NVIDIA's shady trick to boost the GeForce 9600 GT

[fornowagain]

http://www.techpowerup.com/reviews/NVIDIA/Shady_9600_GT/4.html

On "normal" VGA cards, when you increase the PCI-Express bus frequency you increase the theoretical bandwidth available between card and the rest of the system, but do not affect the speed the card is running at. On the GeForce 9600 GT, a 10% increase in PCI-Express frequency will make the card's core clock run 10% faster!

The automatic increase of 25 MHz on the PCI-Express bus frequency yields an increase of 25% or 162.5 MHz over the stock clock (assuming a 650 MHz clock board design). With a final clock of 812.5 MHz you can bet this card will perform much better, when used by an unsuspecting user, on an NVIDIA chipset motherboard with LinkBoost.

It is certainly nice for NVIDIA to see their GeForce 9600 GT reviewed on NVIDIA chipsets with LinkBoost enabled where their card leaves the competition behind in the dust (even more). Also it could send a message to customers that the card performs considerably better when used on an NVIDIA chipset? Actually this is not the case, the PCI-Express frequency can be adjusted on most motherboards, you will see these gains independent of Intel/AMD CPU architecture or Intel/NVIDIA/AMD/VIA chipset.

-------------------------------

Explains the 780i results > http://en.expreview.com/2008/02/23/geforce-9600gt-review/?page=12

 

[fornowagain]

It is a bit weird, we had this on the nvidia chipset 650i and I wondered what is the point when pci-e frequency does not limit the cards anyway.

It's not about the PCI-e or GPU limit, its simply that for the first time they've linked the GPU clock gen to the PCI-e frequency. And that the actual core clock isn't accurately reported from the board.

 

[fornowagain]

So the reason the 9600GT looks faster is due to this? So I can make my 8800GT faster by upping the PCI-E frequency? In the bios there is something called robust graphics booster or something like that.

Its the reason it looks faster in certain NV boards. You can't make the 8800GT faster from increasing the PCI-e frequency because its GPU clock is tied the cards own 27MHz crystal. The 9600GT is referencing the PCi-e frequency as it clock generator.

 

[fornowagain]

So there is a different component on it?

There's lots I should think. The final GPU core clock is based on a reference frequency. Just like the FSB on a cpu. The normal card uses a crystal to generate the base frequency, the 9600GT is using the PCI-e as the base.

So they take the PCI-e 100MHz/4 = 25Mhz x 29 = 725MHz.
Put up the PCI-e to 115MHz/4 = 29Mhz x 29 = 834MHz

 

[fornowagain]

Follow-up to NVIDIA’s shady trick on 9600GT

So, other cards’ performance did not affect by PCIe. For now only 9600GT have this issue. But will it be the only one? no. We ran a test with a 9500GT ES, the issue is the same, the card gain performance when we increase the PCIe clock.Maybe it is a new feature NVIDIA just did not want us know about it.

Will nForce mobo automatically starts LinkBoost?

No, because nFoce board did not have this feature any more.

9600GT runs better when using it with nForce?

No. It is only a 0.3% increase in Multi-texture test, which can be omitted. If compare it with 3Dmark06 total score that Intel P35 won’t have any difference with 780i.

9600 GT GPU clock increase 1:1 when we increase PCIe clock?

No.PCIe clock change not only affects GPU clocks. First they are not 1:1, second the scale pattern are not the same.

Other cards also have this issue?

For now, no. But in the future we will encounter more 9 series card which have this kind of issue.

Now here is a new question which we can not answer: What PCIe clock affects?

It is clear that the GPU clock and PCIe clock patterns are not the same, so PCIe clock not only affect the GPU clock. But we can not sure which part is the key. Maybe only NVIDIA can answer this question.