ROG MAXIMUS X HERO BIOS VER 1101 Feedback Thread

[mr_allroy]

I just noticed this is on the ASUS US website. Has anyone tried it yet? I'm not feeling like being a guinea pig as "bios flashback" does not seem to work.

Anyways, let's keep this friendly and post our experiences with it.

[izakaya]

KeksimusMaximus wrote:
Anybody uses NVMe SSD (Samsung EVO 960) and went to bios 1101?

I just noticed massive write speed/IOPS performance loss, like 4 times drop. From previous 1800MB/s now i barely get 500MB/s. The random IOPS dropped from 200,000 to... 20,000. Tested both with Samsung Magician and Cristal Disk Mark.

And yes, its running on Samsung drivers and M2 is set to PCIE 3 x4.

Ouch! 😕 I have the same NVMe and have been waiting for more info on 1101.
Ehm... seems like I'll be keeping the 1003 for a while.

I'm sorry to hear about your issue and hope there's a simple solution to it :/*

[KeksimusMaximus]

izakaya wrote:
Ouch! 😕 I have the same NVMe and have been waiting for some more info on the 1101, before updating from 1003.

Since all BIOS settings will have been reset, perhaps check:

1) Is the NVMe set to run as*PCIe?

2) Is the NVMe set to use x4?

Does Samsung Magician say: "Gen 4x3 in the main screen?

Yes to all

[izakaya]

KeksimusMaximus wrote:
Yes to all

Yeah, sorry, I read your post right after replying, and edited out my questions. I got interrupted for a while after I had read your post, and I simply forgot that you had already mentioned everything 😕

But yeah, it does seem extremely odd that the performance would simply plummet like that.
A performance loss of say 5% seems to be quite common when updating the BIOS these days, due to the Intel fixes etc, but the numbers you've mentioned are just silly 😕

If you had the "correct" numbers before the BIOS update, then I'm fairly certain that you've had the latest firmware and driver versions before the update. So what on earth could have caused something like that?
Assuming that you've had the latest Everything, and it all was working fine before flashing the BIOS, then the issue probably lies in the flashing BIOS part 😕

1) Perhaps try and flash the BIOS again, but using a different method this time? (something might have gone wrong somewhere)

2) Have you tried using the other M.2 slot to see which speed readings you get?

3) Is the Samsung 960 Evo the only storage drive you have installed?

4) Do you have anything connected to any SATA port on the motherboard?

[Tygerion]

KeksimusMaximus wrote:
Anybody uses NVMe SSD (Samsung EVO 960) and went to bios 1101? I just noticed massive write speed/IOPS performance loss, like 4 times drop. From previous 1800MB/s now i barely get 500MB/s. The random IOPS dropped from 200,000 to... 20,000. Tested both with Samsung Magician and Cristal Disk Mark. And yes, its running on Samsung drivers and M2 is set to PCIE 3 x4.

No problem here, SSD is as fast as ever working within specs after upgrading from 1003 to 1101 (960 EVO/250Gb, M.2_1, PCIe, Samsung drivers, measured with Crystal DiskMark 6.0.0).

[KeksimusMaximus]

Tygerion wrote:
No problem here, SSD is as fast as ever working within specs after upgrading from 1003 to 1101 (960 EVO/250Gb, M.2_1, PCIe, Samsung drivers, measured with Crystal DiskMark 6.0.0).

Yeah, something else is messed up in my OS, these results are pathetic:

-----------------------------------------------------------------------
CrystalDiskMark 6.0.0 x64 (C) 2007-2017 hiyohiyo
                               Crystal Dew World : https://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

   Sequential Read (Q= 32,T= 1) :  3306.939 MB/s
  Sequential Write (Q= 32,T= 1) :   467.836 MB/s
  Random Read 4KiB (Q=  8,T= 😎 :  1037.660 MB/s [ 253335.0 IOPS]
 Random Write 4KiB (Q=  8,T= 😎 :    84.866 MB/s [  20719.2 IOPS]
  Random Read 4KiB (Q= 32,T= 1) :   438.223 MB/s [ 106988.0 IOPS]
 Random Write 4KiB (Q= 32,T= 1) :    48.737 MB/s [  11898.7 IOPS]
  Random Read 4KiB (Q=  1,T= 1) :    43.443 MB/s [  10606.2 IOPS]
 Random Write 4KiB (Q=  1,T= 1) :     2.029 MB/s [    495.4 IOPS]

  Test : 1024 MiB [D: 50.1% (150.4/300.0 GiB)] (x3)  [Interval=5 sec]
  Date : 2018/04/04 20:33:13
    OS : Windows 10 Professional [10.0 Build 16299] (x64)

Like, when i boot Windows from EVO the write performance goes to hell, as you can see above, the write speeds are worse than regular SATA SSD and even worse than some HDD (random write 4KiB)

But when i boot secondary Windows installation from separate drive (SATA SSD Crucial MX100) and then EVO is a "data drive" then when i benchmark it the speeds are all normal (1800MB/s). The most ridiculous thing is that it will reach its full speed when being non-OS drive even with Microsoft drivers.

[SMB_92]

I have the Apex, and I found that my LLC setting of 5 seems to wildly overshoot now, I am set at 4.8ghz 1.24v across the board, was seeing max of 1.28v rarely at first load, but now its hitting into 1.33v territory.

Did LLC change from 1003?

[Shadowdane]

SMB_92 wrote:
I have the Apex, and I found that my LLC setting of 5 seems to wildly overshoot now, I am set at 4.8ghz 1.24v across the board, was seeing max of 1.28v rarely at first load, but now its hitting into 1.33v territory.

Did LLC change from 1003?

Yah Adaptive Voltage seems to be completely broken on now on 1101. Have a X Hero AC board and had been running at 4.8Ghz @ 1.22v since I got this board at launch. Now after upgrading the BIOS it seems to completely ignore the Adaptive voltage setting I set.

4.8Ghz
1.22v - Adaptive Voltage (0 Offset)

Doesn't matter what LLC setting I use.. Auto, Default, 1 or higher. Now it jumps to ~1.28v no matter the setting!
This is of course stable but temps are now hotter compared to the previous BIOS.

Hopefully Asus can get the Adaptive voltage fixed.. not sure why it's overshooting soo much.

[SMB_92]

Well I had this issue originally too (when i first setup the motherboard as new) , and to bring it back in check I had to enabled the SVID setting. So it went 1.MCE enabled, 2. SVID Behavior Best Case scenario and 3. The other SVID setting (can't recall at the minute) which was auto to enabled.

But that is not working now. And besides, the above is a real pita, seems rather, say, "unconventional" just to get the voltage range in order.

Edit: dropped it to level 4, still pretty high. My actual setting is 1.232, still boosting to 1.29 at level 4, and I can see VID is requesting up to 1.36. Perhaps the new rules from 802 and 1003 are not present in this version?

[bass_junkie_xl]

asus x hero wifi here

8700k delided and rockit 88 copper ihs liquid metal between heat sync as well

been using 5 ghz no avx offset

adaptive mode
totaly adaptive voltage - 1.36 v
offset - none ( defult )
svid-best case cinario
svid - on
ia/dc load line - 0.001
load line cali= 6
booting windows i get 1.36v - 1.392 v .....

bf1 i get 1.36v but soon as i get a heavy load over 50 % cpu ussage i get spikes to 1.392 v more 1.372v but spikes to 1.392......

i can counter the spikes by

1.36v totaly adaptive volts
offset negative -
use a - 0.016 v offset gives me

1.344 v on light loads goes to 1.36v 50 % of the time and under heavy loads like bf1 or bf4 spikes to 1.372v witch is perfect but i get 1.344 v 40 % of the time .

if i use 1.344v and run asus real bench i fail after 3 hrs
if u i use 1.36v manual asus real bench passes 8 hrs

so using the - 0.016 v offset is work around and i get 1.372v on loads wich is good but im on the edge of instability if it wants to stay @ 1.344v on load .

if i drop load line cali to 5 instead of 6 i get 1.372 v + on light loads and booting windows but i get 1.31 - 1.324 v on heavy loads .

ugh ...

i cant recall bios 1003 doing the over shoot spikes or not .

but the -0.016v off set with 1.360v works so far just wish intels voltege table in offset mode didnt use 0.016 v increments it kinda sucks i wish it was more 0.005 or 0.0010 increments not 0.016v is there anyway to make adaptive mode voltage work like manual vcore ? for example if i use

1.360v on manual vcore i get 1.360v 95 % of the time load line level 6
wheres
adaptive 1.360v gets load line 6

1.344v , 1.360v, 1.372v 1.392v range isnt this the behavior of offset mode ? seems like adaptive isnt doing what its suposed to imo .

its okay if adaptive mode uses the 0.016v increments but only going to your set max typed in totaly adaptive voltage ( 1.36v with out going over ) why doesn't it work like that ? adaptive is kinda borked atm ( kind of )

can anyone add to this ?

[Carlyle2020]

hi bass junkie xl,

i can confirm higher loads using the same Bios config on both Versions:
i run:

asus x hero wifi

8700k lidded and a Noctua NH-D14 + NT-H1 paste

been using 5 ghz no avx offset

adaptive mode
totaly adaptive voltage - 1.305 v
offset - none ( default )
svid-best case cenario
svid - on
ia/dc load line - 0.001
load line cali= 4
on full load i get :

CPU core:1.312
CPU Vid:1.354 (With Bios ver. 1003 the power capped almost exactly at 1.300)