Normally cards like that get just as hot as air cooled cards. I had one from Sapphire and it was clocked higher but the temps were similar as once the card got up to temp the water pump was them pushing around hot liquid.
Mind you,.. Quad core laptop with 8800GTS on board (can't use the GTX ver... we'd be carrying a something with the weight of a brick if we used that now, can we? )... inter-changeable screen for those "just-cuz" moments of arrogance... mmmmmmm....
Even Larry Ellison would be saying "I'll take ten."
On the Doom 3 page, they state "even though this game is still CPU limited".
That makes NO sense. All that the CPU does is pass on frames to the GPU, for it to do all the work. As long as a CPU can pass over 60 FPS to the GPU, there is no bottleneck, and even an Athlon XP can do that in Doom 3.
The reason the the frame-rate is so low at higher resolutions is because of the GPU.
Cmon Track you've been around long enough to understand the concept of CPU limitation in 3D graphics, haven't you? As an analogy, think of a pellet stove with an auger. If the firepit itself translates to the GPU (meaning some can burn more pellets than others, because they have a greater volume or better airflow), and the CPU is the auger, you can see if you have an auger that can deliver 2 ounces per minute, and you upgrade the firepit from one that can burn 2 ounces per minute, to one that can burn 3, you won't see a significant performance increase. The same holds true with 3D graphics; if your CPU can send geometry information for, say, 188 Frames in a second in Doom3 (which is the CPU-lock in the benchmark we are discussing), but one card could actually generate 190, while another could generate 225, it doesn't matter. It is only when you test with games that are more GPU-intensive, or when you add variables to your test that increase GPU workload (such as AA or AF), that you can accurately see which GPU's can process more frames (burn more pellets).
Hope this helps!
I wonder how the mobile version Go 8800 GTX will perform with the new quad core laptops when it is released next month.
Nvidia has not announced any mobile version of the Geforce 8800GTX, and even if it's on a 65nm die shrink, thermals would be too high for a mobile GPU.
Cmon Track you've been around long enough to understand the concept of CPU limitation in 3D graphics, haven't you? As an analogy, think of a pellet stove with an auger. If the firepit itself translates to the GPU (meaning some can burn more pellets than others, because they have a greater volume or better airflow), and the CPU is the auger, you can see if you have an auger that can deliver 2 ounces per minute, and you upgrade the firepit from one that can burn 2 ounces per minute, to one that can burn 3, you won't see a significant performance increase. The same holds true with 3D graphics; if your CPU can send geometry information for, say, 188 Frames in a second in Doom3 (which is the CPU-lock in the benchmark we are discussing), but one card could actually generate 190, while another could generate 225, it doesn't matter. It is only when you test with games that are more GPU-intensive, or when you add variables to your test that increase GPU workload (such as AA or AF), that you can accurately see which GPU's can process more frames (burn more pellets).
Hope this helps!
So ur saying that it dosent matter if ur GPU can do 225, as long as the CPU can only do 188, u will only get 188 FPS?
I wonder how the mobile version Go 8800 GTX will perform with the new quad core laptops when it is released next month.
Nvidia has not announced any mobile version of the Geforce 8800GTX, and even if it's on a 65nm die shrink, thermals would be too high for a mobile GPU.
I don't think so. Nvidia should be releasing it within the next few weeks and Clevo has already demonstrated two Santa Rosa laptops at CeBit that will accept the new graphics cards, in SLI.
Hey Track it sounds like you're saying that you only need a CPU that's capable of feeding 60 frames of data, because beyond that the user can't detect higher speeds. In that sense, I agree with you. Consider this, however. If you have a rig with CPU and memory subsystems that can provide more performance than you need, then you have two big advantages: First, you can do other things in the background while gaming (playing music, downloading, virus-scanning, etc.) and second, you are somewhat future-proof because you can upgrade the GPU in a rig and get instant better results, when newer/bigger games come out!