Conroe Reviews Released

Status
Not open for further replies.
Here's something that not the source engine, and that does not have very advanced AI or Physics. This is about as normal is a game gets.

I don't want to spend time answering your Bus question before I get this cleared out of the way.

 
You're still running around in circles mate, E6300 vs. X6800 is still only about a 25% framerate increase from what I can tell and you're talking about comparing two cores have about a 50% performance difference.

I don't see what is hard about identifying a bus, as far as I know there's only one bus on a GPU, two if you wanna include the I/O interface
 
Bad decision on benchmark charts.
I don't get what the point of that chart is? What are you trying to prove? That a new chip shows higher frame rates in a time demo?

First of all, it's a time demo. Not Realtime gameplay... so it means nothing. Second of all, Battlefield 2 (Natively) doesn't "support" dual core.

Also, BF2's "FPS" is capped at 100, unless you use ~ game.lockFps 0 to unlock the cap. Which i've done on a OPty 150 @2.8Ghz and a X1900XTX 714/1825 and with patch (1.22) i can reach 130FPS easy. At high settings. I'd be interested in know what GPU they used for that time demo and what it's speeds were?

lol - Flankers is Outofwords - how punning!
 
gaara said:
You're still running around in circles mate, E6300 vs. X6800 is still only about a 25% framerate increase from what I can tell and you're talking about comparing two cores have about a 50% performance difference.

Dude, YOU'RE the one running around in circles. The point - That there is a FPS increase. If the CPU has little to no effect on the gameplay, then there should not be a 33% performance increase due to a greater CPU.
 
Dude, YOU'RE the one running around in circles. The point - That there is a FPS increase. If the CPU has little to no effect on the gameplay, then there should not be a 33% performance increase due to a greater CPU.
ok fine you win, I lose. you're right, I'm wrong, I really don't care, just identify where the bus that's causing the bottleneck is
 
They're running 1600x1200 all maxed out on a Dual X1900XTX Crossfire system.

Even then, its a Timedemo, its supposed to be much more intensive than anything you would face in real gameplay.
 
Oops, didn't see your post.

They communicate via AGP or PCI-e which in turn uses FSB/HT-Bus to NorthBridge to CPU.

This has little to do with the Bus. This has to do with the limited ability of the videocard to calculate what instruction it recieves from the CPU.
 
This has little to do with the Bus. This has to do with the limited ability of the videocard to calculate what instruction it recieves from the CPU.
Then there is no bottleneck if there's no bus operating slower than the information it recieves. A GPU is an independant component, we've gone over this before, it's not a bottleneck, it just happens to be the component that gets put under more load

If you still think a GPU is a bottleneck try running the same game on an X6800, GPUs are much more powerful specialized processors than any CPU which is a centralized logical processor that basically just directs the numerous other smaller specialized processors in your PC.

Like I've said, the only existing bottleneck is the FSB/Mem Bus ratio as the CPU relies or is dependant on extracting instructions from the RAM. If the memory BUS is slower than the front side BUS you've got a bottleneck created as the information starts getting jammed up waiting for the other stuff to get through first. A GPU is never going to have that problem unless you manually modify the clock frequencies of your VRAM and thus create a similar bottleneck between the GPU memory and the GPU itself. If you turn on HDR, AA/AF or try and increase your resolution, it puts more pressure on the GPU. Regardless of video settings the CPU is always performing the same calculations and will thus complete them at the same time regardless of whether it's 640x480 or 1600x1200 or whatever else you wanna cause
 
Status
Not open for further replies.
Back
Top Bottom