• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

[WARNING] ASUS X299 TUF MK1 BIOS issues - DON'T UPDATE TO BIOS VERSION 1704

Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
If you use ASUS X299 motherboard (confirmed on my X299 TUF MK1) don't update to BIOS Version 1704. It disabled BIOS downgrade for me, which isnt much of an issue, but ASUS BIOSes have critical flaw.

I've found out that most BIOSes don't actually let you control voltage of your CPU. Early BIOSes would even report the voltage incorrectly, for example: if i set my CPU to 4.8Ghz and my Vcore to 1.130V it would boot, and be 100% stable. All monitoring software would show approx the same voltage I applied in BIOS. This was obviously false, my guess is that voltage was about 1.3V.
Later BIOS revisions fixed the incorrect Vcore readout(in a way that there is at least some control) but the actual voltage is still "unknown". There is only one BIOS which actually lets you control Vcore and that is Version 1503 and only in manual Vcore mode and disabled downclocking (sync all cores → [your desired CPU multiplier] → manual → [your manual voltage]).

In case you are already on BIOS 1704 and decide to downgrade, then you'll need to use ASUS BIOS Flashback (which can be kinda risky). It did work and my temperatures went back to "normal".
Feel free to test it for yourself if you have TUF X299 MK1: (Had to upload it to mega.nz because apperently 21MB is "too large") - contains now removed BIOS 1602, original 1503 and 1704.

I assume the situation on board from the same line-up will be identical (at least on X299-A Prime, TUF X299 MK2...). Feel free to test it for yourself, you might get some extra performance out of your CPU or cooler temps.

It would be great if mutliple people could confirm this, just to be sure, as one "test sample" isnt very scientific.

Note: Using Ai Suite overclocking makes the BIOS 1503 act like other BIOSes → BIOS reflash is required to get the voltage control back.


My results:
Version 1503
15XX.PNG


Version 1704 (the lower score is because i had steam running in background, as far as i can tell the BIOS versions dont effect score)
17XX.PNG


Both were ran at:
BCLK: 100.1
Multiplier: 47x (Sync All Cores)
Memory: 3600Mhz 18-19-19-34 T1 tRFC 279
Memory voltage: 1.25V
System agent voltage: 0.930V
CPU input voltage: 1.670V
CPU Current Capability: 140%
LLC: Level 2
Vcore: 1.205V
Uncore offset: +0.150V
Mesh multiplier: 32x
Mesh Voltage: 1.010V
Feel free to replicate my settings.
 
Joined
Feb 22, 2019
Messages
1 (0.00/day)
Hi there man,

Thanks for sharing this. I was thinking about upgrading my BIOS from the 1602 version to 1704 and happen to find your post on Google. I build my first PC with ROG STRIX X299-E GAMING motherboard and with Intel i9 7960x CPU. I'm a pretty noob to this stuff. Also, I contacted Asus Rog customer support and ask them about why they removed 1602 version from the x299-e gaming motherboard driver download page, they told me "BIOS 1602 version was removed because it was causing our customer's system to be unstable" but I did not have any problem with the BIOS 1602 version. My question to you is that, should I now upgrade to 1704 bios version from my old 1602 version?

Thank you, any help would be appreciated.
 

Attachments

  • my pc.PNG
    my pc.PNG
    8.6 KB · Views: 459
  • bios.PNG
    bios.PNG
    48.7 KB · Views: 683
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
Hi there man,

Thanks for sharing this. I was thinking about upgrading my BIOS from the 1602 version to 1704 and happen to find your post on Google. I build my first PC with ROG STRIX X299-E GAMING motherboard and with Intel i9 7960x CPU. I'm a pretty noob to this stuff. Also, I contacted Asus Rog customer support and ask them about why they removed 1602 version from the x299-e gaming motherboard driver download page, they told me "BIOS 1602 version was removed because it was causing our customer's system to be unstable" but I did not have any problem with the BIOS 1602 version. My question to you is that, should I now upgrade to 1704 bios version from my old 1602 version?

Thank you, any help would be appreciated.

Depends, if you arent pushing your chip to the max and you must have the latest ME update for some reason then you probably should flash to 1704, but you should understand that there's a quite bigger risk that you'll brick your motherboard if you then decide you want to downgrade since flashback is riskier than upgrade/downgrade. Also i should probably note that i dont know the actual voltage that the motherboard feeds to the CPU and its possible that its even above 1.3V, which will cause long term degradation. This glitch degraded my CPU from 5.1Ghz max to 4.975ish max after running on that unknown voltage for about 8 months. As far as i can tell all BIOSes prior to 1704 can upgrade or downgrade to any other BIOS.

The best thing you can do is run your prefered stress test, monitor the temps and then downgrade to 1503 to check if there is any improvement. It might be just TUF boards but all ASUS X299 seem to have identical VRM and BIOSes. If you are running conservative 4-4.2Ghz its pointless to blast your CPU with +-1.3V all the time.

*I'll try to check what voltage i need to reach similiar temperatures but its still just guessing.

EDIT: Tried couple runs at voltages of 1.290V, 1.300V, 1.310V. 1.310V was too high but both 1.300V or 1.290V seemed identical to what i get on other BIOSes. This is very weird as ASUS recommended YouTubers/Journalists to run their CPUs at 1.250V max at launch of the X299.
All the BIOSes are complete mess anyway, most of them run at >1.9V CPU Input voltage which is absolute overkill, 1.7V was enough for 5Ghz.
 
Last edited:
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
I've found out that most BIOSes don't actually let you control voltage of your CPU.
All ASUS boards do not have the same VRM and BIOS.

Auto voltage usually leaves a lot of meat on the bone and allows voltage to be lowered. This holds true for all boards Ive personally tested.
This is very weird as ASUS recommended YouTubers/Journalists to run their CPUs at 1.250V max at launch of the X299.
They did? I don't recall that.

Anyway, I have an Omega here for testing... its using 0402 with the latest 0504. So, that said, this isn't an "ALL ASUS X299" issue, first of all. Second, it may just be to that board.
 
Joined
Jun 5, 2017
Messages
13 (0.01/day)
Location
Romania
System Name Bruno's
Processor Ryzen 7 2700
Motherboard ASROck X470 Gaming K4
Cooling Noctua NH-U12S
Memory 32GB ADATA 3200 C16
Video Card(s) Galax GTX 1050 Ti EXOC White
Storage 480GB ADATA SX8200 M.2 NVMe + 2TB Seagate HDD
Display(s) 3x Dell 2412
Power Supply Seasonic 750W
Mouse MSI DS200
Keyboard Microsoft Natural Ergonomic 4000
Software Win 10 x64
This is very weird as ASUS recommended YouTubers/Journalists to run their CPUs at 1.250V max at launch of the X299.
That is not true! I've reviewed several Asus X299 boards and Asus hasn't told me that.
Yet, a value above 1.25V will get the CPU hot if cooling is not appropriate.

Also i should probably note that i dont know the actual voltage that the motherboard feeds to the CPU and its possible that its even above 1.3V, which will cause long term degradation. This glitch degraded my CPU from 5.1Ghz max to 4.975ish max after running on that unknown voltage for about 8 months.

You can read the vcore, vin, vccio, etc with HWInfo. It is pretty accurate and it also shows the min and max values so you know how you stand.

From my point of view, your problem may be form LLC value. On Asus boards, LLC 2 means under-volting under load, so 1.2V set in BIOS may went to 1.15V (i don't know exactly how much it over-volts, but use HWInfo to find out) when CPU has workload, but may come to 1.2V when idle, depending on power management settings.
So my advice is this: Use HWInfo and read idle and load voltages (under all BIOSes you need) using LLC 2.
Then try Level 5 and 6 and check the values for those.
Also check the CPU temps in each case.

Regarding CPU temps, in your pics there are cores that reach 100C, which is the reason for CPU degradation. If under CBR15, where load is not that high, your CPU gets to 100C, I don't want to know how it goes under X265 or XTU.


All the BIOSes are complete mess anyway, most of them run at >1.9V CPU Input voltage which is absolute overkill, 1.7V was enough for 5Ghz.
Input voltage is split between other voltages (like vcore, vccin, vram, etc). On auto, BIOS sets it to a value that covers system stability and that means fluctuating it up and down along with vcore, vram, vccin, etc. Those voltages depends on CPU quality so some CPUs need higher values. Manufacturers set VIN to 1.8V+ to be sure ANY system will run stable under load regarding CPU and RAM voltages.
Of course, any user can find the lowest VIN for its system, but that means under/over-clocking and there are not many to do it.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
All ASUS boards do not have the same VRM and BIOS.

Auto voltage usually leaves a lot of meat on the bone and allows voltage to be lowered. This holds true for all boards Ive personally tested.

They did? I don't recall that.

Anyway, I have an Omega here for testing... its using 0402 with the latest 0504. So, that said, this isn't an "ALL ASUS X299" issue, first of all. Second, it may just be to that board.

My apologies, i meant all the motherboards released at launch of X299. The lower-end ones seem to have identical VRM(easy to tell on X299-E and X299-A) and all of them get identical BIOSes at the same time.

As for safe voltage, i dont recall where i saw this figure as i was building my PC almost 2 years ago but im pretty sure it was 1.25V or 1.275V.



That is not true! I've reviewed several Asus X299 boards and Asus hasn't told me that.
Yet, a value above 1.25V will get the CPU hot if cooling is not appropriate.



You can read the vcore, vin, vccio, etc with HWInfo. It is pretty accurate and it also shows the min and max values so you know how you stand.

From my point of view, your problem may be form LLC value. On Asus boards, LLC 2 means under-volting under load, so 1.2V set in BIOS may went to 1.15V (i don't know exactly how much it over-volts, but use HWInfo to find out) when CPU has workload, but may come to 1.2V when idle, depending on power management settings.
So my advice is this: Use HWInfo and read idle and load voltages (under all BIOSes you need) using LLC 2.
Then try Level 5 and 6 and check the values for those.
Also check the CPU temps in each case.

Regarding CPU temps, in your pics there are cores that reach 100C, which is the reason for CPU degradation. If under CBR15, where load is not that high, your CPU gets to 100C, I don't want to know how it goes under X265 or XTU.



Input voltage is split between other voltages (like vcore, vccin, vram, etc). On auto, BIOS sets it to a value that covers system stability and that means fluctuating it up and down along with vcore, vram, vccin, etc. Those voltages depends on CPU quality so some CPUs need higher values. Manufacturers set VIN to 1.8V+ to be sure ANY system will run stable under load regarding CPU and RAM voltages.
Of course, any user can find the lowest VIN for its system, but that means under/over-clocking and there are not many to do it.

I do use HwInfo as well as AIDA64. On the early BIOSes they showed the exact number i typed in to BIOS later BIOSes just show basically random number that makes at least little sense. LLC doesn't change anything, voltage is constant without single milivolt dip at any level, which is suspicious in itself.

Highest temperature with glitched BIOS was 92°C for the whole CPU compared to 76°C on the semi-working BIOS. As for cooling im using 280mm rad, 360mm rad and thick 120mm rad, all of them in push pull and its not problem to run 4.95Ghz at mid 90's, which i obviosly don't do.
I know what CPU input voltage does, but I noted that since i think this is reason why there was the whole "X299 VRM Distaster". When running non-AVX Prime95 the temperature of VRM barely reaches 87°C in the hotspot in closed case with all intakes blocked by radiators.
These issues have been on at least my motherboard since release and it does seem to be entirely BIOS's fault not hardware fault.

I also used the identical settings across all the BIOSes, how do you explain the different results? Literally nothing changed except the BIOS and i get almost delid-like temperature drop? (note: my CPU isnt delided)
 
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
My apologies, i meant all the motherboards released at launch of X299. The lower-end ones seem to have identical VRM(easy to tell on X299-E and X299-A) and all of them get identical BIOSes at the same time.

As for safe voltage, i dont recall where i saw this figure as i was building my PC almost 2 years ago but im pretty sure it was 1.25V or 1.275V.
all asus x299 boards at launch did not have the same power bits. :)

Voltage goes over 1.25v when set on auto on most (all?) chips... I struggle t9 believe they said this as harmful...but like was said above, things get warm up there on these HCC chips.

Perhaps with the new bios you are hitting some power limit and throttling? Software voltage isnt exactly accurate at times, so that can play a role too.
 
Joined
Feb 2, 2015
Messages
2,707 (0.81/day)
Location
On The Highway To Hell \m/
Regarding CPU temps, in your pics there are cores that reach 100C, which is the reason for CPU degradation.
Wrong and wrong. But, since you've already been corrected on the first one, I'll just address the second one. TJ Max for his processor is 99°C. So 100°C is not even possible, for any extended amount of time anyway. It might hit 100°C for a split second before thermal protection kicks in. And anything below that, temperature wise, isn't going to hurt it at all(no matter how long it's run at whatever sub-100°C temperature). TJ Max = the maximum safe operating temperature for the CPU die. There's no need for internet myths when we have manufacturer specifications. ;)
 
Last edited:
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
all asus x299 boards at launch did not have the same power bits. :)

Voltage goes over 1.25v when set on auto on most (all?) chips... I struggle t9 believe they said this as harmful...but like was said above, things get warm up there on these HCC chips.

Perhaps with the new bios you are hitting some power limit and throttling? Software voltage isnt exactly accurate at times, so that can play a role too.

The thing is voltage behaves like on Auto on any settings(Manual, Offset, Adaptive... none of them do anything) in all BIOSes except for 1503. Im not trying to get a solution here as Im fully aware that there just is none other than fixing the BIOS which is in hands of ASUS. I just want to warn other users, there is nowhere mentioned that updating to 1704 will disable dowgrading. Im waiting for a fix for a long time, I didnt mind it being "broken" as long as ASUS is looking into it, but as of now they are silently removing BIOSes and in the new ones preventing users from downgrading so they can't even compare if they have this problem or not. There is no reason why user couldnt downgrade from 1704 to any previous one, as i did it and everything works 100% fine.
I sank a lot of hours figuring out how to actually get this to work as should and i simply wanna provide my knowledge to those who are stuck in the same situation.

As for "throttling", there is none, if there was throttling the results would differ, they dont differ outside of margin of error.

As of now it would be great if someone with similiar board could double check this, i absolutely dont mind sharing my testing methodology.
 
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
I'm sure you worked hard and such... but that doesn't mean something possibly couldn't have been missed. Just throwing things out there to add to the pile. :)

Did you disable SVID?


Anyway, no issues with the Omega on the newest BIOS. Able to change voltage and such without issue. I hope you find someone with the same board to test and confirm.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
I'm sure you worked hard and such... but that doesn't mean something possibly couldn't have been missed. :)

Did you disable SVID?
Yes, there are actually two SVID settings, so i disabled one in Ai Tweaker Tab and second one in Tweaker's Paradise Tab.

Edit: Im glad you are not having any issues with your motherboard, but as far as i can tell your motherboard was released with 9000 generation of X299 chips so there might be differences.
May i ask what CPU, CPU OC settings and CPU cooler are you using? (Your System Specs dont work for me for some reason)
 
Last edited:
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
I just realized I have a X299 TUF Mk1 in the closet... but, I don't know when I will have time to pop it on the bench and test it... I'm pretty behind on other review work. :(

From the review it had 0802 on it and there was voltage control there. I upgraded it from.................(looking..........)..........0402.

EDIT: Honestly, I don't think I can get to it as I need to break down my daily driver and drop that CPU in as the BIOS that is on there doesn't support the 9 series CPU...(unless it has CPUless flash capabilities which I do not recall off hand......)

EDIT2: My system specs I have disabled/not filled out. Currently on the omega is a 9980XE with an OC Cool Eisbar extreme. I overclocked it just a bit ago to 4 GHz all c/t 1.05V. Daily driver is a 7960X 16c/t (HT disabled) at 4.4 GHz 1.18V with 3x120mm cooler (MSI board though).


EDIT3: You may want to change the title to something a bit more applicable. At this time, we only know that it is your/the TUF Mk1 having this issue. Until someone proves otherwise, the thread title is pretty misleading. :)
 
Last edited:
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
I just realized I have a X299 TUF Mk1 in the closet... but, I don't know when I will have time to pop it on the bench and test it... I'm pretty behind on other review work. :(

From the review it had 0802 on it and there was voltage control there. I upgraded it from.................(looking..........)..........0402.

EDIT: Honestly, I don't think I can get to it as I need to break down my daily driver and drop that CPU in as the BIOS that is on there doesn't support the 9 series CPU...(unless it has CPUless flash capabilities which I do not recall off hand......)

EDIT2: My system specs I have disabled/not filled out. Currently on the omega is a 9980XE with an OC Cool Eisbar extreme. I overclocked it just a bit ago to 4 GHz all c/t 1.05V. Daily driver is a 7960X 16c/t (HT disabled) at 4.4 GHz 1.18V with 3x120mm cooler (MSI board though).


EDIT3: You may want to change the title to something a bit more applicable. At this time, we only know that it is your/the TUF Mk1 having this issue. Until someone proves otherwise, the thread title is pretty misleading. :)

TUF MK.1 should be able to be flashed without CPU. Just use BIOS Flashback. (Source)

The reason why i used cinebench is that the run will complete even if the CPU is slightly unstable in lets say Prime95. That way i found out if the voltage control actually works or not. What i did was i tried lowering the cpu voltage by just a little bit (0.010) and kept my eyes on temperatures. With my initial BIOS i kept going down at 4.7Ghz all the way to 1.080V which then just refused to boot but the temperatures were the same all the way from 1.220V.
It would be great if you could try this test too, but as already noted, your newly released board might not suffer from this.

I would absolutely change the title, but i honestly have no clue what to and i dont wanna make it too long.
 
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
Just add the name of the board it is affecting.
[WARNING] ASUS X299 TUF MK1 BIOS issues w/VERSION 1704

Anyway, Cinebench runs for like 5-8 seconds. Voltage moves on my boards, I've seen it happen, there isn't a need to try it in Cinebench to confirm anything. I would prefer to check something like that on anything that runs longer...but as I said, I saw the changes... saw the results in power output as well. Clearly it doesn't affect this board. NONE of the ASUS boards I reviewed for X299 exhibited this behavior, but it was on older BIOSs as I reviewed these in 2017 through mid 2018.

0.01V won't budge temps... besides you are reading software and that can be inaccurate. Your quick blip and minor changes, seemingly aren't really telling you the information needed to come to your conclusion. You really need to get at the voltage with a DMM if possible and confirm that way. You are likely right, but more so by accident than data due to the testing methods.

Maybe next week I can dig the board up and flash it forward to the latest and see what is going on.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
Just add the name of the board it is affecting.
[WARNING] ASUS X299 TUF MK1 BIOS issues w/VERSION 1704

Anyway, Cinebench runs for like 5-8 seconds. Voltage moves on my boards, I've seen it happen, there isn't a need to try it in Cinebench to confirm anything. I would prefer to check something like that on anything that runs longer...but as I said, I saw the changes... saw the results in power output as well. Clearly it doesn't affect this board. NONE of the ASUS boards I reviewed for X299 exhibited this behavior, but it was on older BIOSs as I reviewed these in 2017 through mid 2018.

0.01V won't budge temps... besides you are reading software and that can be inaccurate. Your quick blip and minor changes, seemingly aren't really telling you the information needed to come to your conclusion. You really need to get at the voltage with a DMM if possible and confirm that way. You are likely right, but more so by accident than data due to the testing methods.

Maybe next week I can dig the board up and flash it forward to the latest and see what is going on.

This title would be kind of misleading too, as these issues are on all BIOSes, i just got fed up with ASUS for blocking downgrade on 1704.

This is voltage readout on 1503, it should be the same on all bioses from 802 above.
Note: AIDA64 reads CPU Core voltage as 1/2 of CPU input voltage and CPU VID as the actual CPU voltage.
vDroop.PNG


Zero Vdroop on Vcore, small Vdroop on CPU Input. And I can confirm that even 0.010V drops temperature by couple degrees (1-3°C).

EDIT: to be clear, im just trying to say that the motherboard reports for most readouts what you set in the BIOS and not the real-time value.
 
Last edited:
Joined
Apr 15, 2009
Messages
1,007 (0.18/day)
Processor Ryzen 9 5900X
Motherboard Gigabyte X570 Aorus Master
Cooling ARCTIC Liquid Freezer II 360 A-RGB
Memory 32 GB Ballistix Elite DDR4-3600 CL16
Video Card(s) XFX 6800 XT Speedster Merc 319 Black
Storage Sabrent Rocket NVMe 4.0 1TB
Display(s) LG 27GL850B x 2 / ASUS MG278Q
Case be quiet! Silent Base 802
Audio Device(s) Sound Blaster AE-7 / Sennheiser HD 660S
Power Supply Seasonic Prime 750W Titanium
Software Windows 11 Pro 64
you should understand that there's a quite bigger risk that you'll brick your motherboard if you then decide you want to downgrade since flashback is riskier than upgrade/downgrade.
I'm curious what your basis is for this statement.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
I'm curious what your basis is for this statement.
Sure, here:
https://rog.asus.com/forum/showthread.php?102234-Maximus-X-Hero-Wifi-AC-BIOS-Flashback-not-working
https://hardforum.com/threads/usb-bios-flashback-not-working.1922049/

Back when X99 was a thing, i remember someone got their motherboard bricked after using BIOS flash back, and when he contacted MSI support they told him it might not work all the time and that its experimental. (im fully aware this is MSI board and not ASUS, but still...)
↑ cant find source for this tho
 
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
This title would be kind of misleading too, as these issues are on all BIOSes
There is? The TUF Mk1 does...maybe... but what about the rest? You're making a lot of assumptions and applying it across all x299 Asus boards.

This proves nothing. There is a risk with any bios flash. I dont see where there is more or less risk from those links. Bad flashes happen man... in either direction.

What does msi x99 have to do with Asus x299???

(more assumptions...)
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
There is? The TUF Mk1 does...maybe... but what about the rest? You're making a lot of assumptions and applying it across all x299 Asus boards.

This proves nothing. There is a risk with any bios flash. I dont see where there is more or less risk from those links. Bad flashes happen man... in either direction.

What does msi x99 have to do with Asus x299???

(more assumptions...)
My explanation was poor, i admit that, but if you actually used flashback, its very unclear whats actually going on at given moment. When you update your bios from bios you get a progress bar and at least some information. With flashback you get one blinking LED, ASUS isnt very clear either about explaining whats going on. During my flashbacks I waited 20ish minutes after i assumed it was done just to be sure.

This can absolutely confuse normal user and make him turn off his PC and corrupt his BIOS.
 
Joined
Dec 31, 2009
Messages
19,366 (3.72/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
My explanation was poor, i admit that, but if you actually used flashback, its very unclear whats actually going on at given moment. When you update your bios from bios you get a progress bar and at least some information. With flashback you get one blinking LED, ASUS isnt very clear either about explaining whats going on. During my flashbacks I waited 20ish minutes after i assumed it was done just to be sure.

This can absolutely confuse normal user and make him turn off his PC and corrupt his BIOS.
It isnt exactly a feature for the common folk, the flashback, honestly... but yeah, that is low level flashing without a CPU. That little chip doesnt have the horsepower to do anything but. Most users use the ezflash in the bios which does have graphical interface/display progress.

All we know now is your board is sort of borked and MAYBE this issue is for your board... Novody knows if others are affected (hence why the thread title needs to be changed). The rest are assumptions that arent based off anything concrete.

Anyway, I'll see what I can do during the week - hopefully I get some time. I'm just going to flash to the latest bios and see if I can change voltage. Being able to roll back doesnt really concern me much as there always seems to be a stopping point for that kind of thing IIRC.

Have you tried a normal reflash of the latest bios?
 
Joined
Jun 5, 2017
Messages
13 (0.01/day)
Location
Romania
System Name Bruno's
Processor Ryzen 7 2700
Motherboard ASROck X470 Gaming K4
Cooling Noctua NH-U12S
Memory 32GB ADATA 3200 C16
Video Card(s) Galax GTX 1050 Ti EXOC White
Storage 480GB ADATA SX8200 M.2 NVMe + 2TB Seagate HDD
Display(s) 3x Dell 2412
Power Supply Seasonic 750W
Mouse MSI DS200
Keyboard Microsoft Natural Ergonomic 4000
Software Win 10 x64
Wrong and wrong. But, since you've already been corrected on the first one, I'll just address the second one. TJ Max for his processor is 99°C. So 100°C is not even possible, for any extended amount of time anyway. It might hit 100°C for a split second before thermal protection kicks in. And anything below that, temperature wise, isn't going to hurt it at all(no matter how long it's run at whatever sub-100°C temperature). TJ Max = the maximum safe operating temperature for the CPU die. There's no need for internet myths when we have manufacturer specifications. ;)
99C is the limit of Intel, but many boards usually have the setting on 100-105C, and it can also be set manually, so CPU throttling starts there, not on 99C. Or it may not start at all, if power management settings are disabled.
CPU degradation is not an Internet myth, especially on air/water cooling. Even though it doesn't happen overnight, a constant exposure to high temps can induce degradation in time. It may appear after one year and just increase the VCore with 0.01V, but it is still degradation.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
It isnt exactly a feature for the common folk, the flashback, honestly... but yeah, that is low level flashing without a CPU. That little chip doesnt have the horsepower to do anything but. Most users use the ezflash in the bios which does have graphical interface/display progress.

All we know now is your board is sort of borked and MAYBE this issue is for your board... Novody knows if others are affected (hence why the thread title needs to be changed). The rest are assumptions that arent based off anything concrete.

Anyway, I'll see what I can do during the week - hopefully I get some time. I'm just going to flash to the latest bios and see if I can change voltage. Being able to roll back doesnt really concern me much as there always seems to be a stopping point for that kind of thing IIRC.

Have you tried a normal reflash of the latest bios?
Flashed back 1704 via EZFlash in BIOS and the results were identical. Now that im back to 1503 i can that it effects even the idle temps (1503: 26-35°C; 1704: 33-48°C). Used identical BIOS settings for both.
Also there is one small issue, if you plan on using 9980XE on BIOS 1503 it wont work according to this, but maybe you have some sort of a different plan. I'll just list the settings i change in order to get the temperature drop so if anyone wants to test it out...

Ai Overclock Tuner: Manual (not XMP, set your timings, clocks, voltage manually)
BCLK Frequency: 100.1
CPU Core Ratio: Sync all cores
CPU SVID: Disabled
CPU LLC: Level 2
CPU Current Capability: 140%
BCLK Aware Adaptive Voltage: Disabled (this option isnt present on BIOS 1704)
CPU Input Voltage: 1.670*
System Agent voltage: 0.930*
Uncore Voltage Offset: +0.150*
CPU Core voltage & CPU cache Voltage: Manual


*These could be different between CPUs.

I obviously excluded cache and core clocks & voltages as they rely on CPU itself.
Rest left on Auto or "stock"/unchanged from default.
 
Joined
May 20, 2019
Messages
11 (0.01/day)
System Name Main PC
Processor Intel Core i7 7820X
Motherboard Asus Rog Strix X299-E Gaming
Cooling Corsair Hydro Series H115i
Memory G.Skill Trident Z 32 Go (4x 8 Go) DDR4 3200 MHz CL15
Video Card(s) ZOTAC GeForce GTX 1080 Ti AMP Extreme Core Edition
Storage Samsung 850 Pro 1 To + Western Digital WD Red 4 To SATA 6Gb/s
Display(s) ASUS 27" LED - ROG Swift PG279Q
Case Corsair Obsidian 750D Airflow
Audio Device(s) Asus Xonar DX
Power Supply Corsair RM 1000i
Mouse ROCCAT Kone AIMO
Keyboard Corsair K70 Lux RGB
Software Windows 10 Pro 64
Benchmark Scores ?
Hi,
I've found this topic trying to find some help.
For some reasons I decided to upgrade my Bios version from 0802 to 1704 version, despite the fact that my PC worked perfectly well.
Before updating the bios, I downloaded and installed all the latest versions of the drivers and firmware of my chipset (ITBM Version 1.1.0.1005; Intel Chipset Driver Version 10.1.1.45; Intel ME Version 11.11.65.1590, and so on...).
All the installations went smoothly, I restarted and used the PC for a few hours following these installations, and everything was working normally. So I think my problem is not related to these updates.

Then I updated my bios (from the bios using the tools Asus EZ Flash 3). Before that I renamed the bios file with the program attached in the archive (ROG-STRIX-X299-E-GAMING-ASUS-1704.CAP => X299SE.CAP).
The update went normally, in the end I got the message that the bios had been updated successfully.

Note that in the bios I'm just loading the XMP profile and some minor changes to the operation of the fans.
My CPU is not overclocked because my knowledge does not allow me to do this, and unfortunately I do not have enough time to learn.
So it works at the basic frequencies set by default in the motherboard.
Before booting the PC after the bios update, I went in the Bios settings and I put everything as before (AI Overclock Tuner : XMP instead of Auto; Q Fan control : DC Mode instead of Auto).

After this, impossible to start Windows, I had many blue screen during Windows start up saying that my PC have a problem and needs to restart with such QR Code :
- CRITICAL_PROCESS_DIED
- IRQL_NOT_LESS_OR_EQUAL
- And many others like this.

Windows has made multiple attempts to repair, without success.
Then I tried to reset all Bios settings to default, then to start the system again.
It started, but it seems that the system is unstable and some software does not work properly.

I really don't know what to do now.
I would prefer to return to the old version that worked well.
For Windows, I can restore it from a complete Acronis image of my partitions, which is only 15 days old.
But I would also like to go back to the version 0802 of the bios because I had no problem with it.
Is it possible to go back to bios 0802 and how ?
I tried from the bios with ASUS EZ Flash 3, but it always tells me when reading the bios on my USB drive, that it's not a proper Bios. I tried renaming the .CAP file and without renaming it, I have the same message.
For this risky operation, is the only way to use the USB BIOS FLASBACK button ?
If so, should I rename the .CAP file from bios 0802 to X299SE.CAP or not ?
Otherwise do you think I have to stay in version 1704, but in this case I do not know how to set my bios to make the system working, and I'm afraid of having to reinstall my entire system because the installation seems damaged (it would be weeks of work and I would like to avoid).
Thanks for your help.
 
Joined
Apr 1, 2018
Messages
58 (0.03/day)
Location
Czech Republic
System Name Main | Second PC
Processor 12900K (P-only 5.1Ghz static)| 5800X (4.625Ghz Static)
Motherboard ASUS Z690-F | X570 Unify
Cooling 420mm + 120mm + D5 | NH-D15S
Memory DDR5 2x32GB 6200Mhz CL30| 4x8GB 3800Mhz CL16 1:1
Video Card(s) EVGA RTX 3080 Ti FTW3 (450W/1000W)| MSI RTX 3060 (190W)
Storage 2x 980 Pro 2TB | Micron 3400 1TB OEM
Display(s) ASUS PG32UQ (32" 4K 155Hz) + Samsung G7 27" | Acer Nitro XV273K (4K 144Hz 10bit)
Case Corsair 7000D Airflow| Meshify 2 XL
Audio Device(s) Custom HD600 (4.4mm BAL Pentacon, sheepskin pads) + Fiio Q3 | Audio-Technica MSR7b + FiiO Q1 Mk.II
Power Supply Corsair AX1600i | SilverStone SX1000 Platinum
Mouse Logitech G303 SE | Razer Viper Ultimate
Keyboard Razer Huntsman | Custom w/ thick white PBT keycaps, Kaihl Copper & BOX Black switches
Found a fix for the unreasonably high temperatures:

1. Uninstall AI Suite 3 (Thermal Radar 3)
2. Stop all these services & prevent them from starting ever again (they will remain there even after uninstall):
123486

3. Reboot
4. Open Task Scheduler
5. Block all events in ASUS folder
6. Reflash your BIOS

Results in consistent temperature but 1704 still causes +-5°C higher temps in all scenerios(Could be hidden Offset in 1704 as TjMax in 1503 is 105°C while in 1704 is 110°C).

This fix has one disadvantage, which is being able to change fan curves only in BIOS.

- CRITICAL_PROCESS_DIED
- IRQL_NOT_LESS_OR_EQUAL
I get these if my IMC or RAM isnt stable.

The best thing you could do is downgrade to 1503 and if you still get crashes increase System Agent Voltage to 0.975V (I run mine at 0,930V for 3733Mhz).


I tried from the bios with ASUS EZ Flash 3, but it always tells me when reading the bios on my USB drive, that it's not a proper Bios. I tried renaming the .CAP file and without renaming it, I have the same message.

It says this even if the BIOS is completely fine, i used Flashback more than 10 times and it went well every time, i just recommend not doing anything with the PC as long as the LED light is blinking, ASUS manuals have some contradictory statements.


For this risky operation, is the only way to use the USB BIOS FLASBACK button ?
If so, should I rename the .CAP file from bios 0802 to X299SE.CAP or not ?

Yes Flashback is the only way.
The name of the file should be X299SE and the extention .CAP, make sure its not X299SE.CAP.CAP.

This is how my Flashback drive is configured
123492


Otherwise do you think I have to stay in version 1704
I really wouldnt stay in 1704, doesn't have any benefits over 1503, and i dont think OS re-install would help with your issue.
You could also try to not use XMP to see if its stable, but you will take huge performance hit.
 
Last edited:
Top