Yet another sneaky Nvidia trick

[Tempered81]

So you've heard about the 174 forceware boosting the 9600gt, well check this out...


http://www.techpowerup.com/rev...A/Shady_9600_GT/1.html

 

[krnmastersgt]

Interesting read, thanks for the link. I wonder if we'll see something similar once the 9800s are out.

 

[Tempered81]

i think so. I think it explains the core clock increase on the 9800gtx over the 8800gts

 

[krnmastersgt]

Well personally I don't care that they didn't advertise this little tidbit, and I suspect a lot of the users with 9600 GTs are going to be upping that PCI-E bus speed when they overclock, or if they have an nVidia chipset they can rest easy knowing it was designed to automatically do it.

 

[JACKDRUID]

so is this a good thing or bad?...

 

[taltamir]

i don't see the problem here...

 

[ArchAngel777]

Originally posted by: krnmastersgt
Well personally I don't care that they didn't advertise this little tidbit, and I suspect a lot of the users with 9600 GTs are going to be upping that PCI-E bus speed when they overclock, or if they have an nVidia chipset they can rest easy knowing it was designed to automatically do it.

What does the PCI-E bus speed have to do with this?

 

[ttnuagadam]

Originally posted by: krnmastersgt
Well personally I don't care that they didn't advertise this little tidbit, and I suspect a lot of the users with 9600 GTs are going to be upping that PCI-E bus speed when they overclock, or if they have an nVidia chipset they can rest easy knowing it was designed to automatically do it.

uping the pci-e bus is the best way to fry your card

 

[ArchAngel777]

Originally posted by: ttnuagadam

Originally posted by: krnmastersgt
Well personally I don't care that they didn't advertise this little tidbit, and I suspect a lot of the users with 9600 GTs are going to be upping that PCI-E bus speed when they overclock, or if they have an nVidia chipset they can rest easy knowing it was designed to automatically do it.

uping the pci-e bus is the best way to fry your card

Not to mention the lack of performance increase it would bring...

 

[krnmastersgt]

Did you read the article? The card was automatically raising the pci-e bus and it actually showed performance gains with it, it only does this automatically on nVidia chipset boards but manually overclocking the pci-e bus on a non-nVidia board would yield the same or similar results.

 

[krnmastersgt]

As to whether this is good or bad, well it shows nVidia performing on par with an 8800 GT in some cases even with less physical hardware backing it, a lot of people were confused as to how the 9600 GT was performing so well and this may very well be the reason, performance wise none of us can really complain, however the fact that they didn't market this info is rather strange.

 

[ttnuagadam]

it only improves performance because the clard overclocks itself based on the pcie frequency. overclocking pcie bus is NOT a wise move.

 

[Narse]

Originally posted by: krnmastersgt
Did you read the article? The card was automatically raising the pci-e bus and it actually showed performance gains with it, it only does this automatically on nVidia chipset boards but manually overclocking the pci-e bus on a non-nVidia board would yield the same or similar results.

Where in the article does it talk about the pci-e speed at all? looks to my like they are talking about the crystal multiplier, witch has nothing at all to do with the pci-e speed.

In case of the GeForce 9600 GT the strap says "27 MHz" crystal frequency and Rivatuner Monitoring applies that to its clock reading code, resulting frequency: 783 MHz = 27 MHz * 29 / 1. The NVIDIA driver however uses 25 MHz for its calculation: 725 MHz = 25 * 29 / 1

 

[Martimus]

Originally posted by: Narse

Originally posted by: krnmastersgt
Did you read the article? The card was automatically raising the pci-e bus and it actually showed performance gains with it, it only does this automatically on nVidia chipset boards but manually overclocking the pci-e bus on a non-nVidia board would yield the same or similar results.

Where in the article does it talk about the pci-e speed at all? looks to my like they are talking about the crystal multiplier, witch has nothing at all to do with the pci-e speed.

In case of the GeForce 9600 GT the strap says "27 MHz" crystal frequency and Rivatuner Monitoring applies that to its clock reading code, resulting frequency: 783 MHz = 27 MHz * 29 / 1. The NVIDIA driver however uses 25 MHz for its calculation: 725 MHz = 25 * 29 / 1

The article shows that their is no 25MHz crystal, and that the card actually uses the PCI-e speed divided by 4 to get their base speed. So increasing the PCI-e bus speed will automatically increase the core clock speed by the same percentage. The memory clock speed is actually controlled by an onboard crystal. I think you just didn't realize that the article was more than one page long.

 

[Narse]

missed the 2nd page
heh

 

[smithkt]

Originally posted by: Narse
Where in the article does it talk about the pci-e speed at all? looks to my like they are talking about the crystal multiplier, witch has nothing at all to do with the pci-e speed.

Top of page two.

As many of you know, the PCI-Express bus clock frequency which connects the card to the rest of the system is running at 100 MHz by default. What NVIDIA did is to take this frequency and divide it by four to get their reference frequency for the core clock.

ETA: Beat me to it.

 

[Zap]

Originally posted by: krnmastersgt
As to whether this is good or bad
...
the fact that they didn't market this info is rather strange.

<foil hat on>

I think they didn't market this because they want their 9xxx series cards to outperform their 8xxx series cards in benchmarks, plus they want reviews to show that the best platform for their GPUs are their motherboard chipsets.

Bad that it seems kind of shady, but good in that someone who knows about this can exploit it in their own machine, regardless of chipset and without having to run software for overclocking.

 

[krnmastersgt]

That makes a lot of sense actually, someone will probably say that isn't entirely fair but I like the way the 9600 is shaping out, and it's big brothers are soon to come out which if they follow the same designs, will be monstrous cards. Maybe we can finally play Crysis at 19x12 with max settings at over 60 fps.

 

[ArchAngel777]

Originally posted by: Narse
missed the 2nd page
heh

I didn't see it either...

 

[krnmastersgt]

I think a lot of people didn't see the second page, the title of the first page is 1/4 sooooo read on. I know almost everyone thinks upping the pci-e bus is a bad idea, but that was mainly since there was no performance gains doing it and you shortened the life of the board, however we now have a card that takes advantage, so I'm guessing anyone with a 9600 GT without an nvidia chipset board is going to experiment with pci-e bus speeds.

 

[vanvock]

Whether or not their stuff can take this treatment for prolonged periods remains to be seen but maybe they didn't promote it because they didn't want to have to start buying whatever other hardware you have plugged in there that can't.

 

[krnmastersgt]

I like Zap's reason, and it's doubtful they would release the cards knowing that it would kill boards out there that they built, otherwise they're going to be spending an awful lot of money replacing systems.

 

[thilanliyan]

Isn't HDD corruption possible with increased PCI-e speed?

 

[vanvock]

Originally posted by: krnmastersgt
I like Zap's reason, and it's doubtful they would release the cards knowing that it would kill boards out there that they built, otherwise they're going to be spending an awful lot of money replacing systems.

But what about sound, network, raid etc. in a slot that they didn't build? It oc's the whole bus not just 1 slot right?

 

[krnmastersgt]

Originally posted by: vanvock

Originally posted by: krnmastersgt
I like Zap's reason, and it's doubtful they would release the cards knowing that it would kill boards <<out there that they built, otherwise they're going to be spending an awful lot of money replacing systems.

But what about sound, network, raid etc. in a slot that they didn't build? It oc's the whole bus not just 1 slot right?

Yes but it doesn't OC the components that are plugged in, as the article stated when you increased the bus speeds, as well as the theoretical bandwidth the 8800 card operated as it normally does, increasing bus speeds would damage the board if anything, but assuming it could take that stress theres no reason anything plugged into the pci bus would fail.