GTX 1070 memory VRM driver or Bios Bug in Micron Memory 1070 cards
  1 / 56    
Update 6 Nov: Bios Updates have been released for he majority of card partners that resolves the Micron artifact and crashing problem. You should check with the support pages of your card to check for requirements and downloads of your specific model of card. Many vendors are distributing a single file that contains bioses for the different models in their range but you need to check. The update utilities I have seen Cheks the version of the card before it will allow the correct update to be applied but your vendors utility may not do that. Please make read any documentation that comes with the download or utility. Links for manufacturers cards I have found are: EVGA: http://forums.evga.com/EVGA-GeForce-GTX-1070-BIOS-Update-v8604500070-m2565056.aspx Gainward: http://www.gainward.com/main/vgapc.php?vgapc_id=79&lang=en Palit: http://www.palit.com/palit/vgapc.php?mid=2&subid=282&lang=en MSI: https://www.msi.com/Graphics-card/support/GEFORCE-GTX-1070-GAMING-X-8G.html#down-bios Gigabyte: http://www.gigabyte.com/products/product-page.aspx?pid=5916#bios Asus: https://www.asus.com/us/support/Download/9/2/36/1/IkF5VLiS13T4hUIG/45/ I have no information about Inno3d, Galax/KAF2, Zotac or PNY so please check their support pages for the bios update availability. [url]https://forums.geforce.com/default/topic/963768/geforce-drivers/gtx-1070-memory-vrm-driver-or-bios-bug-in-micron-memory-1070-cards-/post/4986835/#4986835[/url] This issue has actually been around since the release of the 2nd manufactured batch AIB cards that are using Micron GDDR5 memory modules instead of the original Samsung modules and as far as I can tell, is applicable to all brands. All the Samsung memory cards that I have seen to date, are running bios versions in the range 86.04.1e.00.xx and have no issue if you overclock the memory to high levels +500 and above when the card is sitting in a low voltage powersaving state. If the card is idling at the 300mv level when you click apply to overclock the vram from default levels to somewhere in the range of +400 to +500, The Micron memory cards will produce checkerboard artifacts and either crash the driver or create a video scheduler BSOD. If you lock the voltage in Afterburner first, them you can quite easily run +500 to +600 mhz offset overclocks without the card crashing. That suggests to me that the software that controls memory VRMs on the micron based cards is not increasing the voltage fast enough to supply the higher voltage required to meet the needs of the overclocked Micron vRam. I would imagine that it is a relatively easy fix in the driver, inserting a few milliseconds delay between the commands to increase the memory VRM voltage supply and start running the memory at the higher clock speeds
Update 6 Nov: Bios Updates have been released for he majority of card partners that resolves the Micron artifact and crashing problem. You should check with the support pages of your card to check for requirements and downloads of your specific model of card. Many vendors are distributing a single file that contains bioses for the different models in their range but you need to check. The update utilities I have seen Cheks the version of the card before it will allow the correct update to be applied but your vendors utility may not do that. Please make read any documentation that comes with the download or utility.

Links for manufacturers cards I have found are:

EVGA: http://forums.evga.com/EVGA-GeForce-GTX-1070-BIOS-Update-v8604500070-m2565056.aspx

Gainward: http://www.gainward.com/main/vgapc.php?vgapc_id=79&lang=en

Palit: http://www.palit.com/palit/vgapc.php?mid=2&subid=282&lang=en

MSI: https://www.msi.com/Graphics-card/support/GEFORCE-GTX-1070-GAMING-X-8G.html#down-bios
Gigabyte: http://www.gigabyte.com/products/product-page.aspx?pid=5916#bios
Asus: https://www.asus.com/us/support/Download/9/2/36/1/IkF5VLiS13T4hUIG/45/

I have no information about Inno3d, Galax/KAF2, Zotac or PNY so please check their support pages for the bios update availability.

https://forums.geforce.com/default/topic/963768/geforce-drivers/gtx-1070-memory-vrm-driver-or-bios-bug-in-micron-memory-1070-cards-/post/4986835/#4986835

This issue has actually been around since the release of the 2nd manufactured batch AIB cards that are using Micron GDDR5 memory modules instead of the original Samsung modules and as far as I can tell, is applicable to all brands.

All the Samsung memory cards that I have seen to date, are running bios versions in the range 86.04.1e.00.xx and have no issue if you overclock the memory to high levels +500 and above when the card is sitting in a low voltage powersaving state.

If the card is idling at the 300mv level when you click apply to overclock the vram from default levels to somewhere in the range of +400 to +500, The Micron memory cards will produce checkerboard artifacts and either crash the driver or create a video scheduler BSOD. If you lock the voltage in Afterburner first, them you can quite easily run +500 to +600 mhz offset overclocks without the card crashing.

That suggests to me that the software that controls memory VRMs on the micron based cards is not increasing the voltage fast enough to supply the higher voltage required to meet the needs of the overclocked Micron vRam.

I would imagine that it is a relatively easy fix in the driver, inserting a few milliseconds delay between the commands to increase the memory VRM voltage supply and start running the memory at the higher clock speeds

#1
Posted 09/09/2016 01:18 PM   
You don't fix hardware bugs with a driver. new vbios needed or recall the faulty cards (as msi has)
You don't fix hardware bugs with a driver.

new vbios needed or recall the faulty cards (as msi has)



In Memory of Chris "ChrisRay" Arthington, 1982-2010

Specs:
CPU:Intel Xeon x5690 @ 4.2Ghz, Mainboard:Asus Rampage III Extreme, Memory:48GB Corsair Vengeance LP 1600
Video:EVGA Geforce GTX 1080 Founders Edition, NVidia Geforce GTX 1060 Founders Edition
Monitor:BenQ G2400WD, BenQ BL2211, Sound:Creative XFI Titanium Fatal1ty Pro
SDD:Crucial MX300 275, Crucial MX300 525, Crucial MX200 250
HDD:500GB Spinpoint F3, 1TB WD Black, 2TB WD Red, 1TB WD Black
Case:NZXT Phantom 820, PSU:Seasonic X-850, OS:Windows 7 SP1
Cooler: ThermalRight Silver Arrow IB-E Extreme

WIP:
CPU:Intel Xeon x5660, Mainboard:Asus Rampage II Gene, Memory:16GB Corsair Vengeance 1600 LP
Video:EVGA Geforce GTX 680+ 4GB, Palit Geforce GTX 550ti
Monitor:Pending, Sound:Pending
SDD:Pending
HDD:Pending
Case:NZXT Guardian 921RB, PSU:Corsair 620HX, OS:Windows 7 SP1
Cooler: ThermalRight True Spirit 120M

#2
Posted 09/09/2016 04:12 PM   
I have the exact same issue with my Micron memory on my Palit Super Jetstream 1070 which results in instant freeze on desktop with checkerboard and reboot even with small overclocking like +200. But, if I first increase the voltage, then memory becomes rock stable even at +600 ! There is definitely something not right with voltage management with Micron memory. Please fix it.
I have the exact same issue with my Micron memory on my Palit Super Jetstream 1070 which results in instant freeze on desktop with checkerboard and reboot even with small overclocking like +200.
But, if I first increase the voltage, then memory becomes rock stable even at +600 !
There is definitely something not right with voltage management with Micron memory.
Please fix it.

#3
Posted 09/09/2016 05:17 PM   
[quote="reflex75"]I have the exact same issue with my Micron memory on my Palit Super Jetstream 1070 which results in instant freeze on desktop with checkerboard and reboot even with small overclocking like +200. But, if I first increase the voltage, then memory becomes rock stable even at +600 ! There is definitely something not right with voltage management with Micron memory. Please fix it.[/quote] You do realize the this is a Palit problem and not one of NVIDIA? You should contact Palit directly about this issue.
reflex75 said:I have the exact same issue with my Micron memory on my Palit Super Jetstream 1070 which results in instant freeze on desktop with checkerboard and reboot even with small overclocking like +200.
But, if I first increase the voltage, then memory becomes rock stable even at +600 !
There is definitely something not right with voltage management with Micron memory.
Please fix it.


You do realize the this is a Palit problem and not one of NVIDIA? You should contact Palit directly about this issue.

Intel Core i7-6700K
Asus STRIX Nvidia GTX 1070 @ 378.49
Asus Z170-A
Auzentech X-FI Prelude 7.1
16 GB DDR 4-3000 Corsair Vengeance
Samsung 950 Pro M.2 PCIe 256 GB
Windows 10 x64 17033
formerly known as Roland01 ;)

#4
Posted 09/09/2016 05:30 PM   
It's not only about Palit, MSI has the same problem, even at stock speed. It's a global issue with Micron memory on 1070 from Nvidia.
It's not only about Palit, MSI has the same problem, even at stock speed.
It's a global issue with Micron memory on 1070 from Nvidia.

#5
Posted 09/09/2016 06:34 PM   
[quote="Sora"]You don't fix hardware bugs with a driver. new vbios needed or recall the faulty cards (as msi has)[/quote] It is not an MSI specific problem, other brands with the mircon memory and bios are exhibiting the same symptoms. Both Samsung and Micron cards use the same Nvidia mandated voltage controllers. It may be resolved with an updated base bios, but all of the base bios files come originally from Nvidia and are only tweaked to suit the AIB. Nvidia also writes the drivers used by all cards. Afterburner and the other utilities use the control hooks provided by the driver to adjust the card settings If it was a hardware fault, the cards would crash/artifact regardless of having the voltage locked high or not. The artifacts and crashes only appear when taking the card from a low voltage/default state to a high memory over clocked. If the voltage is already running above about 800mv the artifact issue is not apparent. and the cards run quite happily at at 500-600 mhz or higher over the stock reference clocks. Regardless of where the issue is raised, Nvidia has to do something for the issue to be resolved. That could be either issuing a new version of base bios for micron memory 1070 cards to the Partners for modify and distribute or by adding a line or two to the driver code that adds a slight delay between increasing the voltage and increasing the memory clock. So please, get off the fanboy train and start using observation and facts rather than regurgitating anecdotal "evidence" to solve the issue
Sora said:You don't fix hardware bugs with a driver.

new vbios needed or recall the faulty cards (as msi has)


It is not an MSI specific problem, other brands with the mircon memory and bios are exhibiting the same symptoms. Both Samsung and Micron cards use the same Nvidia mandated voltage controllers. It may be resolved with an updated base bios, but all of the base bios files come originally from Nvidia and are only tweaked to suit the AIB. Nvidia also writes the drivers used by all cards.

Afterburner and the other utilities use the control hooks provided by the driver to adjust the card settings

If it was a hardware fault, the cards would crash/artifact regardless of having the voltage locked high or not. The artifacts and crashes only appear when taking the card from a low voltage/default state to a high memory over clocked. If the voltage is already running above about 800mv the artifact issue is not apparent. and the cards run quite happily at at 500-600 mhz or higher over the stock reference clocks.

Regardless of where the issue is raised, Nvidia has to do something for the issue to be resolved. That could be either issuing a new version of base bios for micron memory 1070 cards to the Partners for modify and distribute or by adding a line or two to the driver code that adds a slight delay between increasing the voltage and increasing the memory clock.

So please, get off the fanboy train and start using observation and facts rather than regurgitating anecdotal "evidence" to solve the issue

#6
Posted 09/09/2016 07:35 PM   
You're wrong about that, each AIB has tools to modify vbios modules, of which vram timings and drive strengths are just a component of that. They can change the outputs priority, voltages, timings, fan curve, clock rates - all without needing to contact nvidia. Why do people always jump to calling fanboyism when they are so wrong its stupid. The vendors are at fault for slapping subpar micron modules on the card and not adjusting the bios for it. Not to mention: You are not guaranteed the ability to overclock. Reference cards are sold as is, tough shit if you can't overclock it because of vendor hardware limits.
You're wrong about that, each AIB has tools to modify vbios modules, of which vram timings and drive strengths are just a component of that.

They can change the outputs priority, voltages, timings, fan curve, clock rates - all without needing to contact nvidia.


Why do people always jump to calling fanboyism when they are so wrong its stupid.

The vendors are at fault for slapping subpar micron modules on the card and not adjusting the bios for it.


Not to mention: You are not guaranteed the ability to overclock. Reference cards are sold as is, tough shit if you can't overclock it because of vendor hardware limits.



In Memory of Chris "ChrisRay" Arthington, 1982-2010

Specs:
CPU:Intel Xeon x5690 @ 4.2Ghz, Mainboard:Asus Rampage III Extreme, Memory:48GB Corsair Vengeance LP 1600
Video:EVGA Geforce GTX 1080 Founders Edition, NVidia Geforce GTX 1060 Founders Edition
Monitor:BenQ G2400WD, BenQ BL2211, Sound:Creative XFI Titanium Fatal1ty Pro
SDD:Crucial MX300 275, Crucial MX300 525, Crucial MX200 250
HDD:500GB Spinpoint F3, 1TB WD Black, 2TB WD Red, 1TB WD Black
Case:NZXT Phantom 820, PSU:Seasonic X-850, OS:Windows 7 SP1
Cooler: ThermalRight Silver Arrow IB-E Extreme

WIP:
CPU:Intel Xeon x5660, Mainboard:Asus Rampage II Gene, Memory:16GB Corsair Vengeance 1600 LP
Video:EVGA Geforce GTX 680+ 4GB, Palit Geforce GTX 550ti
Monitor:Pending, Sound:Pending
SDD:Pending
HDD:Pending
Case:NZXT Guardian 921RB, PSU:Corsair 620HX, OS:Windows 7 SP1
Cooler: ThermalRight True Spirit 120M

#7
Posted 09/10/2016 09:26 AM   
Im wrong? You just paraphrased what I said in the first paragraph of my last post. keep on trolling.
Im wrong? You just paraphrased what I said in the first paragraph of my last post.

keep on trolling.

#8
Posted 09/10/2016 06:36 PM   
return your card, accept it won't overclock or just stop posting. Nvidia has no obligation to you in your issue.
return your card, accept it won't overclock or just stop posting.

Nvidia has no obligation to you in your issue.



In Memory of Chris "ChrisRay" Arthington, 1982-2010

Specs:
CPU:Intel Xeon x5690 @ 4.2Ghz, Mainboard:Asus Rampage III Extreme, Memory:48GB Corsair Vengeance LP 1600
Video:EVGA Geforce GTX 1080 Founders Edition, NVidia Geforce GTX 1060 Founders Edition
Monitor:BenQ G2400WD, BenQ BL2211, Sound:Creative XFI Titanium Fatal1ty Pro
SDD:Crucial MX300 275, Crucial MX300 525, Crucial MX200 250
HDD:500GB Spinpoint F3, 1TB WD Black, 2TB WD Red, 1TB WD Black
Case:NZXT Phantom 820, PSU:Seasonic X-850, OS:Windows 7 SP1
Cooler: ThermalRight Silver Arrow IB-E Extreme

WIP:
CPU:Intel Xeon x5660, Mainboard:Asus Rampage II Gene, Memory:16GB Corsair Vengeance 1600 LP
Video:EVGA Geforce GTX 680+ 4GB, Palit Geforce GTX 550ti
Monitor:Pending, Sound:Pending
SDD:Pending
HDD:Pending
Case:NZXT Guardian 921RB, PSU:Corsair 620HX, OS:Windows 7 SP1
Cooler: ThermalRight True Spirit 120M

#9
Posted 09/10/2016 06:46 PM   
Customers feedback is always valuable in order to improve Nvidia experience present and future.
Customers feedback is always valuable in order to improve Nvidia experience present and future.

#10
Posted 09/11/2016 07:49 AM   
[quote="reflex75"]Customers feedback is always valuable in order to improve Nvidia experience present and future.[/quote] You need to learn what a subvendor is.
reflex75 said:Customers feedback is always valuable in order to improve Nvidia experience present and future.


You need to learn what a subvendor is.



In Memory of Chris "ChrisRay" Arthington, 1982-2010

Specs:
CPU:Intel Xeon x5690 @ 4.2Ghz, Mainboard:Asus Rampage III Extreme, Memory:48GB Corsair Vengeance LP 1600
Video:EVGA Geforce GTX 1080 Founders Edition, NVidia Geforce GTX 1060 Founders Edition
Monitor:BenQ G2400WD, BenQ BL2211, Sound:Creative XFI Titanium Fatal1ty Pro
SDD:Crucial MX300 275, Crucial MX300 525, Crucial MX200 250
HDD:500GB Spinpoint F3, 1TB WD Black, 2TB WD Red, 1TB WD Black
Case:NZXT Phantom 820, PSU:Seasonic X-850, OS:Windows 7 SP1
Cooler: ThermalRight Silver Arrow IB-E Extreme

WIP:
CPU:Intel Xeon x5660, Mainboard:Asus Rampage II Gene, Memory:16GB Corsair Vengeance 1600 LP
Video:EVGA Geforce GTX 680+ 4GB, Palit Geforce GTX 550ti
Monitor:Pending, Sound:Pending
SDD:Pending
HDD:Pending
Case:NZXT Guardian 921RB, PSU:Corsair 620HX, OS:Windows 7 SP1
Cooler: ThermalRight True Spirit 120M

#11
Posted 09/11/2016 01:37 PM   
. First I would like to say thank you for the hard work and thought out explanation of the issue with the Micron Memory. It is not easy to track these down. [quote][color="orange"]It is not an MSI specific problem[/color], other brands with the mircon memory and bios are exhibiting the same symptoms. [/quote] But, NVIDIA makes a REFERENCE chipset / card and the ""Partners"" are responsible for the actual components. The use of MICRON memory instead of the SAMSUNG is an attempt to make their cards cheaper and their profits higher. In doing so, the changes to the vBIOS should come from the ""Partners"" and not NVIDIA due to the fact that it was the ""Partners"" choice to use subpar hardware at the cost of the USERS experience. So, in this case, you really should be hitting MSI up for an updated vBIOS specific to the MICRON memory. Now I would like to point out why I believe you are looking for the VENDOR to supply the vBIOS. If it were coming DIRECTLY from NVIDIA, all cards would have the SAME vBIOS..... [color="orange"]NVIDIA[/color]: 86.04.1E.00.01 : Founders [color="orange"]<------[/color] [color="orange"]MSI:[/color] 86.04.1E.00.34 : Aero OC 86.04.1E.00.5E : Aero OC 86.04.1E.00.01 : Founders [color="orange"]<----- [/color] 86.04.1E.00.70 : Gaming 86.04.26.00.3F : Gaming 86.04.1E.00.40 : Gaming X 86.04.1E.00.41 : Gaming X 86.04.26.00.3E : Gaming X 86.04.1E.00.6C : Gaming Z 86.04.26.00.3B : Gaming Z 86.04.1E.00.55 : Seahawk [color="orange"]Gainward:[/color] 86.04.1E.00.2E : Pheonix [color="orange"]Gigabyte:[/color] 86.04.1E.00.01 : Founders [color="orange"]<-----[/color] 86.04.1E.00.68 : G1 Gaming 86.04.1E.00.78 : WF2OC [color="orange"]EVGA:[/color] 86.04.1E.00.70 : [color="green"]Founders[/color] / FTW / SC [color="green"]<---Where is the .01 vBIOS?!?!?![/color] 86.04.1E.01.70 : FTW 86.04.26.00.70 : FTW So, as you can see, NVIDIA made and released the FOUNDERS edition as the REFERENCE CARD with a REFERENCE vBIOS. After that, it is the ""Partners"" that make all MODIFICATIONS to the vBIOS. .
.

First I would like to say thank you for the hard work and thought out explanation of the issue with the Micron Memory.

It is not easy to track these down.

It is not an MSI specific problem, other brands with the mircon memory and bios are exhibiting the same symptoms.


But, NVIDIA makes a REFERENCE chipset / card and the ""Partners"" are responsible for the actual components.

The use of MICRON memory instead of the SAMSUNG is an attempt to make their cards cheaper and their profits higher.

In doing so, the changes to the vBIOS should come from the ""Partners"" and not NVIDIA due to the fact that it was the ""Partners"" choice to use subpar hardware at the cost of the USERS experience.

So, in this case, you really should be hitting MSI up for an updated vBIOS specific to the MICRON memory.


Now I would like to point out why I believe you are looking for the VENDOR to supply the vBIOS.


If it were coming DIRECTLY from NVIDIA, all cards would have the SAME vBIOS.....


NVIDIA:
86.04.1E.00.01 : Founders <------



MSI:
86.04.1E.00.34 : Aero OC
86.04.1E.00.5E : Aero OC
86.04.1E.00.01 : Founders <-----
86.04.1E.00.70 : Gaming
86.04.26.00.3F : Gaming
86.04.1E.00.40 : Gaming X
86.04.1E.00.41 : Gaming X
86.04.26.00.3E : Gaming X
86.04.1E.00.6C : Gaming Z
86.04.26.00.3B : Gaming Z
86.04.1E.00.55 : Seahawk

Gainward:
86.04.1E.00.2E : Pheonix

Gigabyte:
86.04.1E.00.01 : Founders <-----
86.04.1E.00.68 : G1 Gaming
86.04.1E.00.78 : WF2OC

EVGA:
86.04.1E.00.70 : Founders / FTW / SC <---Where is the .01 vBIOS?!?!?!
86.04.1E.01.70 : FTW
86.04.26.00.70 : FTW




So, as you can see, NVIDIA made and released the FOUNDERS edition as the REFERENCE CARD with a REFERENCE vBIOS.

After that, it is the ""Partners"" that make all MODIFICATIONS to the vBIOS.

.

My Opinion is no more important or right than yours. But if your using the LATEST drivers, go back to the prior set.

Piss Poor tech support blame all issues on drivers and assumes that your an idiot. Find a set that you know worked and see if the problem exists still.

""Don't f with it if it ain't broken!""
Constantly updating drivers is a good way to F'up the whole system.

If the driver has an issue, don't be a fool, remove it.

Just because you can, does not mean you should.

#12
Posted 09/11/2016 03:23 PM   
Interesting. I have a Gigabyte Founders 1080, which has MICRON memory. And I get artifacting under certain situations. Just did some more testing and i can also hit much higher in mem clocks without issues. Seems to be something to do with the performance modes changing. So yep, sounds like this issue, or a very similar one, also effects, Gigabyte Founders Edition 1080's...
Interesting. I have a Gigabyte Founders 1080, which has MICRON memory. And I get artifacting under certain situations.

Just did some more testing and i can also hit much higher in mem clocks without issues. Seems to be something to do with the performance modes changing.

So yep, sounds like this issue, or a very similar one, also effects, Gigabyte Founders Edition 1080's...

#13
Posted 09/11/2016 08:36 PM   
[quote="gozu"]Interesting. I have a Gigabyte Founders 1080, which has MICRON memory. And I get artifacting under certain situations. Just did some more testing and i can also hit much higher in mem clocks without issues. Seems to be something to do with the performance modes changing. So yep, sounds like this issue, or a very similar one, also effects, Gigabyte Founders Edition 1080's...[/quote] 1080s all have Micron GDDR5X Ram. 1070 use GDDR5 ram which is a different animal.
gozu said:Interesting. I have a Gigabyte Founders 1080, which has MICRON memory. And I get artifacting under certain situations.

Just did some more testing and i can also hit much higher in mem clocks without issues. Seems to be something to do with the performance modes changing.

So yep, sounds like this issue, or a very similar one, also effects, Gigabyte Founders Edition 1080's...


1080s all have Micron GDDR5X Ram.

1070 use GDDR5 ram which is a different animal.

#14
Posted 09/12/2016 06:04 AM   
You forgot to mention the micron bioses from vendors other than MSI such as: EVGA 86.04.26.01.70 Zotac 86.04.26.00.4B Gainward 86.04.26.00.27 Asus 86.04.26.00.61 KFA2 86.04.26.00.9C So you are suggesting that EVGA, ASUS, MSI, Palit/Gainward, Zotac, Galaxy all independently wrote their own bioses from scratch? They all contain sections of code that are identical with only relatively minor changes for device IDs and clock timing changes. I agree that the vendors customize the bios files they supply with their branded cards. I intimated that in my original post. The original base firmware with the core hardware interface code is still supplied by nVidia. The issue is common to all AIB cards using the 86.04.26 bios files/Mircon memory not a single vendor. None of the vendors will be able to issue an update that resolves the problem if the foundation code supplied by Nvidia is intrinsically flawed. The reasons for the change to Micron GDDR5 memory is irrelevant, nVidia has accepted that change by providing the different firmware base code to the vendors to support the different brand of memory. They need to create a fix a the bug in the base firmware that can then be modified by the partners for distribution to customers. That fix could be as simple as a document to each partner telling them to edit some hex in the base file all the way to sending them rewritten base code to work with. This is not a big deal, bug fixes are rolled out to solve problems or improve device performance in just about every piece of technology all the time. Problem is that nVidia cannot make running improvements to their product if they don't know that the problem exists. I don't understand the attitude exhibited by some members here who seem to want to live in denial and seem to want to blame everyone else. Nvidia is getting a reputation problem based mostly on ignorance that people have heard that the micron cards are crap, and the memory cant overclock, they are worried about the "memory lottery" all because of a really minor issue that could/should be dealt with in the normal course of business. The truth is that the Micron memory cards aren't crap, They work to spec and overclock fine if you know how to work around the bug in the firmware, but the bug can be fixed and all the "lottery" talk will go away! Anyway, why should we want facts to get in the way of a good emotional internet emotional melt down? [quote="Greybear"]. First I would like to say thank you for the hard work and thought out explanation of the issue with the Micron Memory. It is not easy to track these down. [quote][color="orange"]It is not an MSI specific problem[/color], other brands with the mircon memory and bios are exhibiting the same symptoms. [/quote] But, NVIDIA makes a REFERENCE chipset / card and the ""Partners"" are responsible for the actual components. The use of MICRON memory instead of the SAMSUNG is an attempt to make their cards cheaper and their profits higher. In doing so, the changes to the vBIOS should come from the ""Partners"" and not NVIDIA due to the fact that it was the ""Partners"" choice to use subpar hardware at the cost of the USERS experience. So, in this case, you really should be hitting MSI up for an updated vBIOS specific to the MICRON memory. Now I would like to point out why I believe you are looking for the VENDOR to supply the vBIOS. If it were coming DIRECTLY from NVIDIA, all cards would have the SAME vBIOS..... [color="orange"]NVIDIA[/color]: 86.04.1E.00.01 : Founders [color="orange"]<------[/color] [color="orange"]MSI:[/color] 86.04.1E.00.34 : Aero OC 86.04.1E.00.5E : Aero OC 86.04.1E.00.01 : Founders [color="orange"]<----- [/color] 86.04.1E.00.70 : Gaming 86.04.26.00.3F : Gaming 86.04.1E.00.40 : Gaming X 86.04.1E.00.41 : Gaming X 86.04.26.00.3E : Gaming X 86.04.1E.00.6C : Gaming Z 86.04.26.00.3B : Gaming Z 86.04.1E.00.55 : Seahawk [color="orange"]Gainward:[/color] 86.04.1E.00.2E : Pheonix [color="orange"]Gigabyte:[/color] 86.04.1E.00.01 : Founders [color="orange"]<-----[/color] 86.04.1E.00.68 : G1 Gaming 86.04.1E.00.78 : WF2OC [color="orange"]EVGA:[/color] 86.04.1E.00.70 : [color="green"]Founders[/color] / FTW / SC [color="green"]<---Where is the .01 vBIOS?!?!?![/color] 86.04.1E.01.70 : FTW 86.04.26.00.70 : FTW So, as you can see, NVIDIA made and released the FOUNDERS edition as the REFERENCE CARD with a REFERENCE vBIOS. After that, it is the ""Partners"" that make all MODIFICATIONS to the vBIOS. .[/quote]
You forgot to mention the micron bioses from vendors other than MSI such as:

EVGA 86.04.26.01.70
Zotac 86.04.26.00.4B
Gainward 86.04.26.00.27
Asus 86.04.26.00.61
KFA2 86.04.26.00.9C

So you are suggesting that EVGA, ASUS, MSI, Palit/Gainward, Zotac, Galaxy all independently wrote their own bioses from scratch? They all contain sections of code that are identical with only relatively minor changes for device IDs and clock timing changes.

I agree that the vendors customize the bios files they supply with their branded cards. I intimated that in my original post. The original base firmware with the core hardware interface code is still supplied by nVidia.

The issue is common to all AIB cards using the 86.04.26 bios files/Mircon memory not a single vendor. None of the vendors will be able to issue an update that resolves the problem if the foundation code supplied by Nvidia is intrinsically flawed.

The reasons for the change to Micron GDDR5 memory is irrelevant, nVidia has accepted that change by providing the different firmware base code to the vendors to support the different brand of memory. They need to create a fix a the bug in the base firmware that can then be modified by the partners for distribution to customers.

That fix could be as simple as a document to each partner telling them to edit some hex in the base file all the way to sending them rewritten base code to work with. This is not a big deal, bug fixes are rolled out to solve problems or improve device performance in just about every piece of technology all the time. Problem is that nVidia cannot make running improvements to their product if they don't know that the problem exists.


I don't understand the attitude exhibited by some members here who seem to want to live in denial and seem to want to blame everyone else. Nvidia is getting a reputation problem based mostly on ignorance that people have heard that the micron cards are crap, and the memory cant overclock, they are worried about the "memory lottery" all because of a really minor issue that could/should be dealt with in the normal course of business.

The truth is that the Micron memory cards aren't crap, They work to spec and overclock fine if you know how to work around the bug in the firmware, but the bug can be fixed and all the "lottery" talk will go away!

Anyway, why should we want facts to get in the way of a good emotional internet emotional melt down?







Greybear said:.

First I would like to say thank you for the hard work and thought out explanation of the issue with the Micron Memory.

It is not easy to track these down.

It is not an MSI specific problem, other brands with the mircon memory and bios are exhibiting the same symptoms.


But, NVIDIA makes a REFERENCE chipset / card and the ""Partners"" are responsible for the actual components.

The use of MICRON memory instead of the SAMSUNG is an attempt to make their cards cheaper and their profits higher.

In doing so, the changes to the vBIOS should come from the ""Partners"" and not NVIDIA due to the fact that it was the ""Partners"" choice to use subpar hardware at the cost of the USERS experience.

So, in this case, you really should be hitting MSI up for an updated vBIOS specific to the MICRON memory.


Now I would like to point out why I believe you are looking for the VENDOR to supply the vBIOS.


If it were coming DIRECTLY from NVIDIA, all cards would have the SAME vBIOS.....


NVIDIA:
86.04.1E.00.01 : Founders <------



MSI:
86.04.1E.00.34 : Aero OC
86.04.1E.00.5E : Aero OC
86.04.1E.00.01 : Founders <-----
86.04.1E.00.70 : Gaming
86.04.26.00.3F : Gaming
86.04.1E.00.40 : Gaming X
86.04.1E.00.41 : Gaming X
86.04.26.00.3E : Gaming X
86.04.1E.00.6C : Gaming Z
86.04.26.00.3B : Gaming Z
86.04.1E.00.55 : Seahawk

Gainward:
86.04.1E.00.2E : Pheonix

Gigabyte:
86.04.1E.00.01 : Founders <-----
86.04.1E.00.68 : G1 Gaming
86.04.1E.00.78 : WF2OC

EVGA:
86.04.1E.00.70 : Founders / FTW / SC <---Where is the .01 vBIOS?!?!?!
86.04.1E.01.70 : FTW
86.04.26.00.70 : FTW




So, as you can see, NVIDIA made and released the FOUNDERS edition as the REFERENCE CARD with a REFERENCE vBIOS.

After that, it is the ""Partners"" that make all MODIFICATIONS to the vBIOS.

.

#15
Posted 09/12/2016 07:01 AM   
  1 / 56    
Scroll To Top