BIOS updates for AM4 motherboards

This might come as a surprise but it seems that bios 1201 is actually far more predictable and works much better around the cold boot problem. Even if it does not "boot" it does not stop always allowing interaction with the bios even if the "overclock" failed. Last 3 days been better where I just have to adjust the voltage for dram (normal and boot) until it finds something that it will run (sadly this still has some variation from 1.375 to 1.39 and it is never the same.
 
So after further playing around I can't get 3200 stable, random cold boot issues. 2933 for right meow.

Msi b350 tomahawk Arctic and geil Evo x (hynix chips)
 
1006 might be the mark. Lot of memory suppliers are still not supported. Need a chipset or processor revision. AVX512 is a sticking point
 
pgaster said:
I hate to burst your bubble but I'm not sure that update will do much good. Somewhere else I read that the long awaited memory updates are coming with AGESA 1005 or 1006.
Gigabyte has a 1005 release already and maybe others do as well. Asus seems a bit behind here.
Click to expand...

What bubble ? It is just a new BIOS another that been out for a while (week) in beta form. It shows what has been patched. You can look at the link that I posted a few pages back which holds the post when this bios was released on overclockers.net

The rumour goes that Asus will do a 1006 AGESA based bios and skip 1005.
 
Pieter3dnow said:
What bubble ? It is just a new BIOS another that been out for a while (week) in beta form. It shows what has been patched. You can look at the link that I posted a few pages back which holds the post when this bios was released on overclockers.net

The rumour goes that Asus will do a 1006 AGESA based bios and skip 1005.
Click to expand...

Sorry, didn't mean anything offensive, and wasn't really even replying to you, it was a more general statement.
Lots of news sites have been saying for a month or so that there would be a bios update in May that would enable higher RAM speeds. It's supposed to really help with Hynix IC ram. If it's still AGESA 1004a then anyone thinking that this update was that mythical May update for extra ram speed will be sorely disappointed.
I'm in a similar boat. I've got an Asus Prime B350 Plus and I'm checking the site every day hoping for the update.
 
pgaster said:
Sorry, didn't mean anything offensive, and wasn't really even replying to you, it was a more general statement.
Lots of news sites have been saying for a month or so that there would be a bios update in May that would enable higher RAM speeds. It's supposed to really help with Hynix IC ram. If it's still AGESA 1004a then anyone thinking that this update was that mythical May update for extra ram speed will be sorely disappointed.
I'm in a similar boat. I've got an Asus Prime B350 Plus and I'm checking the site every day hoping for the update.
Click to expand...

Even then I'm thinking not everyone gets it right the first time so even if they used the latest AGESA we will still see updates that matter after 1005/1006 if not for all but for some.
 
Pieter3dnow said:
Even then I'm thinking not everyone gets it right the first time so even if they used the latest AGESA we will still see updates that matter after 1005/1006 if not for all but for some.
Click to expand...

Better take a deep breath. The ansiness will not bring the bios faster. I have been told by reliable sources that the updates will be coming near the very end of the month.
 
os2wiz said:
Better take a deep breath. The ansiness will not bring the bios faster. I have been told by reliable sources that the updates will be coming near the very end of the month.
Click to expand...

As long as people can find something that works(where they have some things working but not all) for them then people know they won't have to update until then.

After this morning cold boot failure at 1.39 normal 1.39 boot , tried 1.355 boot and 1.36 normal and it all works again at 3200. This is what makes me wonder what exactly is going on with crosshair VI hero bios settings.
 
Last edited:
Might have something to do with the low power process where higher voltages cause instability. Lot of stuff to iron out. These chips seem to like lower voltages, depends on the sample you have.
 
Pieter3dnow said:
As long as people can find something that works(where they have some things working but not all) for them then people know they won't have to update until then.

After this morning cold boot failure at 1.39 normal 1.39 boot , tried 1.355 boot and 1.36 normal and it all works again at 3200. This is what makes me wonder what exactly is going on with crosshair VI hero bios settings.
Click to expand...


Asus used to be good about their bios. When I bought the Crosshair VI on launch day, March 3rd, It was a complete brick. I had to return it to Micro Center and I bought the MSI X370 Titanium. Never failed me once. I do have issues with their slow bios updates but it is a whole lot better than issuing official bios releases that turn out to be bricks. Until I see a better Asus I will no longer purchase their boards.
 
gigaxtreme1 said:
Might have something to do with the low power process where higher voltages cause instability. Lot of stuff to iron out. These chips seem to like lower voltages, depends on the sample you have.
Click to expand...


I see too many people recommending high SOC/ CPU-NB voltages and too high cpu core voitages here like they know what they are talking about. I would never run above 1.10 v SOC and never above 1.40 volts for 24/7 use. Right now I am at 3.85 GHZ with 1.38 core voltage . When you spend $500 for a cpu you have to be concerned about it degrading over time unless you are a fool with money to burn who only cares about bragging rights
 
os2wiz said:
Asus used to be good about their bios. When I bought the Crosshair VI on launch day, March 3rd, It was a complete brick. I had to return it to Micro Center and I bought the MSI X370 Titanium. Never failed me once. I do have issues with their slow bios updates but it is a whole lot better than issuing official bios releases that turn out to be bricks. Until I see a better Asus I will no longer purchase their boards.
Click to expand...

The odd thing was is that I ordered 3 different ones the gigabyte I cancelled. The Taichi took so long the person at the shop told me look at another one , it was the last one on the list. weird thing is about a few days later the Taichi was in stock after the brick stories by several people all across the web I was not looking forward to Asus crosshair VI hero motherboard at all.
 
gigaxtreme1 said:
Might have something to do with the low power process where higher voltages cause instability. Lot of stuff to iron out. These chips seem to like lower voltages, depends on the sample you have.
Click to expand...
I'm talking about the dram voltages , haven't even touched the cpu.
 
Asus just updated their QVL for the B350M-A. Last time they did that a bios release came out a few days later that increased memory speed, so I'd expect the same. My ram is now listed at full speed in the QVL, but running one step lower on current bios, so good news there.
 
Which manufacturer has better memory support (up to 3000) in general for B350 mb's? Not looking to overclock ram.
 
bobzdar said:
Asus just updated their QVL for the B350M-A. Last time they did that a bios release came out a few days later that increased memory speed, so I'd expect the same. My ram is now listed at full speed in the QVL, but running one step lower on current bios, so good news there.
Click to expand...

The Prime X370 QVL was also updated on the 15th. Can't really tell what has changed. There are still a lot of 3000 - 3200 MHz kits that are only listed at 2133 MHz :(
http://dlcdnet.asus.com/pub/ASUS/mb...-PRO/PRIME_X370-PRO_Memory_QVL_2017051511.pdf
 
Last edited:
os2wiz said:
I see too many people recommending high SOC/ CPU-NB voltages and too high cpu core voitages here like they know what they are talking about. I would never run above 1.10 v SOC and never above 1.40 volts for 24/7 use. Right now I am at 3.85 GHZ with 1.38 core voltage . When you spend $500 for a cpu you have to be concerned about it degrading over time unless you are a fool with money to burn who only cares about bragging rights
Click to expand...

i mean sure if you want to use the processor for the next 10+ years that's fine.. but if you tend to upgrade every 2-3 years it's really not that big of a deal increasing those voltages slightly over stock spec. modern processors are pretty damn tough to kill or degrade unless you go completely ape shit with voltages, the motherboard components are the weakest link of the entire setup these days.
 
sirmonkey1985 said:
i mean sure if you want to use the processor for the next 10+ years that's fine.. but if you tend to upgrade every 2-3 years it's really not that big of a deal increasing those voltages slightly over stock spec. modern processors are pretty damn tough to kill or degrade unless you go completely ape shit with voltages, the motherboard components are the weakest link of the entire setup these days.
Click to expand...
Yeah, my mobo is a recent casualty, though I don't know the status of the cpu yet, I'm sure it's fine.
 
sirmonkey1985 said:
i mean sure if you want to use the processor for the next 10+ years that's fine.. but if you tend to upgrade every 2-3 years it's really not that big of a deal increasing those voltages slightly over stock spec. modern processors are pretty damn tough to kill or degrade unless you go completely ape shit with voltages, the motherboard components are the weakest link of the entire setup these days.
Click to expand...
Slightly like 1.20 volts instead of 1.1 volts SOC? Or 1.45 volts cpu core voltage instead of under 1.40 volts?? That is complete bullshit. Thise voltages will likely degrade these chips within one to 2 years not 3 years like you suggest. Keep up the stupdity. This is NOT Bulldozer or Piledriver. Even those chips degraded in less than two years if you pumped above 1.48 volts cpu core voltage 24/7.
 
Many are likely increasing SOC voltage more than what is necessary. I know some boards even boost it to 1.1V automatically when you run factory overclocked memory. I only had to increase it by 0.02V to get my RAM stable at 3200.
 
JimmiG said:
Many are likely increasing SOC voltage more than what is necessary. I know some boards even boost it to 1.1V automatically when you run factory overclocked memory. I only had to increase it by 0.02V to get my RAM stable at 3200.
Click to expand...

Now here is a man that makes sense. Too many in this crowd are trying to fit a square peg into a round hole. A small soc/cpu-nb voltage increase like .02 or .03 volts is fine, but jacking it up higher is more likely to screw your chip or motherboard than make your memory work right. In 95 out of 100 cases the problem is not the a-xmp voltage , it is memory sub timings that we do not have access to. When the Agesa 1.06-based bios updates come out near the end of this month we will have a proper fix for most people. They can stand on their heads and swear at the heavens for all the good it will do until then.
 
bobzdar said:
They just updated the QVL again yesterday, not sure what changed in 3 days unless there was a mistake in the previous one (or they made some tweaks to the yet to be released bios).
Click to expand...

Well it does seem like more modules are getting listed at 3200.
G.SKILL F4-4000C19D-16GTZSW
GALAXY HOF4CALCS3600K17LD162C
CORSAIR CMK16GX4M2B3600C18 16GB
G.SKILL F4-3300C16Q-16GRKD 16GB
G.SKILL F4-4133C19D-16GTZKW

Before the last two updates, there were only 2, the Corsair 3600's and GALAXY 3600's.
Also, 9 modules listed at 2933.

Could just be that they have tested a bunch more modules, though.
 
JimmiG said:
Well it does seem like more modules are getting listed at 3200.
G.SKILL F4-4000C19D-16GTZSW
GALAXY HOF4CALCS3600K17LD162C
CORSAIR CMK16GX4M2B3600C18 16GB
G.SKILL F4-3300C16Q-16GRKD 16GB
G.SKILL F4-4133C19D-16GTZKW

Before the last two updates, there were only 2, the Corsair 3600's and GALAXY 3600's.
Also, 9 modules listed at 2933.

Could just be that they have tested a bunch more modules, though.
Click to expand...


good to see, hopefully more start getting added.. ram supports pretty much the only thing holding me back from finishing my new build.. having to search through different forums to find what unlisted ram people have been able to get working on different boards is driving me nuts.
 
sirmonkey1985 said:
good to see, hopefully more start getting added.. ram supports pretty much the only thing holding me back from finishing my new build.. having to search through different forums to find what unlisted ram people have been able to get working on different boards is driving me nuts.
Click to expand...

newsflash! I was scanng the AMD motherboard forums specifically the ROG Crosshair VI overclocking forum on overclock.net. Elmor, the master overclocker and Asus employee gave hotlinks for two beta bioses for Crosshair VI that will each help certain specific memory modules including G.Skill Hynix dimms and Samsung B-die dimms with four 8GB dimm configurations. Please run, do not walk to that foprum and download the appropriate beta bioses for your ROG Crosshair VI motherboard.
 
gigaxtreme1 said:
Need a chipset or processor revision. AVX512 is a sticking point
Click to expand...
What would a chipset or processor revision do to improve the BIOS? Most of Ryzen's bugs and issues seem to only require firmware updates to fix. It just taking a while to do so.

And pretty much nothing uses AVX512 in the consumer space. Nor will it for years. Only HPC cares much about it since that is the work load that wins big with it. Even AVX256 has poor use in the consumer market. And likely will still for years too.

The use cases for very wide and ultrawide SIMD bringing big performance improvements aren't widespread unfortunately.
 
os2wiz said:
newsflash! I was scanng the AMD motherboard forums specifically the ROG Crosshair VI overclocking forum on overclock.net. Elmor, the master overclocker and Asus employee gave hotlinks for two beta bioses for Crosshair VI that will each help certain specific memory modules including G.Skill Hynix dimms and Samsung B-die dimms with four 8GB dimm configurations. Please run, do not walk to that foprum and download the appropriate beta bioses for your ROG Crosshair VI motherboard.
Click to expand...

***Oops seems a double post ;)

These were posted in here :
http://www.overclock.net/t/1624603/rog-crosshair-vi-overclocking-thread/15900#post_26106654
 
mesyn191 said:
What would a chipset or processor revision do to improve the BIOS? Most of Ryzen's bugs and issues seem to only require firmware updates to fix. It just taking a while to do so.

And pretty much nothing uses AVX512 in the consumer space. Nor will it for years. Only HPC cares much about it since that is the work load that wins big with it. Even AVX256 has poor use in the consumer market. And likely will still for years too.

The use cases for very wide and ultrawide SIMD bringing big performance improvements aren't widespread unfortunately.
Click to expand...
Thank you Mesyn191. A Chipset revision could bring us more PCIe lanes more along the lines of x399. A few extra features and resources would sell more chips, maybe with regard to Ryzen pro, as yet unlaunched. As far as AVX 512 being a sticking point, it is only so for company detractors. AMD in the past has bumped the platform a little as it matured. Memory and stability are coming with Bios updates and I can't help feeling that the launch was a little rushed probably because AMD doesn't have the resouces to test every vendor's modules in every use case.
I look forward to the whole AMD ecosystem and platform growing. Avid fan here.
 
New MSI bios beta version 1.72 with a myrioad of new memory timings and options up to ddr4 4000 for memory. I just wish I knew what to do to get my 4 dimms up to 3200mhz on my samsung b-die dimms. Too many subtimings available and no clue to what values I should p[lug in there. The voltages for SOC/cpu nb and dram voltage are the easiest to deal with but insufficient to raise my memory with 4 dimms to 3200mhz. Even raising 2 dimms to 3466 or 3600mhz is beyond me. I tried but not successful.
 
I've been dormant with my Ryzen due to little I can tinker with on the Titanium, and newer BIOSes being flaky with Cold Boot issues (but it's not as bad as before... but is still not as functional as 1.10 shipping BIOS was), but I do keep an eye on the Beta threads on MSI...

NEW 1.72 BETA BIOS is out for the Titanium! I tossed it into AMIBCP to see if visually anything is new and... :eek::woot::happy:
First up, the big additions... Check out those SPPEEEEEEEEDS! (Highlighted items = Changed or NEW, compared to past versions)
Titanium-1.72-DRAM.png

And timings... I hope these are editable this time! They were there in the past (the primary and secondary) but they were only Viewable, not changeable, and looked just like this. In 1.6 they vanished, so I assume their re-appearance is because of being user-adjustable now :D
1.72-New Timing Options #1.png 1.72-New Timing Options #2.png
1.72-New Timing Options #3.png
(SSPACE and ESPACE I presume are indicators for which options are shown/hidden by the "Nested Menu" feature that the Timings page uses, standing for START SPACE and END SPACE, so everything in between will be shown/hidden upon expending the menus, indicated by the [*])

ALSO new are some voltages! No more CrossHair master-race! ;) Or, at least, we're moving a bit closer... lol They still have a leg up in multiple other areas... (I doubt the BCLK that's available [hidden] will be tunable beyond 103MHz still, but I'll report back)
1.72-New Voltage Options.png

I presume that CLDO is "CPU Low Drop Out", which as I understand, was speculated to be disabled on Desktop Ryzen chips, but otherwise still present in them. Maybe this turns them on?? *fingers crossed*
Also we see... two new DRAM options! All three "DRAM Voltage" choices have the exact same Help String, which is why I've not bothered capturing it exclusively, but by the looks of it they refer to: Operating Voltage, POSTing Voltage (or in-BIOS volt?) and then Initial Power-Up Voltage (Training)... which I think is what'll solve the Cold Boot issues on our Titaniums for those of us encountering them (like myself, when using any BIOS version beyond shipping's)

[NOTE: The CPU_VDDP option, from all I've been able to determine, is non-functional on our X370 Titaniums. Whether or not it'll become so in this update, I dunno. However, the reading in-BIOS has previously reported Zero Volt, my adjustments to it have always resulted in no change to any of the HWiNFO64 readings, and even the on-board test point below the Chipset reads 0.00V]

I am actually excited! This is the update I've been waiting for... it'll allow me to actually FUTZ WITH STUFF :pompous: I've not even bothered overclocking my CPU yet until I'd managed to have a stable system, and while it's 98% stable in most regards, the Cold Boot issue and some very random infrequent CTD or BSoDs have been enough to put me off.

So with that... off I go to flash this baby! :shame:

EDIT: Damn you os2wiz! lol Serves me right for being too detailed and going over the damn BIOS with a fine-tooth comb.. :p
 
os2wiz said:
New MSI bios beta version 1.72 with a myrioad of new memory timings and options up to ddr4 4000 for memory. I just wish I knew what to do to get my 4 dimms up to 3200mhz on my samsung b-die dimms. Too many subtimings available and no clue to what values I should p[lug in there. The voltages for SOC/cpu nb and dram voltage are the easiest to deal with but insufficient to raise my memory with 4 dimms to 3200mhz. Even raising 2 dimms to 3466 or 3600mhz is beyond me. I tried but not successful.
Click to expand...
Since I didn't really plan to flash the default MSI version (I edit mine to manually change a bunch of user settings, since it's easier to do from Windows), can you comment on which of the SubTiming choices show up for you compared to what is visible in my above list?

Lots of times these are present, but hidden to us. Case in point, the BCLK option on the main page which has been in the BIOS since I think BETA v1.3x (release 1.3 was an older version). So I'm curious if there are any options I'll need to manually un-hide before I go about flashing this one. :)

Thanks!
 
Formula.350 said:
Since I didn't really plan to flash the default MSI version (I edit mine to manually change a bunch of user settings, since it's easier to do from Windows), can you comment on which of the SubTiming choices show up for you compared to what is visible in my above list?

Lots of times these are present, but hidden to us. Case in point, the BCLK option on the main page which has been in the BIOS since I think BETA v1.3x (release 1.3 was an older version). So I'm curious if there are any options I'll need to manually un-hide before I go about flashing this one. :)

Thanks!
Click to expand...

Okay but how much of the hidden stuff is actually working , it is nice you can put values to things you want to change but if nothing happens ;) .
 
You must log in or register to reply here.
Back
Top