• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

CPU Bottlenecking (8800's / 2900xt's)

Teaf

Teaf

Teaf

Soldato
Joined
17 Dec 2006
Posts
8,261
Anyone per chance got a link I could read up on about CPU bottlenecking in regards to the 8800 and 2900xt card ranges? I see a lot of people mentioning the bottlenecking and it greatly reduces performance of the cards in some cases but I've had a look around the forum and googled somewhat, but still can't get any solid info about how much of a reduction happens for particular chips.

Cheers
 
I've got an opty 170 @ 2.5ghz, memory at 1:1 and a BFG 8800GTX OC and I'm getting just below 9200marks.

Compare this to other peoples' scores with c2d setups, and you'll realise that this is poor performance for a GTX.
People with c2d & 320mb 8800s are getting higher scores than me on a mild overclock
 
flibby said:
Anyone per chance got a link I could read up on about CPU bottlenecking in regards to the 8800 and 2900xt card ranges? I see a lot of people mentioning the bottlenecking and it greatly reduces performance of the cards in some cases but I've had a look around the forum and googled somewhat, but still can't get any solid info about how much of a reduction happens for particular chips.

Cheers
Click to expand...

What are you worried for? you have a Conroe @ 2.8GHZ, you will be fine. Though why not push it to 3GHZ for the win :D
 
Not worried really, but yeah I'll stick my CPU back to > 3Ghz at some later point:)

It was more reading purposes I guess, just to see if there was some documentation to read on it, and thought it may have given other people an idea of how much they have to overclock or what they should upgrade too depending on what card they wanna get.
 
fornowagain said:
Basically once the resolution and AA gets up a bit the cpu get less and less important. Makes you laugh how well the X2 does in that test.
Click to expand...


To a degree - the CPU still has to shunt increasingly complex wireframes/whatnots around the 3D environment, does it not? :confused:
 
CPU vertex geometry and physics information gets shunted into the GPU front end, the command processor. From what I've read the CP's are getting better are buffering data from the host interface, so less bottlenecks. Data does of course increase with more geometric entities in object space (points and edges). But it's minor compared to the extra GPU load of an increased number of rasterised triangles in screen space. With higher resolutions the balance changes and GPU processes like pixel shading or antialiasing soon become the bottleneck.
 
Last edited:
fornowagain said:
http://www.tomshardware.co.uk/2007/05/11/system_builder_marathon_uk/page10.html

Budget X2 3800/N4 vs Mid E6600/P965 with 8800

Click to expand...


That graph doesn't make any sense. (assuming both pcs are running the same GPU)

At higher resolutions whilst a cpu becomes less important a E6600 based system with a 8800 will still be quicker than a X2 3800+ based system with the same graphics card.

There also won't be a 29fps difference at 2560x1600 4xAA/8xAF especially if most of the load is on the GPU.


Also personally the only real world situation where I have noticed cpu bottlenecking on my 8800GTS (e4300) is under synthetic benchmarks. (comparing 2.4ghz and 3ghz) as in game I can't see any difference between the two cpu clocks as I am already getting more than playable framerates.
 
They're not running the same GPU. Theyre simply making the point that if you want to game, spend more money on the GPU rather than other components.

The cheaper budget pc with a better graphics card destroys the more expensive pc at higher resolutions. Thats the point.

Also, 3dmark 06 scores should never be used to find cpu bottleknecking since it incorporates the cpu score into the final score.
 
Last edited:
Surely they should have both been using the same GPU for it to be worthwhile. I know that it illustrates CPU bottlenecking to a certain degree, but with the same GPU, you would have seen the effects better.
 
Actually really suprised no big review place has done a bottlenecking article yet, maybe like getting a full system for Core2's and AMD's, then keeping everything the same except the chips and just benchmarking, overclock it, benchmark again, then try the next chip.
 
You must log in or register to reply here.
Back
Top Bottom