Video Card Decision (**Update**)

[josh6079]

16xAA with quake4? Come on. It's an old game but it's not that old. It's a FPS you don't need that much AA to begin with.

Yeah, but "need" and "want" are different terms.

My X1950XTX is all I "need" to play it.

That 9800 GTX+ might fit the bill perfectly. It's on par with the 4850 and allows the better AF and soft particles. Sure, it probably won't be able to handle xS modes in some of the games I play, but neither could the GTX 260.

 

[AzN]

Originally posted by: chizow

Yep, and many of them show higher resolutions with AA as well, the point which you continually miss however is that if you have 40 FPS at 1280 with a slower CPU, like a C2D at 2.13GHz and you're not happy with those frame rates, you're certainly not going to see higher FPS as you increase resolution and AA if you're already CPU bottlenecked at 1280. This is CPU scaling/bottlenecking basics though, and something I'm not going to argue with you about.

Who said anything about seeing higher FPS as you increase resolution?

When you raise resolution and become more gpu bound the faster CPU lead shrink then it would have had at lower resolution. Obviously you didn't pay hard enough attention in CPU bottlenecking basics or else you wouldn't be arguing about it with me.

Here are some examples:

http://www.guru3d.com/article/...quad-core-processors/6

AT 1024x768 the fastest CPU QX9770 @ 3.2 ghz is pulling 100fps.

the slowest CPU x2 4850e is pulling 64fps.

34% difference.

At 1600x1200 that 34 difference becomes much smaller 3.5%.

QX9770 58fps
X2 4850e 56fps

Yep, his results are pretty typical of what you'd expect from a title that is mostly GPU limited with occasional dips in FPS in scenes where the CPU is limiting frame rate. Again, Crysis is one of the most repeatable, even predictable gaming benchmarks around so questioning his results are laughable, especially when you're unwilling to post some of your own.

Did you even bother reading his results? So you agree with what he got in that benchmark with him minimum frame rates? At 30% less clocks his minimum frame rates dropping over 50%. yeah really. :roll: Did you know that he later acknowledged there were errors with his minimum frame rates? :laugh:

You come to a thread you have no idea what the fuck is going on when in fact I posted Crysis benches few posts after his benchmarks. Then again all you ever do is run your mouth. :laugh:

As for overclocking, I'm quite sure there's much better value gained from overclocking low and mid-range CPUs than with GPUs so that there's really no need to spend $200 more for a processor. If anything you'd have to put that towards a higher-end GPU as there's clearly more differentiation with regard to GPUs than CPUs in pricing and actual chip/card differences.

yeah right. 10% gains from 33% less CPU clocks. That's what it showed in Toyota and my benchmark. Those settings weren't stressing his GTX 260 limitation BTW. Typical GPU overclock 10% or so. It doesn't take a genius GPU makes more dramatic difference in games than CPU does.

 

[chizow]

Originally posted by: Azn
Who said anything about seeing higher FPS as you increase resolution?

When you raise resolution and become more gpu bound the faster CPU lead shrink then it would have had at lower resolution. Obviously you didn't pay hard enough attention in CPU bottlenecking basics or else you wouldn't be arguing about it with me.

LMAO, again, its obvious you still don't understand the point of using low resolutions when testing CPU bottlenecks. If you understood the point, you wouldn't be complaining about the use of low resolutions in those benchmarks as you will never get higher frame rates as you increase resolutions. So again, if you're not happy with your FPS at 1280, why would you care what FPS you get at 1680 or 1920 when that FPS can only be the same or lower?

As for STALKER, lol, why are you linking a 2 year old game? Its still scales with faster CPUs but there's clearly more recent games that are more CPU and GPU intensive, like any of the games I've linked.

Did you even bother reading his results? So you agree with what he got in that benchmark with him minimum frame rates? At 30% less clocks his minimum frame rates dropping over 50%. yeah really. :roll: Did you know that he later acknowledged there were errors with his minimum frame rates? :laugh:

You come to a thread you have no idea what the fuck is going on when in fact I posted Crysis benches few posts after his benchmarks. Then again all you ever do is run your mouth. :laugh:

Are you claiming differences in CPU performance are always going to scale linearly with FPS? Of course not, in fact you're arguing the opposite without even knowing it. He cut and pasted the results from the built-in benchmark which tend to be accurate and repeatable, so no, I have no reason to doubt his results.

Also he later ran a custom benchmark where he got higher minimums, which doesn't mean his original results are invalid, it just means his benchmark content was different and got different results as expected. Similarly, you don't know if you ran the same bench loop as he did when comparing results, and it doesn't look like either of you confirmed which test you ran either.

yeah right. 10% gains from 33% less CPU clocks. That's what it showed in Toyota and my benchmark. Those settings weren't stressing his GTX 260 limitation BTW. Typical GPU overclock 10% or so. It doesn't take a genius GPU makes more dramatic difference in games than CPU does.

Clearly not in any of the cases I linked where CPU bottlenecks clearly prevent faster parts from performing as they should. Also its clearly much easier to achieve similar CPU performance levels through overclocking than with GPU overclocking. You can spend $230 on a Core i7 920 and achieve the performance of a $1000 i7 965 but you'll never get a $100 9600GT to perform like a $350 GTX 285, no matter how much you overclock it, unless perhaps you're heavily CPU bottlenecked.

 

[josh6079]

There really doesn't need to be anymore discussion concerning the CPU. For the games I play a 2.33 GHz C2D CPU seems fine. If I see the need to overclock it after upgrading my GPU, I'll do so.

As far as GPUs, the 9800 GTX+ or 4850 might be a better fit for the price, as some have mentioned. I'd welcome a GTX 260 over them, but it is more expensive and still may not provide the kind of benefit over those two cards that I was wanting (Supersampling playability).

FEAR 2 is another game I will probably get soon, and it looks as though both a 4850 and a 9800 GTX+ can play it well at my resolution. HardOCP recently did an analysis of it and found the 4850 could play it at 24xCFAA with max settings @ 1680x1050, which is sounds very awesome. Does anyone know if the 9800 GTX+ could score similar results with 16xCSAA?

 

[toyota]

Originally posted by: josh6079
There really doesn't need to be anymore discussion concerning the CPU. For the games I play a 2.33 GHz C2D CPU seems fine. If I see the need to overclock it after upgrading my GPU, I'll do so.

As far as GPUs, the 9800 GTX+ or 4850 might be a better fit for the price, as some have mentioned. I'd welcome a GTX 260 over them, but it is more expensive and still may not provide the kind of benefit over those two cards that I was wanting (Supersampling playability).

FEAR 2 is another game I will probably get soon, and it looks as though both a 4850 and a 9800 GTX+ can play it well at my resolution. HardOCP recently did an analysis of it and found the 4850 could play it at 24xCFAA with max settings @ 1680x1050, which is sounds very awesome. Does anyone know if the 9800 GTX+ could score similar results with 16xCSAA?

thats with an overclocked qx9770 so your results will be lower of course with your much slower cpu. it wont be too much off though.

this is with a faster gtx280 and no AA and you can see a cpu around your level getting about what Hardocp got with a faster cpu and 4850. of course those benchmarks are probably at different parts of the game but does give you an idea. http://www.pcgameshardware.com...tings-compared/?page=3

 

[AzN]

Originally posted by: chizow

LMAO, again, its obvious you still don't understand the point of using low resolutions when testing CPU bottlenecks. If you understood the point, you wouldn't be complaining about the use of low resolutions in those benchmarks as you will never get higher frame rates as you increase resolutions. So again, if you're not happy with your FPS at 1280, why would you care what FPS you get at 1680 or 1920 when that FPS can only be the same or lower?

As for STALKER, lol, why are you linking a 2 year old game? Its still scales with faster CPUs but there's clearly more recent games that are more CPU and GPU intensive, like any of the games I've linked.

ROFL.. You can test all you like only to prove yourself wrong again. :laugh: This happens every time you want to say something stupid which you do every often.

in the real world people play at a setting where it becomes GPU bound not CPU bound. That was the whole point but obviously you don't seem to get it. You just want to push your low resolution benchmarks to show CPU bound.

Are you claiming differences in CPU performance are always going to scale linearly with FPS? Of course not, in fact you're arguing the opposite without even knowing it. He cut and pasted the results from the built-in benchmark which tend to be accurate and repeatable, so no, I have no reason to doubt his results.

Also he later ran a custom benchmark where he got higher minimums, which doesn't mean his original results are invalid, it just means his benchmark content was different and got different results as expected. Similarly, you don't know if you ran the same bench loop as he did when comparing results, and it doesn't look like either of you confirmed which test you ran either.

What you should have done was shut your mouth especially when you don't know what the fuck was posted and were discussed instead of butting in like you know something. :laugh:

I posted benches when you said I didn't. Which is Chizow? Did I post benchmarks or not?

He cut and pasted. Exactly that was the problem. He used some automatic benchmarking program without thoroughly looking at his results and doing more investigations on them. You obviously have trouble with simple math as well or else you wouldn't agree with the results he got.

Clearly not in any of the cases I linked where CPU bottlenecks clearly prevent faster parts from performing as they should. Also its clearly much easier to achieve similar CPU performance levels through overclocking than with GPU overclocking. You can spend $230 on a Core i7 920 and achieve the performance of a $1000 i7 965 but you'll never get a $100 9600GT to perform like a $350 GTX 285, no matter how much you overclock it, unless perhaps you're heavily CPU bottlenecked.

Yeah low resolution with no AA on GTX 280. :laugh:

Seriously why do you bring up such useless things to prove worthless shit? Why not compare at a setting where GTX 280 card will be used majority of it's life time @ 1920x1200 4xAA? At least that's what I hope people use to play on a GTX280 and not 1280x1024 no AA. You are GTX 280 user are you not? Tell me what settings do you play with your GTX 280?

 

[chizow]

Originally posted by: Azn
ROFL.. You can test all you like only to prove yourself wrong again. :laugh: This happens every time you want to say something stupid which you do every often.

in the real world people play at a setting where it becomes GPU bound not CPU bound. That was the whole point but obviously you don't seem to get it. You just want to push your low resolution benchmarks to show CPU bound.

Like I said, its a simple concept I'm not going to argue with you. Not sure why I expected you to understand it to begin with. I've clearly proved my point, I'm sure the interested parties understand as well.

 

[BFG10K]

Originally posted by: Leyawiin

No, I used full quality on almost every game I play and either 4X or 8X AA & AF. Right now I'm pretty immersed in Fallout 3. I use exactly the same settings with the 8800 GTX and GTX 260 (4X AA, 8X AF, all quality sliders and options maxed) and the gameplay is barely smoother with the GTX 260 - to the point of being almost unnoticeable unless I'm watching very closely. I haven't benchmarked it with FRAPS, but they must be withing 5 FPS of each other.

Fair enough; I just noticed you had a Athlon X2 6400+, so what you describe definitely sounds like CPU bottlenecking in this instance.

 

[AzN]

Originally posted by: chizow

Originally posted by: Azn
ROFL.. You can test all you like only to prove yourself wrong again. :laugh: This happens every time you want to say something stupid which you do every often.

in the real world people play at a setting where it becomes GPU bound not CPU bound. That was the whole point but obviously you don't seem to get it. You just want to push your low resolution benchmarks to show CPU bound.

Like I said, its a simple concept I'm not going to argue with you. Not sure why I expected you to understand it to begin with. I've clearly proved my point, I'm sure the interested parties understand as well.

You've seen the results with you own eyes but that's the nature with you. You will argue even though it's been proven. You'll bring up useless benches at 1280x1024 with GTX 280 to get your point across. That wasn't even what i was proposing to begin with. :laugh:

 

[josh6079]

Well, caught a deal and pulled the trigger: This is what I got

Hopefully it will be in working order.

Things that made me decide: price, image quality, cooler, and power.

It was very difficult to ignore the prices the 4850s were going for, especially when several could be found open box but all of the 9800 GTX+s were not. (For days and days the only open box 9 series were at the 9800 GT or lower). When looking at retail 9800 GTX+ prices, I was even more tempted to go for the 4870 512 MB models as they were about $10-$20 more than the 9800 GTX+ version I wanted. (EVGA dual 6-pin connector for $160).

The 4850 would have been lower on the power consumption, but not by too terribly much. While it would have offered Adaptive AA in OpenGL titles it also lacked Soft Particles.

Making the decision even more difficult was Firing Squad's review of F.E.A.R. 2, showing a 4850 beat a GTX 260-216 at my resolution with 8xAA. (A game I intend on playing a lot) Nonetheless, even though there isn't as big of a performance reservoir I considered the 9800 GTX+ to still give playable frames while being capable of better Alpha-texture AA and AF.

Thanks again all for helping me decide. I'll update this thread if I run into any problems once it's here.

 

[josh6079]

Got it.

First impressions:
[*] Noticeably lighter in weight.
[*] Definitely longer in length.
[*] Much more quiet

After getting it running, I ran the Quake Wars HoC benchmark utility. I selected the most difficult map and settings, at least, the most difficult according to my previous X1950XTX benches. Results were:

X1950XTX: 1680x1050 / Salvage map / 4xAA / 16xAF = 28 fps
9800 GTX+: 1680x1050 / Salvage map / 4xAA / 16xAF = 64 fps

It was like night and day.

Playing through the F.E.A.R. demo wasn't as noticeable of a difference, but in the areas where the X1950XTX tanked this one didn't even hiccup. Also, I don't know if I just didn't notice it before or what, but it seemed like in the beginning of the demo - when everything is fiery and the sky is red - there were more red ash particles in the air. I don't know if I remember there being that many or not, just something I noticed right away this time.

The only issue I'm currently having atm is getting BF2 to run. It just won't do it for some reason.

I just downloaded Rivatuner so I can see what the temps are like and stuff.

So far, looks good though. Way good.

 

[josh6079]

Got BF2 to load - had to reinstall it though. Don't know what the issue was.

Now, how do I force the xS modes? I'm using RivaTuner and trying to do it there. AA is set to "Off" in the application and I'm trying to force it to 8xS in RivaTuner, but that hasn't worked.

Also, even when in the normal nVidia control panel if I "Override any application setting" and designate "4x", I still don't get AA once in the application.

Could use some guidance in getting used to how AA is forced, especially the xS modes.