2nd GPU operating at "Gen3" while 1st operating at 1.1?

[P A U L]

I'm using a GTX 690... which has 2 GPU on a single board.

I put my card into the lowest PCI slot and GPU-Z says

GPU1 PCIe x16 1.1
GPU2 PCIe x16 3.0


I was quite convinced until now that he bottom slot would be x8 only

...anyway, those of you SLI users... does the GPU1 operating at 1.1 make sense?

 

[EarthDog]

A picture is worth a thousand words.

 

[dfonda]

What is the other GPU? Does it do SLI....Which at this point has been replaced with NV Link.

 

[Mr.Scott]

What is the other GPU? Does it do SLI....Which at this point has been replaced with NV Link.

690 is a x2 card Dave. Both GPU's share the same card.

 

[dfonda]

ah tru antiques...

 

[Mr.Scott]

@ Paul - put a load on it and check if both are then running at 3.0.

 

[EarthDog]

That's where I was going, Mr. Scott.

GPUz allows you to put a load on it to see. Those are likely idle results.

 

[P A U L]

After swapping PCIe slots back and forth I have realized that which DVI port I use determines which GPU is being used. So now GPU1 stays in 3.0 most of the time. Apparently settling down to 1.1 is an intended function.

GPU2 is mostly at 1.1 unless something on that monitor is doing graphical stuff. I can't get GPUZ to stress that monitor... it only activates on the primary. WHich is fine by me at the moment.

However, I was pretty damn convinced that the bottom slot would force the card into 8x. But it doesn't seem to be true. I also have an m.2 drive from my layman's understanding that would affect the PCIe's botom slot speed?

ANyway, i moved the GPU back to top slot because I noticed a 4C drop in temp at the bottom slot. One of the fans sits directly on top of the PSU and is clearly drawing in no air---maybe drawin gin hot air from the PSU.

 

[EarthDog]

I also have an m.2 drive from my layman's understanding that would affect the PCIe's botom slot speed?

That depends. Which socket is it in? You can find out for sure by checking the manual. But one slot is cou connected and wouldn't make a difference. Only the chipset connected m.2 sockets would share lanes like that.

Post magically merged: Jan 30, 2022

ANyway, i moved the GPU back to top slot because I noticed a 4C drop in temp at the bottom slot. One of the fans sits directly on top of the PSU and is clearly drawing in no air---maybe drawin gin hot air from the PSU.

I'd like an image of this....

 

[P A U L]

ANyway, i moved the GPU back to top slot because I noticed a 4C drop in temp at the bottom slot. One of the fans sits directly on top of the PSU and is clearly drawing in no air---maybe drawin gin hot air from the PSU.

I'd like an image of this....

I miss wrote... the fan that is on top of the PSU is not drawing in air. It's pulling away from the heatsink and blowing downward. Anyway, if the GPU were in the bottom slot, there's damn near no room for the hot air to be sent anywhere except barely sideways.

 

[Mr.Scott]

I don't understand. Accelero fans are intake and PSU fan is intake. Neither of those blow down.

 

[P A U L]

Oh, then I must've looked at it wrong.

 

[EarthDog]

The psu fan really has nothing to do with it. It's literally taking air in from the bottom and sending out the back never mixing with case air.

The accellro cooler is taking in air from the bottom fans pointing up.

 

[P A U L]

I wasn't suggesting the PSU itself was the cause/source, per-se. The GPU in the picture is in the top slot, but when I previously had it in the bottom slot,.. there was maybe 2~3mm of clearance between the fan and PSU. Which I assume probably contributed to the 4C increase.

 

[EarthDog]

I wasn't suggesting the PSU itself was the cause/source, per-se. The GPU in the picture is in the top slot, but when I previously had it in the bottom slot,.. there was maybe 2~3mm of clearance between the fan and PSU. Which I assume probably contributed to the 4C increase.

Well yeah, absolutely logical. Less room to freely breathe.

But it isn't drawing in air from the PSU. It may be slightly warmer since it's the top of the PSU, but these are things we cannot measure. I don't want to say it's obvious...but yeah, if you slam the cooler on top of the PSU with little/no room to breathe, I expect a few C warmer temps.

I also expect a few % performance drop since it isn't the 4.0 x16 slot, but something notably less.