be running at PCIe x8 electrically, you will lose about 5% performance overall compared to SLI GTX 570s in two x16 slots. Now, your other option you mentioned for this generation of nVidia GPU's would be to buy a GTX 590. However, although it is an enthusiast's dream to own a pair of those bad-boys, there's a little more to their story than meets the eye. The advantage of the card is that it only populated one PCIe x16 slot, but the way it does this is a little sneaky. nVidia manufactured a chip, called the NF200 that takes 16 PCIe lanes and splits them into two x8 lanes. Now, the GTX 590 is actually two underclocked GTX 580's, running in SLI, with the NF200 taking the x16 slot of the motherboard and converting it two 8x lanes. But what they don't tell you is that the both GTX 580s are running at x8 electrically, and since they are in SLI on the GPU's board, only the 1.5 GB of VRAM available to the top GTX 580 is usable. That said, the advertised 3 GB of VRAM is a misleading advertisement. So, I wouldn't recommend a GTX 590 for your future upgrade. This is the reason AMD's competitor for the GTX 580, the HD Radeon 6970, even exists in the GPU market, as it comes with 2 GB of VRAM natively. So, what is your solution without switching to AMD's shoddy drivers, terrible support, and extensive heating issues? I have one for you! EVGA is the best sub vendor for nVidia products I've seen, and it offers a $550 - $600 version of the GTX 580 that comes with 3 GB of VRAM. This solves all your problems, and it is the option I fully recommend you take. In fact, I'll letcha in on a secret. I ordered one yesterday to replace my primary GTX 580, while keeping the other two 1.5 GB ones!

So, good luck with your decision in the future, and let me know if all goes well!

Cheers.

@Dillionaire

Excellent and thanks for that info. I had been blissfully ignorant about how SLI worked and that helped clear up a lot. You did a really good job of explaining things in a thoughtful and easy to understand manner. I won't be upgrading any time soon as I require all my hardware to earn it's keep and this new stuff hasn't paid for itself yet but when I do, I'll know to be looking for a single-GPU card with maximum VRAM (Yes, Nvidia as they're all I've used post-3dfx).

I'll try setting prerendered frames to 1 and 2 and see how I like it. I played with setting to 0 and 8 yesterday but didnt really know what to look for and they both felt wrong. Now I can be a litle more informed about what I'm observing.