Settle a bottlenecking argument

No F$#king SH!T!!! Now you understand? What, did you have a revelation or something? Yes the performance of the GPU was not raised, only the CPU which in turn netted higher FPS, thereby revealing a CPU bottleneck at the given speed of 2.8ghz.

 

Sure, @ 1024x768 you have proven that your GPU is bottlenecked by your CPU. I do not disagree on that. BUT. is this real world example.

But @ 1600x1200, we will see that the CPU is bottlenecked by the GPU instead. So who is right?

So it all depends on what resolution you want to run at.

I have included the link to thread about this discussion above

***edit***
if you play the game @ 1440x900 then you should run the test at that resolution too. Lower the cpu to 2.6, then to 2.8, then to 3.0 as you had for 1024x768. Come back with your results. Does your CPU still bottleneck your GPU? That is what we want to know.

AND PLEASE, everyone calm down!!! no need to call each other idiots. I really am debating for the sole purpose of knowledge. I thank you spectacular for taking the time to run the test for us. Anyone else want to do the same, it would be much appreciated.

 

Lol.

 

haha its not a CPU bottleneck with the GPU. All you are doing is increasing how well your CPU calculates physics and AI and that is increasing the FPS. This has nothing to do with a GPU. That is in fact what this discussion is about isnt it?

GODzillaSDM he does not understand what the purpose of running at higher res to load the GPU has to do with showing a bottleneck. The kid is just running his mouth with half and half knowledge of computer basics.

 

You are a moron. No, by raising the CPU you are not only increasing the physics calculations, you are also increasing ther rate at which the CPU feeds the GPU with data....

 

NO YOU ARENT IF YOU ARENT STRESSING THE GPU. God are you that dense. Ive already reported you to the forums for your language and general attitude. So drop it. You are wrong here and it makes me laugh that you dont know it.

 

well, physics and such is part of bottleneck. It all winds down to FPS. If a game requires much phsyics, then it is bottlenecked by CPU.

I also made an edit to my previous post, but afraid you guys didn't read, cuz we have been posting like crazy.. here's what i said.

***edit***
if you play the game @ 1440x900 then you should run the test at that resolution too. Lower the cpu to 2.6, then to 2.8, then to 3.0 as you had for 1024x768. Come back with your results. Does your CPU still bottleneck your GPU? That is what we want to know.

AND PLEASE, everyone calm down!!! no need to call each other idiots. I really am debating for the sole purpose of knowledge. I thank you spectacular for taking the time to run the test for us. Anyone else want to do the same, it would be much appreciated.

 

That is true but he believes that if you increase the speed of the CPU and the FPS go up then the GPU must be bottlenecked. That is his thinking and this is wrong.

This is only true if the GPU is truly being put to its full potential because then you are showing the bottleneck. Why is this so hard for him to understand? If the GPU isnt needing more information than the CPU can send then you dont have a bottleneck. It cant need more than your CPU is sending if you are running at such a low res. If you would read that article it tells you everything I am saying.

 

I can't run the in game test @ 1440x900 because as soon as you slesct performance options it will immediatly default to it's standard resolutions. In order to play FEAR @ widescreen settings you have to edit the cfg file. I can only offer you the same performance test with Fettel or I can wait a bit more and lookt for a more stressing point in the intro.

But you guys are missing the point. There is still a CPU bottleneck even up to 1600x1200 resolutions with a 7900GT and a 2.8Ghz CPU.

I'll run the test once more just for the sake of arguement.......brb.....

 

Look, you abviously have no idea what a bottleneck is.

I can't run the in game test @ 1440x900 because as soon as you slesct performance options it will immediatly default to it's standard resolutions. In order to play FEAR @ widescreen settings you have to edit the cfg file. I can only offer you the same performance test with Fettel or I can wait a bit more and lookt for a more stressing point in the intro.

But you guys are missing the point. There is still a CPU bottleneck even up to 1600x1200 resolutions with a 7900GT and a 2.8Ghz CPU.

I'll run the intro once more just for the sake of arguement.......brb.....

 

you are about to prove yourself wrong. When you run this in a higher resolution the FPS will not be directly dependent on the CPU speed. If you notice on your last test every 200mhz the FPS went up 4 fps. It may still go up (if you go back and read in the beginning I said if I was wrong I would be OCing more) but it wont be in a pattern like that if it does.

 

Report me all you want, it's crying little babies like you that can't handle when they are wrong so they go and tell daddy that "the bad man is picking on me".....somebody call a whaaaambulance

 

No Spectacle, this is a computer forum we come on here to learn and share computer knowledge not put up with drama queens with dirty mouths like yourself. Many users just like yourself have been removed for acting like childs on here. If you want to act like that then go to a public forum or a general chat not a computer tech forum. Also this is a family based forum so fowl language like you have displayed isnt tolerated.

 

Ugh, I just read everything up from my post last night, this thread really exploded.

Spectacle, I would appreciate your calming down a bit, regaining some maturity, and countering this arguent:

In the article both LedHed and I posted, http://digital-daily.com/video/processor_dependency/ , we see on the fourth page, http://digital-daily.com/video/processor_dependency/index04.htm , a graph of a FEAR benchmark. We quite clearly see the shelving effect talked about in the article, where the CPU is waiting for the GPU to finish its intense 1600*1200/soft shadow calculations [and CPU frequency clearly has almost no effect on the fps count].
I hope you understand this very basic concept LedHed has been trying to explain to you.
Now tell me why, whether my proc is a peice of **** or not, I wouldn't want to 'play catch up' with a 2nd 7800, seeing that I play all my games at 1600*1200 with 'max settings', AA and AF etc etc? [and consequentially, get 'shelved' as shown in the article]


Edit: Have to go, won't be back till tomorrow or Thursday. I'm quite busy on school days.

 

OK first off all I apologize for the name calling, but the kid is dense. But anyway here are some more screens of...

2.8ghz.....1440/900 4x AA All settings max



3Ghz.....1440/900 4xAA All settings max.....



Now let me guess....that resolution is still not to your standards?

This was done with a single 7900GT @ 600/1600. The GPU clocks were not adjusted. Only the CPU......


...but I wash my hands of this. If you are too blind to see what is right there in front of you, you are beyond all help....

 

1440/900? where are the screens of your clocks? Did you run SLI? You gave almost no information unlike last time

 

No SLI, Single card @ 600/1600 only thing that was raised was the CPU clocks.

 

hmm loks to me like spectacle proved his point

 

here is my question, why didnt do you do this an hour ago rather than make a fool out of yourself on here. Just because you showed the 7900GT @600/1600 is bottlenecked does not mean anything you said in this topic has made any sense. I admited from the beginning the FPS could in fact go up, but I was told other wise and dont have the ability to test it. You flat out made a fool out of yourself in here and your theories on computer hardware are way off from the truth. However it looks like I might be looking into overclocking futher now.

 

No he didnt prove his point. He just finally showed us what we have been trying to get him to show us. Thats all I wanted to see.