1080ti Strix reporting PCI 2.0 16x on a PCI 3.0 16x slot in GPU Z

[azazl187]

Hi All,

As the title says, GPU Z is reporting PCI 2.0 in a PCI 3.0 slot.... If i put the GPU to 100% load it still reports the same.

I have a sound card and several hard drives using Sata ports as well as an over clocked i7 2600k sitting at 4.5ghz.

What do you think the reason could be for this? Am i choking up the bandwidth somehow?

Thanks

 

[pandem0nium]

as well as an over clocked i7 2600k sitting at 4.5ghz.

There's your reason. Sandy bridge is only capable of Gen 2 speeds. Dropping in, say a 3770k would give you the gen 3.

 

[r7slayer]

AFAIK Sandybridge on the LGA 1155 socket does not support PCI_E 3.0 only 2.0. It was Sandybridge E And the X79 platform that started with it. Ivybridge started with the PCI_E support.

edit -

Ahh beaten to it above, just haha.

Also i would like to add i would not be worried to be fair because this has been tested before going from 2.0 to 3.0 and it had almost zero benefit to single card solutions.

 

[azazl187]

lol ok so that explains it! Its a Z68 board..... Ill have to take a look and see what he highest rating CPU is tat i can drop into it.... I remember reading that Ivybridge was a pretty poor CPU and the as you say, the benefits of moving from PCI 2.o to PCI 3.0 at 4k is next to nothing.

A new 3770k is 300 quid anyway so i may as well just start saving and get a new Mobo, CPU and RAM

 

[r7slayer]

lol ok so that explains it! Its a Z68 board..... Ill have to take a look and see what he highest rating CPU is tat i can drop into it.... I remember reading that Ivybridge was a pretty poor CPU and the as you say, the benefits of moving from PCI 2.o to PCI 3.0 at 4k is next to nothing.

A new 3770k is 300 quid anyway so i may as well just start saving and get a new Mobo, CPU and RAM

Unless you seem CPU limited in games i wouldn't worry too much. I've gone from a 3770k at 4.6Ghz to a 8700k which has two more cores and a total of 4 more threads at 5Ghz (was 5.1 but had slight instability) and only noticed a slight improvement in games. Only games that were noticeable were ones that favoured more threads such as GTA and battlefield series etc. Now this is me at 1440p, if you're gaming at 2160p then you'll be putting more stress on your GPU meaning you'll be GPU limited before CPU. Some food for thought

 

[MagicBoy]

How do you know that you're running out of PCI-E bandwidth?

 

[azazl187]

Unless you seem CPU limited in games i wouldn't worry too much. I've gone from a 3770k at 4.6Ghz to a 8700k which has two more cores and a total of 4 more threads at 5Ghz (was 5.1 but had slight instability) and only noticed a slight improvement in games. Only games that were noticeable were ones that favoured more threads such as GTA and battlefield series etc. Now this is me at 1440p, if you're gaming at 2160p then you'll be putting more stress on your GPU meaning you'll be GPU limited before CPU. Some food for thought

Yes this is why i have not taken the dive into the latest gen of hardware. Its simply not worth me upgrading the CPU, MOBO and RAM for a couple more frames at 4k. Maybe when the next 9xxx series is released.....

 

[azazl187]

How do you know that you're running out of PCI-E bandwidth?

if you mean the difference between PCI 2.0 and 3.0 then there is plenty of citation to look up. Do i know that im losing bandwidth on my personal rig? No not really and if i was it would be about 1 or 2%

 

[Kaapstad]

if you mean the difference between PCI 2.0 and 3.0 then there is plenty of citation to look up. Do i know that im losing bandwidth on my personal rig? No not really and if i was it would be about 1 or 2%

A good trick by pro benchmarkers in some software is to switch from PCIE 3.0 to 2.0 as this can give better results due to less stress on the CPU.

OP stick with what you have.

 

[MagicBoy]

if you mean the difference between PCI 2.0 and 3.0 then there is plenty of citation to look up. Do i know that im losing bandwidth on my personal rig? No not really and if i was it would be about 1 or 2%

Well aware of that, was asking how the OP was pinpointing PCIe as his bottleneck.

Same discussion was had with PCIe 1.0 and 2.0. I ran a PCIe 2.0 card (AMD 5850) in a 1.0 board for a couple of years. Performance drop due to the bus was an insignificant 5% on a really bad day. The lower IPC of the C2Q vs an i5 was more of a bottleneck to the GPU.

 

[azazl187]

Well aware of that, was asking how the OP was pinpointing PCIe as his bottleneck.

Same discussion was had with PCIe 1.0 and 2.0. I ran a PCIe 2.0 card (AMD 5850) in a 1.0 board for a couple of years. Performance drop due to the bus was an insignificant 5% on a really bad day. The lower IPC of the C2Q vs an i5 was more of a bottleneck to the GPU.

Yes i did wonder if it was a loaded question. My GPU is the bottleneck at this point that is for sure

 

[azazl187]

A good trick by pro benchmarkers in some software is to switch from PCIE 3.0 to 2.0 as this can give better results due to less stress on the CPU.

OP stick with what you have.

OK thanks