Real World Gameplay CPU Scaling @ [H]

This article is now very much out of date and not valid moving forward due to changes in the PC gaming industry.

What I would like to see is a re-visit with the many co-developed or console port games we have seen released in the last "6 months" that are extreme CPU hogs even at higher resolutions.

Like it or not most major PC game release will be Co-developed or ports, Core 2 duo's even at 3Ghz are now struggling. Users paying attention to this article will now be getting bad info when considering dual cores as viable cheap option in a gaming rig.


You don't need a I7 but you now need a quad.
 
Perhaps you are right that the article is now nearly one year old, but there probably were not really many games released so far that maxes out more than 2 cores. Unless one is running high-end GPU's in crossfire or SLI, an overclocked C2D @ 4GHz would do very nicely IMHO. I still think that it's better having an overclocked dual-core @ 4GHz than a quad-core @ 3GHz, according to the CPU-core usage graphs of the latest games.
 
I'd have to agree with Battleneter2, I think very many of the recently released games will actually show substantial improvement with a quad CPU over a dual CPU. Granted, a 4GHz dual-core should still be fine in most games, and might outperform a quad@3GHz in some of them, but as the article states it's really easy to get most/all quad CPUs to 3.6GHz or even higher.

As such, I'd love to see a dual-core vs quad-core vs hexa-core article following the same format as this one.
 
Raap said:
I'd have to agree with Battleneter2, I think very many of the recently released games will actually show substantial improvement with a quad CPU over a dual CPU. Granted, a 4GHz dual-core should still be fine in most games, and might outperform a quad@3GHz in some of them, but as the article states it's really easy to get most/all quad CPUs to 3.6GHz or even higher.

As such, I'd love to see a dual-core vs quad-core vs hexa-core article following the same format as this one.
Click to expand...

A stock Q6600 outperforming an E8400 @ 3.6ghz in Dragon Age was a big eye opener for me. (Source: PCGH)
 
Battleneter2 said:
This article is now very much out of date and not valid moving forward due to changes in the PC gaming industry.

What I would like to see is a re-visit with the many co-developed or console port games we have seen released in the last "6 months" that are extreme CPU hogs even at higher resolutions.

Like it or not most major PC game release will be Co-developed or ports, Core 2 duo's even at 3Ghz are now struggling. Users paying attention to this article will now be getting bad info when considering dual cores as viable cheap option in a gaming rig.


You don't need a I7 but you now need a quad.
Click to expand...

This article may be marginally out of date but, still relevant. Also, your assessment of future demands games will be placing on CPU's is perhaps wishful thinking. To my knowledge, we haven't seen any games that utilize all 4 cores. In fact, in a pure gaming environment (no apps running in the background), even a dual core is not fully utilized. So, as the article points out, performance gains in a gaming environment by the latest quad processors from AMD or Intel appear to be from the architecture/structure of the units, speed and the cache and not from the use of the extra cores. I hope that will change but I don't see it happening. For now, the GPU rules the roost for gamers. CPU's are old school thinking.
 
ICOM said:
This article may be marginally out of date but, still relevant. Also, your assessment of future demands games will be placing on CPU's is perhaps wishful thinking. To my knowledge, we haven't seen any games that utilize all 4 cores. In fact, in a pure gaming environment (no apps running in the background), even a dual core is not fully utilized. So, as the article points out, performance gains in a gaming environment by the latest quad processors from AMD or Intel appear to be from the architecture/structure of the units, speed and the cache and not from the use of the extra cores. I hope that will change but I don't see it happening. For now, the GPU rules the roost for gamers. CPU's are old school thinking.
Click to expand...
this article did not even use dual cores which are most certainly getting pushed as of now. and just so you know a cpu doesnt have to be pushed to 100% to mean that faster one is needed. there are some games that only use 70-80% of my cpu but will still perform better when I overclock it or if I had a quad.

GTA 4, Anno 1404, Red Faction Guerrilla, Ghostbusters, Dirt 2, RE 5, Bad Company 2, Prototype and a few others will all perform better on a quad. in fact some of those games need a quad just to be perfectly playable the whole time.

on top of that an i5/i7 quad would be needed to fully push a multi gpu setup. you can see here that some modern cpus arent even up to the task of running a 5870 crossfire setup very well and they didnt even use any real cpu intensive games.. http://www.legionhardware.com/artic...ossfire_cpu_scaling_performance_part_2,1.html
 
Last edited:
cannondale06 said:
this article did not even use dual cores which are most certainly getting pushed as of now. and just so you know a cpu doesnt have to be pushed to 100% to mean that faster one is needed. there are some games that only use 70-80% of my cpu but will still perform better when I overclock it or if I had a quad.

GTA 4, Anno 1404, Red Faction Guerrilla, Ghostbusters, Dirt 2, RE 5, Bad Company 2, Prototype and a few others will all perform better on a quad. in fact some of those games need a quad just to be perfectly playable the whole time.

on top of that an i5/i7 quad would be needed to fully push a multi gpu setup. you can see here that some modern cpus arent even up to the task of running a 5870 crossfire setup very well and they didnt even use any real cpu intensive games.. http://www.legionhardware.com/artic...ossfire_cpu_scaling_performance_part_2,1.html
Click to expand...

I selected one of the games you mentioned, Bad Company 2 and did a bit of footwork about it. I ran across this article discussing amongst other things, CPU scaling of an I7 in BC2: http://www.techspot.com/article/255-battlefield-bad-company2-performance/page7.html

While BC2 is one of the minuscule number of modern games will actually utilize all 4 cores of an I7 (or any quad I'd imagine), I was surprised to see how insignificant the scores were between the I7 at 2.2 ghz and an i7 at 2.9 ghz. In fact, it wasn't until the clock speed of this processor reached 3.3 ghz before there was enough difference worth mentioning. The article went on to show that even an I7 920 with 2 cores disabled still showed that there was CPU headroom in a dual core! http://www.techspot.com/mediagallery.php?f=255&sub=images&img=dual.png&full=true

The conclusion I drew from this particular analysis is that you'd better get your clock speeds up. The type of CPU or the number of cores (...well, at least two) is not as relevant as the clock speed at which they operate. So, get the baddest GPU you can afford and don't spend too much time anguishing over the CPU...because evidently, CPU's aren't as significant in a gaming environment as some of us would like to think.
 
Last edited:
ICOM said:
I selected one of the games you mentioned, Bad Company 2 and did a bit of footwork about it. I ran across this article discussing amongst other things, CPU scaling of an I7 in BC2: http://www.techspot.com/article/255-battlefield-bad-company2-performance/page7.html

While BC2 is one of the minuscule number of modern games will actually utilize all 4 cores of an I7 (or any quad I'd imagine), I was surprised to see how insignificant the scores were between the I7 at 2.2 ghz and an i7 at 2.9 ghz. In fact, it wasn't until the clock speed of this processor reached 3.3 ghz before there was enough difference worth mentioning. The article went on to show that even an I7 920 with 2 cores disabled still showed that there was CPU headroom in a dual core! http://www.techspot.com/mediagallery.php?f=255&sub=images&img=dual.png&full=true

The conclusion I drew from this particular analysis is that you'd better get your clock speeds up. The type of CPU or the number of cores (...well, at least two) is not as relevant as the clock speed at which they operate. So, get the baddest GPU you can afford and don't spend too much time anguishing over the CPU...because evidently, CPU's aren't as significant in a gaming environment as some of us would like to think.
Click to expand...
thats hilarious that you looked at the i7 only and think somehow that it relates to another cpu at the same speed. Bad Company 2 is just one of many games that I mentioned that uses all four cores even though you claimed no games even push two cores. BC 2 is NOT even one of the games I was referring to that NEEDS a quad to keep the framerate decent at least not in DX10 or DX9. now in DX11 the cpu is more important than it is in DX10. and the same goes for DX10 as its more demanding on the cpu than DX9. it must be just the way they coded the game but thats how it is in BC 2.

here you can see a Phenom II X4 at the SAME speed gets a massive boost over a Phenom II X2 even at 1920. hell the average goes from 31.8 fps to 48 which is 50% just from going with the quad at 1920 DX9 high settings. there is also a 50% boost in DX10 too so dont tell me a quad doesnt help. not to mention that is just a gtx260 they are using so with a higher end card the gap would be even larger. http://www.legionhardware.com/articles_pages/battlefield_bad_company_2_tuning_guide,3.html

anyway I am not going to go through every game one at a time and what I said earlier still stands.
 
Last edited:
My e6420 @ 3.2 ghz is currently bottlenecking my system.

With everything on high and MSAA from anywhere between 1-4 and res between 1680x1050 to 1400x900, I'm still seeing ~38fps avg.
 
cannondale06 said:
...anyway I am not going to go through every game one at a time and what I said earlier still stands.
Click to expand...

We're all grateful for sparing us the paragraph it would have taken to discuss the trifling number of those games...

Despite the May 9, 2009 date of the article, much of what was originally stated is still true today. I used the I7 because it represents the pinnacle of CPU's (or close enough) and also because I don't have time to review every CPU. Anyhow, I highlighted the insignificant performance difference in the quad CPU's (I7 920 2.2 ghz - 3.0 ghz) because it indicates how important clock frequency is for quad OR dual core CPU's.

Yes, we're finally beginning to see all the cores of a quad CPU being utilized by a smattering of games, it doesn't detract for the fact that CPU clock speed and the GPU remain predominate factors in modern gaming performance. That's a good thing. :cool:
 
Bo_Fox said:
Friggin' awesome article.. bump so I can subscribe! The video on Infernal Engine physics in the Ghostbusters video game flat-out rocks, dude!
Click to expand...

You're tellin me! I witnessed it first hand how badly a cpu can affect your framerate even though you have a pretty good video card.

I was running a dual core athlon 64 X2 system overclocked to about 2.8 ghz. Didn't really matter because they were slow as shit. Played ghostbusters most of the way through and it was fairly smooth, but the framerate dropped down quite often. I finally upgraded and got my i7, installed windows 7 on it. and the game came alive. It was easily triple the framerate from before.
 
What this review doesn't take into account is smoothness, I didn't notice an FPS change going from 4ghz Yorkfield to 4ghz Lynnfield, but it was a hell of a lot smoother.
 
You must log in or register to reply here.
Back
Top