Bottleneck Here, Bottleneck There, Bottleneck Bottleneck Everywhere

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

BFG10K

Lifer
Aug 14, 2000
22,674
2,824
126
There's bottlenecking aright, and it's the GPU in modern games in the vast majority of cases. That?s not to say a faster processor makes no difference, but by far the best upgrade to yield a higher framerate is a new GPU, even with a relatively modest processor.

This is one of the few things nVidia?s marketing page gets right when they talk about building a balanced system.
 

biostud

Lifer
Feb 27, 2003
18,194
4,675
136
to me a bottleneck only exists if the fps dips below 30fps at a given resolution and graphical resolution.
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
I get tired of some of these generalized comments about how its just the gpu thats important and not the cpu because that is bs. there are just too many scenarios where it does matter. sure if you have a 3.00 Core 2 Duo then a faster cpu may not make much difference but if your sporting an older slower cpu then hell yes its going to have a huge impact on performance.

there has to a baseline somewhere and thats what most of the comments fail to address. if someone out there has 3800 X2 and is thinking about getting getting a GTX280 they need to know that they arent going to get close to the performance they see in reviews where the card is coupled with a high end Core 2 cpu. heck a cpu upgrade and a lesser card would deliver a much better gaming experience in that case.

as a former 5000 X2 owner I speak from experience. no matter what card I tried to use in my system the minimum framerate was usually no better than with an overclocked 8600gt. I tried an 9600gt, 8800gt, and 4670 in that system and was never really happy with any of the overall performance. sure my averages and highs were much better especially above 1280 but the cpu just seem to keep the minimums too low.

while using even a mild 4670 in my 5000 X2 I never got even close to the benchmark scores I saw on the web for that card. yet when I used that same 4670 in a system with an E7300 and especially a Q9300 there was a nice increase in minimum framrerates in all games and average framerates in newer ones. I then actually got those benchmark scores review sites were getting. heck with with the Q9300 Far Cry 2 had a massive increase in performance over the 5000 X2. with a stronger card the differences would have been even more dramatic.

I learned a lot about how different cpus can affect gaming. I got so fed up with trying to overcome what my 5000 X2 cpu couldnt do that I went ahead a built a decent computer. theres nothing better than a well balanced pc for gaming and now I finally know what thats like.

heres a nice little chart showing how much the cpu can make a difference even at 1680 and high settings with a GTX260 in Warhead. http://www.hardwarecanucks.com...formance-review-4.html that chart doesnt even cover how much difference in minimum framerates the cpu can make but at least you get an idea of performance difference.

heres another chart showing how much a cpu can bottleneck Far Cry 2 even on Very High settings at 1680. http://www.pcgameshardware.com...CPU_benchmarks/?page=2 thats right even at 1680 on Very High settings an E8400 can deliver 50% better minimum framerates than a 5000 X2.

bottom line is that the cpu does matter but the difference can range from small to huge depending on your current setup. differnent architectures and even cache sizes can play a huge role in your overall gaming experience. the stronger your graphics card the more important it is to back that up with comparable cpu. anyone that doesnt get that is only kidding themselves.
 

chizow

Diamond Member
Jun 26, 2001
9,537
2
0
Originally posted by: BFG10K
There's bottlenecking aright, and it's the GPU in modern games in the vast majority of cases. That?s not to say a faster processor makes no difference, but by far the best upgrade to yield a higher framerate is a new GPU, even with a relatively modest processor.

This is one of the few things nVidia?s marketing page gets right when they talk about building a balanced system.
Actually the reviews in this thread indicate a CPU and GPU would result in similar scaling if either was a significant bottleneck.

Originally posted by: toyota
while using even a mild 4670 in my 5000 X2 I never got even close to the benchmark scores I saw on the web for that card. yet when I used that same 4670 in a system with an E7300 and especially a Q9300 there was a nice increase in minimum framrerates in all games and average framerates in newer ones. I then actually got those benchmark scores review sites were getting. heck with with the Q9300 Far Cry 2 had a massive increase in performance over the 5000 X2. with a stronger card the differences would have been even more dramatic.

I learned a lot about how different cpus can affect gaming. I got so fed up with trying to overcome what my 5000 X2 cpu couldnt do that I went ahead a built a decent computer. theres nothing better than a well balanced pc for gaming and now I finally know what thats like.
Very good points and all too true. It is truly amazing how some people refuse to embrace change and discard common misconceptions. Glad you were curious enough to test on your own, as that's ultimately the only way to truly change one's opinion, through first-hand experience.