Official GIGABYTE Forum

The PCIe 3.0 'Bug', why no fix?

Azari

  • 8
  • 0
The PCIe 3.0 'Bug', why no fix?
« on: May 12, 2014, 10:52:30 pm »
Hi,
I'd like to get to the point straight off the bat, my current system specs are as follows:

MOBO: Gigabyte GA-Z68AP-D3 Rev2.0 (Flashed with UEFI UA9 BIOS)
CPU: Intel Core i5-2400 @ 3.8Ghz
RAM: 8GB Corsair Vengeance 1866Mhz
GPU: Gigabyte Nvidia Geforce 650 Ti (GV-N65TOC-2GI)
PSU: OCZ 700W
HDD: Samsung HD502HI

Around December of last year I upgraded my old age pensioner XFX 8400GS to a Gigabyte 650 Ti. I was thinking the upgrade would be quick and painless, turns out I was wrong, I wasted weeks trying to get the new card to run in the PCIe x16 slot then finally gave up and unwillingly made do with the slower x4 slot below it.

I have recently been searching around for solutions again and found out it was not the fault of my own, it's a so called PCIe 3.0 'bug'. When purchasing the card, I was aware that the i5-2400 only supports PCIe 2.0. This was fine as I knew the graphics card I had bought was backwards compatible with all PCIe interfaces.

Just wondering if there are any advancements in this area? Or have Gigabyte not cared to explain what the deal is with this problem.

This leaves my loyalty with Gigabyte very questionable.

Cheers.
« Last Edit: May 12, 2014, 10:59:00 pm by Azari »
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

dmdilks

  • 3094
  • 43
  • "If it isn't broke don't fix it"
    • http://dmdcomputerservice.webs.com/
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #1 on: May 13, 2014, 02:02:13 am »
I know you are not going to like the answer I going to give you. This board was not made for the UEFI bios. People wanted it yes they got it.

The thing is that the board doesn't support 3.0 cards. This is not uncommon for some motherboards to do that.

The only Z68 board that might support 3.0 cards is the ver 3.0. When You started you had 2 strikes against you with the board & CPU.

I don't know if this link will help you  - http://forum.giga-byte.co.uk/index.php/topic,13818.0.html

Plus one other thing is try and boot from legacy mode. UEFI was really made for windows 8. Plus this board is really too old that they are not going to anything.

« Last Edit: May 13, 2014, 02:04:36 am by dmdilks »
X299X Aorus Master, i9-9940x-3.30Ghz, 64gb G-Skill DDR4-2400, MSI RTX-3070 8GB, Cooler Master case, Thermal-take PSU 850w, 1-M2-NMVe SSD-512gb, 3-Pny 1TB SSD, 2-WD Raptors 1TB, Win 10 pro 64bit, Asus 35" 144Mhz Monitor.

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #2 on: May 13, 2014, 09:05:40 am »
Thanks for the reply, although I don't know what you mean by the board doesn't have PCIeGen3.0, funny thing is, it states it on here in the specification (http://www.gigabyte.com/products/product-page.aspx?pid=4015&m=n#ov) and it also brags the fact that it does support PCIe3.

So in theory, my having an Intel Core i5-2400 shouldn't stop anything from running at PCIe2.0 in the x16 slot, as the CPU, motherboard AND the GPU all state they are backwards compatible with all previous PCIe interfaces.

So, this being said;
I have a motherboard that supports PCIe3.0
BUT I have a sandy bridge which only supports 2.0
Now both the GPU and Motherboard states in the specification on gigabyte they can work in PCIe2.0 mode for unsupported CPU's.

So, what's the deal?

Cheers again.
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

dmdilks

  • 3094
  • 43
  • "If it isn't broke don't fix it"
    • http://dmdcomputerservice.webs.com/
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #3 on: May 13, 2014, 10:01:24 pm »
(All PCI Express x16 slots conform to PCI Express 3.0 standard.) * To support PCI Express 3.0, you must install an Intel 22nm CPU.

Yes that is what it says, but you need a Ivory bridge CPU to do that & you have a Sandy bridge that doesn't.

That is why it isn't working. Not all Z68 support 3.0 card no matter what you do. You are one of the lucky ones to have one.

If you change the CPU it might work or it still might not work sorry. I didn't really look at the spec of the board I was only going by other Z68 boards.
« Last Edit: May 13, 2014, 10:02:39 pm by dmdilks »
X299X Aorus Master, i9-9940x-3.30Ghz, 64gb G-Skill DDR4-2400, MSI RTX-3070 8GB, Cooler Master case, Thermal-take PSU 850w, 1-M2-NMVe SSD-512gb, 3-Pny 1TB SSD, 2-WD Raptors 1TB, Win 10 pro 64bit, Asus 35" 144Mhz Monitor.

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #4 on: May 14, 2014, 02:16:49 am »
I'm not interested in getting the PCIe3.0 interface, I just find it a little hard to get my head around why it can't just run at PCIe2.0 in the x16 slot.

What I got from what you have said, is any newer graphics card that is inserted into x16 requires an Ivy Bridge, notably I used to have an old PCIe1.1 GPU that ran just fine in the x16 slot @ x16 speed.

Every single forum I've checked has said the CPU you own has no knock on effect on how to top PCI slot utelises your GPU. For example, if your CPU doesn't support 3.0 it should run the top slot at 2.0 instead. Or is the 'backwards compatible' statement on the specification on the products I bought just there for show?
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

dmdilks

  • 3094
  • 43
  • "If it isn't broke don't fix it"
    • http://dmdcomputerservice.webs.com/
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #5 on: May 14, 2014, 04:07:07 am »
Here are a couple of screen shots of a old board 775 and your board Z68. This might bring some light what I'm saying.

How the CPU is used and not used. Plus why it works in the x4 slot and not the x16 slot. All video today runs through the CPU on the x16 slot.

On the older boards the x16 & x4 slot ran through the chip-set first. On your board too the x4 is running through the chip-set first.



X299X Aorus Master, i9-9940x-3.30Ghz, 64gb G-Skill DDR4-2400, MSI RTX-3070 8GB, Cooler Master case, Thermal-take PSU 850w, 1-M2-NMVe SSD-512gb, 3-Pny 1TB SSD, 2-WD Raptors 1TB, Win 10 pro 64bit, Asus 35" 144Mhz Monitor.

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #6 on: May 14, 2014, 06:19:00 pm »
Okay I understand.

This is kind of funny though, I'm currently speaking to Gigabyte support.

They sent me a screenshot of the exact build I have with the exact GPU in the top PCI slot (Except I use a 2400, they used a 2500K)

They later told me it could be the x16 slot on my motherboard, this is getting strange  :-\

EDIT: Just want to say thanks for helping me understand the concept, I just find it bizarre how Gigabyte got it to work ???
« Last Edit: May 14, 2014, 07:10:29 pm by Azari »
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

dmdilks

  • 3094
  • 43
  • "If it isn't broke don't fix it"
    • http://dmdcomputerservice.webs.com/
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #7 on: May 14, 2014, 08:09:02 pm »
That is nice but maybe that CPU can run the card where yours can't. They should have done it with everything the same that isn't a good test.

X299X Aorus Master, i9-9940x-3.30Ghz, 64gb G-Skill DDR4-2400, MSI RTX-3070 8GB, Cooler Master case, Thermal-take PSU 850w, 1-M2-NMVe SSD-512gb, 3-Pny 1TB SSD, 2-WD Raptors 1TB, Win 10 pro 64bit, Asus 35" 144Mhz Monitor.

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #8 on: May 16, 2014, 01:21:19 am »
Possibly.

Replied to them saying it's nothing wrong with hardware, as I have tried and tested all my components in multiple builds with different set ups.

Not heard a peep from them in 2 days, customer service at its best ::)
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

shadowsports

  • 2259
  • 67
  • Xbox One, Drives STI, Use QVL RAM For Best Results
    • Gigabyte US
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #9 on: May 16, 2014, 04:33:20 pm »
Azari,
All of what dmdliks has said is true....

My z68 rig is v3 and my "newer' card is PCIe 3 running at 2.0 spec, x16 slot  If I were to put an Ivybridge CPU in, I'd get PCIe 3 support, but this would not be $$$ well spent (marginal performance increase).  Time to start eyeing your next upgrade, sir.  8)  I attribute my success to 1, having a rev 3 board and 2, higer end 2600k cpu.

    
« Last Edit: May 16, 2014, 04:42:18 pm by shadowsports »
Z390 AORUS PRO (F10) \850w, 9900K, 32GB GSkill TriZ RGB - 16-18-18-38, RTX 3080Ti FTW3 Ultra, 960 Pro_m.2, W11
Z370-HD3P (F5) \750w, 8350K, 8GB LPX 3200 - 16-18-18-38, GTX 970 FTW SC, Intel SSD, 2TB RAID1, W11
Z97X-UD5H \850w, 4790K, 32GB Vengeance, RTX 2080 FTW

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #10 on: May 17, 2014, 11:44:40 am »
I'd love an i7 :-[

So to conclude; I have an i5-2400 which means I CAN'T run at PCIe3.0 in the x16 slot due to sandy bridge (Obviously) but still CAN'T run at PCIe2.0 speeds in the x16 slot, I can only run in the bottom x4 slot (which when I've ran benchmarks doesn't actually go over x1 speeds).

It's just baffling I can't even get it to run in the x16 slot at all, card gets power and I can hear everything functioning normally except I have no display, but the x4 slot works just fine, but at x1 speeds.

Extremely annoying :'(

Cheers.
« Last Edit: May 17, 2014, 11:45:43 am by Azari »
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)

shadowsports

  • 2259
  • 67
  • Xbox One, Drives STI, Use QVL RAM For Best Results
    • Gigabyte US
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #11 on: May 17, 2014, 04:37:04 pm »
I'm sure its a combination of factors including how the various manufacturers have chosen to implement support of a feature.  Different CPU, and/or another card might work in the x16 slot, but I'm not going to suggest you do that because in the end, I believe its likely the CPU.  It might be something as silly as the stepping.  Feature set implemented on "K" series for example.  These are the gotcha's that give us our love/hate relationships with all hardware.

You have looked at all your BIOS settings yes?  Int Display First PCIe x16, and Enable if no PEG, and so forth.  ??

I'm looking at your first post I see you said this problem has left you questioning your brand loyalty.  Truth is.. The grass isn't always greener on the other side...  except you should be running EVGA video cards.  (kidding)

And all this... said by the man who is dumping Asus after 13 yrs. I'm waiting for the right board/Rev before I leap.  But I am going to give gigabyte a try.

« Last Edit: May 17, 2014, 05:30:32 pm by shadowsports »
Z390 AORUS PRO (F10) \850w, 9900K, 32GB GSkill TriZ RGB - 16-18-18-38, RTX 3080Ti FTW3 Ultra, 960 Pro_m.2, W11
Z370-HD3P (F5) \750w, 8350K, 8GB LPX 3200 - 16-18-18-38, GTX 970 FTW SC, Intel SSD, 2TB RAID1, W11
Z97X-UD5H \850w, 4790K, 32GB Vengeance, RTX 2080 FTW

Azari

  • 8
  • 0
Re: The PCIe 3.0 'Bug', why no fix?
« Reply #12 on: May 18, 2014, 08:05:59 pm »
Haha, I'd have thought both a Gigabyte motherboard and GPU combined would have reduced little bugs like these. Turns out it's just the CPU, I was so close to buying a 2500K but never bothered as I don't really overclock that much. I guess this is something I'll have to deal with in the future, annoying as it is :-[

I understand, I have quite a love/hate relationship when it comes to my machine. Troubleshooting is just so tedious and frustrating when it's little niggly bits like this.

Thanks for all the help understanding anyhow.

BOIS settings don't change a thing. Clearing CMOS and changing to settings you mentioned don't do anything, just get the same issue and have no onboard graphics when PEG only is applied.

Regarding my loyalty to Gigabyte, we'll just have to wait and see how long this build lasts before I take the plunge and buy from a different manufacturer, although it's always just a matter of time before you run into some problems with those as well ;)

Thanks.
Windows 7 Ultimate 64bit & OSX Mavericks 10.9.3
Intel Core i5-2400 @ 3.8GHz
Gigabyte Z68AP-D3 Rev2.0 (Ua9 UEFI Bios)
8GB Corsair Vengeance Blue @ 1866MHz
Gigabyte Nvidia GeForce 650 Ti 2GB Dual WindForce Edition
Samsung HD502I HDD
OCZ 700W
NZXT H230 (Cooled by x5 Arctic Cooling F12 PWM fans)