Crosshair VI Hero: UEFI build update thread

[Raja]

Directly from Elmor:

Beta BIOS 3008 for C6H/C6HWIFI/C6E:

AGESA 1.0.7.1, temperature offset fixed after S3 resume, GPU Post function fixed, 0d with some GPUs fixed

http://www.mediafire.com/file/f95motjmh211e7h/CROSSHAIR-VI-HERO-ASUS-3008.zip
http://www.mediafire.com/file/ntn6i3jiub610ai/ROG-CROSSHAIR-VI-HERO-WIFI-AC-ASUS-3008.zip
http://www.mediafire.com/file/yj22ld2rmmedp01/ROG-CROSSHAIR-VI-EXTREME-ASUS-3008.zip





C6H Test BIOS 0020 with AGESA 1007 (405200c4e299c1eed9a0044edec9aba51f37cee1d70caabe40b9485b0604521a)

In order to get back to an older version after flashing, you have to use USB BIOS Flashback.

Let me know how it works for you, especially regarding cold boot issues. Any confirmed bugs in the bug report form please




Beta BIOS 1501

* Workaround for some CPUs stuck at 22x ratio if using override voltage
* Fixes PCIEX4 bandwidth setting getting lost after power is removed
* Same DRAM boot behavior as 1403 (no cold boot fix)
* Same DRAM profiles as in 9920
* Still has the Vcore value issue when booting with Offset Mode and switching to Manual Mode

Crosshair VI Hero 1501 SHA256 EDE223DC6897B7199C93D9985E28B7A2CD1B8A8DB2DCBF3D3555A521DB4F045D
Crosshair VI Hero Wifi AC 1501 SHA256 0D9F51F43AA3A56A4AC984B11A52F58451B76F8A7CCB9A04E1C3194231C9D4DA






UEFI build 9920 for the C6H:

* Improved DRAM cold boot, results in slightly longer POST time
* Fix for CPU Ratio stuck at 22x on some CPUs when using Vcore override/offset
* SenseMi Skew is now Disabled by default. If you want to return to previous behavior set SenseMi Skew = Enabled and Offset = 272.
* Added DRAM profiles for Samsung B-based DIMMs with tuned subtimings, including The Stilt's settings



C6H beta UEFI 1403

* Fixed W_PUMP and AIO_PUMP speeds during POST
* Fixed Fan tuning sometimes failing
* Fixed a few issues with AMD USB3.1 ports
* Some tuning on DRAM settings, let us know how they work for you. tRDRD_Sc is still at 5 above 3500 MHz as it helps with stability. For performance you want to force this to 1. We'll consider changing this in future releases as the performance impact can be quite noticeable in certain applications.

An update on DRAM Boot Voltage, currently it should be 1.35V by default if the DRAM Voltage is changed. So if you're setting DRAM Voltage to above 1.35V, you might want to sync this setting. Additionally there might be scenarios where you will have better luck by syncing DRAM Boot Voltage to DRAM Voltage even at lower values.




AGESA 1006 RC4 official UEFI 1401

Just tested quickly 3600 memory and cold booting, seems good but you guys are going to have to help me test this before we have a judgement. Seems CPU temp reading from SIO now has -20*C offset for XFR enabled CPUs.





UEFI build 1107





Older test builds:


Test BIOS 0079

Test BIOS with new AGESA 1004a, with a couple of bugfixes. Up to 5% performance increases in specific applications. Also has P-state overclocking working with BCLK increase. If you want to keep C-states, make sure to set Advanced\AMD CBS\Zen Common Options\Global C-state Control = Enabled. There are two new settings under AMD CBS\UMC Common Options\DDR4 Common Options\ you might want to take a look at, Fail_CNT and ProcODT. Fail_CNT decides how many times to retry when DRAM training fails (F9 -> 0d), default is now 1. ProcODT can help improve your DRAM overclocking. There's a setting available also on previous BIOSes under AMD CBS\UMC Common Options\DRAM Memory Mapping named BankGroupSwap. If you have 2x Single-Rank modules you can try setting this to Disabled and you might see some performance boost in certain applications.

Test BIOS 0081

Same as 0079 but has ProcODT default = 53.3 ohm

Test BIOS 0082

Same as 0081 but with a DRAM compability patch for the below part numbers

CMK16GX4M4A2133C13
AHU08GGB13CGT7G
PV416G240C5QK
F4-2400C15Q-32GRR
TCD44G2400C14BK
F4-2666C16Q2-64GRB
AVD4UZ126661504G
BLT8G4D26AFTA.16FAD
IMA41GU6MFR8N-C F0
MD16GK4D4266615AXR
HX430C15PB2K4/16
HX430C15PBK4/32
AX4U3333W4G16-QGZ
GAM4DBLBM2133D15IE041C
TC48G24S817
SP004GBLFU213N02
78.C1GM3.AF10B
F4-3200C16D-8GVKB

MTA4ATF51264AZ-2G6B1
MTA8ATF1G64AZ-2G3B1
MTA16ATF2G64AZ-2G6B1
HMA41GU7AFR8N-TF T0
HMA451U7AFR8N-TF T0
HMA41GU6AFR8N-UH N0
M378A2K43BB1-CTD
M378A1K43BB1-CRC
M378A1G43EB1-CRC

Test BIOS 0083

Same as 0081 but with "2T" DRAM Mode when using above 2666 DRAM Ratio.

[Clouseau]

looncraz wrote:
The 'subjective' feel would almost certainly be a result of lower latency allowing updates to occur on screen just a little faster, but allowing the update to occur on an entirely earlier frame (modern OSes use compositing, so they act like games in their rendering behavior).

If you are running 60hz on the desktop, then you are moving from having completion on-screen by 16.7ms - which you can absolutely feel. If you took 16.8ms to compute something before, you wouldn't see it for 33.4ms... but, now, computing it in 16.6ms, you see it twice as quickly.

The old way of rendering (direct) allowed you to feel a 10% performance increase in everything you did. But, now, you need BIG changes to feel it everywhere... but, once you get close, it only takes a tiny bit extra to put you over the top - the changes are accumulative.

Using 120hz refresh rate on the desktop allows you to feel the performance jumps even better, but you still have 8ms chunks of time.

Thank you for the information. Will try to keep it in mind.

[sbakic]

Guys I have one more question.

On BIOS 1701, with default and overclocked cpu/ram, fast boot and csm off, when I turn on my pc from 'shut down' state and watch q-codes on motherboard there is no repetition between codes. And colors goes as: orange, red, white, green.

But with 3008 and 3502 BIOS only when I overclock cpu/ram and turn on pc, I see that it repeat some codes. Color goes as: orange, red, orange red, white green. When it come to 3b q-code (red color), it start from start and after that normal sequence.

Is it a bug or it is supposed to be like that? It even happens in 3008 , 3502 when I change turn off fast boot and csm off, so I dont need to overclock it.

[Clouseau]

sbakic wrote:
Guys I have one more question.

On BIOS 1701, with default and overclocked cpu/ram, fast boot and csm off, when I turn on my pc from 'shut down' state and watch q-codes on motherboard there is no repetition between codes. And colors goes as: orange, red, white, green.

But with 3008 and 3502 BIOS only when I overclock cpu/ram and turn on pc, I see that it repeat some codes. Color goes as: orange, red, orange red, white green. When it come to 3b q-code (red color), it start from start and after that normal sequence.

Is it a bug or it is supposed to be like that? It even happens in 3008 , 3502 when I change turn off fast boot and csm off, so I dont need to overclock it.

iirc 1701 did not have the cold boot fix included. Think the cold boot fix was not included again till 3008 and has been part of the bios releases since.

[sbakic]

Clouseau wrote:
iirc 1701 did not have the cold boot fix included. Think the cold boot fix was not included again till 3008 and has been part of the bios releases since.

So that means that you got same q-code and light sequence when you turn pc (orange - red - orange -red - white - green(, but with restart you get (orange -red- white -green no repeat), right?

Can you all check it just for current bios you have, sequence of colors when you turn on pc from "shut down" state. Thanks

[Darth_Ender]

If people are doubting the truth of what you are saying, throw some phoronix or similar benchmarks up that test the system instead of just graphics or one subsystem and it should be pretty clear when all you change is the ram speed that performance is improved significantly.

No need to rely on feelings and it shuts everyone up.

[Mace_ya_face]

Just dropping to say that both the 3502 and 6001 lock my 1700 CPU to 2.25GHz/2.7GHz. The 2.2GHz lock happens if the CPU multiplier or CPU voltage is changed from Auto, the 2.7GHz lock occurs if the multiplier and voltage are left on Auto.

• RYZEN 1700
  
• G.Skill TRIDENT Z RGB 2x8GB
  
• ASUS CROSSHAIR VI HERO
  
• ASUS ROG STRIX GTX 1080TI OC Edition
  
• SAMSUNG 960 EVO 250GB (Windows 10 Pro 64-bit)
  
• AMD chipset drivers are up-to-date (17.80 package)
  

[Clouseau]

Darth Ender wrote:
If people are doubting the truth of what you are saying, throw some phoronix or similar benchmarks up that test the system instead of just graphics or one subsystem and it should be pretty clear when all you change is the ram speed that performance is improved significantly.

No need to rely on feelings and it shuts everyone up.

How can one get phoronix test suite to run under Windows 10? Downloaded what Git Hub had to offer and also the PHP files from Windows.PHP.Net and was a no go. Do not use Linux and no time is being devoted to the Windows platform due to unpopularity. Any help or guidance would be appreciated.

[Darth_Ender]

Clouseau wrote:
How can one get phoronix test suite to run under Windows 10? Downloaded what Git Hub had to offer and also the PHP files from Windows.PHP.Net and was a no go. Do not use Linux and no time is being devoted to the Windows platform due to unpopularity. Any help or guidance would be appreciated.

Assuming you're not going to be compiling your own kernel and all that fun stuff since you dont use linux already, you can make do with a live install usb disk. You should be able to boot without issue and install phoronix and run some tests. reboot, change bios settings, re-run the test you had installed and run again. Then link people to the test results that you were given the option to upload. Without actually installing or changing your system at all (aside from the bios).

That's the simple way if you really wanted to get some numbers behind the claims and there aren't any good windows alternatives for you to try out. There may be windows alternatives out there for something like that, i dont generally care about windows benchmarks so i'm not familiar with what suites are freely available.

[Syaoran]

Any word on a new BIOS that actually fixes more than it breaks?

[HnS-CHH]

When bios 1703 to 6001. Memory completed changed, memtest error popping under 100 % from non error 10.000%+. It takes me weeks to get 3466c14 stable. Now I am looking for shortcut what you guys changed to keep memory stable.

There is definity something wrong with this bios 3466c14 14-14-14-14-30-44 rest auto(really high timing) error within 20 %.