PCI-E Gen 2.0 Bug X38/ X48/X58

[donmarkoni]

Just run nvidia inspector and select the P12 state from the OC side. You will see the apparent clock ranges. You can also use Fermi BIOS Editor to see how some values are actually set to 0MHz. Don't call nonsence on things when you don't have access to all the tools to look at things properly yourself

I have both tools.
When you select P12 and see range from 13 to 101 it doesn't mean that VGA will change clocks by itself. Just click Unlock Min and Unlock Max and you will see range from 0 to 3000MHz for everything. It doesn't mean it will work and for sure doesn't mean that VGA will change speed to anything that is not preset.
And all the boxes in FeBiTor are just BIOS registers to be changed or not, and doesn't represent actual clocks. Newer ATI VGA's might have 0MHz performance level, as it is known that they can go to sleep, but NVidia's can't.

[NBK]

Hi,

Do i have this Nvidia bug???

I am using 182.08 Nvidia driver.

thanks
NBK

[NBK]

If i use the 258.96 Nvidia drive, i appear to be running 1.0 PCI-E.

[donmarkoni]

If i use the 258.96 Nvidia drive, i appear to be running 1.0 PCI-E.

Is it like that right after start, or it drops couple of seconds later?
Check this example on first page, post 21.

[JohnZS]

Why the thread still says "PCI-E Gen 2.0 Bug X38/ X48/X58"?

Good question donmarkoni
At the time this thread was posted there was an issue (perhaps related or not?) which meant PCI-E Gen 2.0 did not work on ANY non nForce based motherboard, however this WAS resolved and it may have only been an issue in one, perhaps two 190 series Forceware drivers. This is confirmed with PCI-E Gen 2.0 NOT working on a P5E3 Premium (X48) using a GTX 285 graphics card. And also not working on a Biostar X58A motherboard. However 195 Forcewares and newer appear to not have this bug/issue/lock on the X48/X58 platform, but the X38 is confirmed locked out of Gen 2.0 by nVidia t the moment due to a PCI-E Link pState "bug"
nVidia are considering making changes to their driver code to accommodate for PCI-E Gen 2.0 on X38 based systems and would like our opinions on the matter.
John

[STEvil]

Indeed STEvil, I must admit I was concerned when I saw 102C and informed Mike that his GTX 480 was running extremely toasty. For comparisons sake a GTX 295 (Single PCB) in my system runs upto 86-88C after one whole hour of Furmark @ 1920*1200 16XFSAA. I think that is "safe" as during heavy gaming the card never goes over 70C (according to eVGA precision tool).

You are right, I think the DX58BSO had this issue fixed in a BIOS update.


Listed as a fix for July 2009?

Also in May 2009


But then again I know nothing
John

If Intel can fix the bug with an ACPI table update then all that needs to be done is to club some engineers at assus/msi/etc and get them to properly validate their ACPI tables...

edit - For the record, ACPI tables have a long history of being horrible quality, just look how badly they're broken under *nix most of the time lol.

[NBK]

Is it like that right after start, or it drops couple of seconds later?
Check this example on first page, post 21.

It does not drop to 1.0 after a couple of minutes...

[slim142]

I have nForce 790i drivers 196.21

GPU Z reports 2.0
CPU Z reports just PCI Express X16

nvidia control panel also reports PCI Express X16 GEN 2.0, so anything wrong here?

is CPU Z suppose to report 2.0 or GEN 2.0 or something like that too?

[STEvil]

CPU-Z does not differentiate between PCi-E 1.x or 2.x.... nobody can tell anything is wrong without testing during GPU usage while monitoring. Also the 790i chipset is not an X series chipset, though I guess there is a possibility the problem could exist if nVidia didnt bother to do qualification testing in their own labs with their own chipset..

[donmarkoni]

Good question donmarkoni
At the time this thread was posted there was an issue (perhaps related or not?) which meant PCI-E Gen 2.0 did not work on ANY non nForce based motherboard, however this WAS resolved and it may have only been an issue in one, perhaps two 190 series Forceware drivers. This is confirmed with PCI-E Gen 2.0 NOT working on a P5E3 Premium (X48) using a GTX 285 graphics card. And also not working on a Biostar X58A motherboard. However 195 Forcewares and newer appear to not have this bug/issue/lock on the X48/X58 platform, but the X38 is confirmed locked out of Gen 2.0 by nVidia t the moment due to a PCI-E Link pState "bug"
nVidia are considering making changes to their driver code to accommodate for PCI-E Gen 2.0 on X38 based systems and would like our opinions on the matter.
John

As I can't say with certainty, I won't contradict you, but I'm pretty sure I would notice it when I bought DFI X58. It was about a 2-3 months after release and I was using all BIOS'es from the first one on...
I do understand possibility of bug not being fixed for long time on X38/X48, but X58 was (SLI) certified by NVidia...
Anyway, it's nice that NVidia is helping solve problem with older chipset.

It does not drop to 1.0 after a couple of minutes...

I'm not sure I understand you... Please take the screenshot right after starting GPU-Z (within the first 15 seconds) and post it.

[NBK]

As I can't say with certainty, I won't contradict you, but I'm pretty sure I would notice it when I bought DFI X58. It was about a 2-3 months after release and I was using all BIOS'es from the first one on...
I do understand possibility of bug not being fixed for long time on X38/X48, but X58 was (SLI) certified by NVidia...
Anyway, it's nice that NVidia is helping solve problem with older chipset.



I'm not sure I understand you... Please take the screenshot right after starting GPU-Z (within the first 15 seconds) and post it.

@donmarkoni: Here is the screenshot GPU-Z after about 15 seconds...

[donmarkoni]

@donmarkoni: Here is the screenshot GPU-Z after about 15 seconds...

No problems for you then! As you see, there is remark 2.0 besides 16x. Could be that the BIOS fixes it. It is X38 board with X48's board BIOS, right?

[NBK]

No problems for you then! As you see, there is remark 2.0 besides 16x. Could be that the BIOS fixes it. It is X38 board with X48's board BIOS, right?

Yes, I am using 1001 BIOS for Rampage Formula X48. I am also using an oldl Nvidia driver. IIRC, i think when i was using Maximus Formula's BIOS, i had that PCI-E 1.1 bug.

I tried the most recent Nvidia driver. The PCI-E was reset to 1.x in GPU-Z.

Whew! At least i do not have this issue!

I am also not going to upgrade my video card to the GeForce GTX470 until this issue is resolved...

NBK

[JohnZS]

FYI Forceware driver 186.18 or older does not have this issue as it does not have the pState feature and X38 chipset block coded into the driver.
I recall Rampage Formula itself does not suffer from the PCI-E Gen 2.0 bug, BUT Maximus Formula flashed to Rampage Formula DOES suffer from the bug/lock.
Install any 190+ series or newer forceware to see for yourself...
John