Thursday, September 1st 2022

ASRock X670E Steel Legend Motherboard Needs Hundreds of Seconds at First Boot or Clear CMOS to Train Memory

At this point, we don't know if this is a limitation at AMD's level or ASRock's, but someone with access to a retail ASRock X670E Steel Legend motherboard, with all its packaged paraphernalia in place, spotted an interesting sticker covering the board's four DDR5 DIMM slots. The sticker has some info on the ideal DIMM slot selection for dual-channel memory (4x sub-channels); but what catches our eye is a table which states just how long the motherboard will take to train the memory the first time it's booted up, or after a clear-CMOS operation (where your BIOS settings are erased).

The table says that a typical setup with two 16 GB modules (read: two single-rank modules in a 1 DIMM per channel/1DPC configuration), takes 100 seconds to train (or until first boot). Two 32 GB modules (typically a pair of dual-rank modules in 1DPC configuration) take 200 seconds, as do four 16 GB modules (four single-rank modules in a 2DPC configuration). The least optimal config, four dual-rank modules in a 2DPC configuration, takes a whopping 400 seconds (almost 7 minutes) to train. That's 100 to 400 seconds of a black screen, or no display signal, enough to unnerve anyone and assume something is DOA.
Update Sep 2nd: The source behind this story confirmed that this is an ASRock-level issue, and that it's been "fixed" with the latest BIOS.

Update Sep 8th: This has been fixed according to ASRock.

Here's the kicker—since UEFI BIOS updates typically clear CMOS, you'll have yourselves some nerve-racking hundred(s) of seconds until the display lights up, letting you know that the BIOS update went through. Interestingly, we haven't yet seen anything to suggest that memory overclock (which involves dozens of reboots and re-training of memory), takes hundreds of seconds—not unless you clear CMOS for some reason.
Source: HXL (Twitter)
Add your own comment

89 Comments on ASRock X670E Steel Legend Motherboard Needs Hundreds of Seconds at First Boot or Clear CMOS to Train Memory

#26
ymdhis
There's something seriously wrong there, boot times keep increasing between chipsets. My b450m boots so fast that I sometimes have to reset to enter the bios because there's less than 1 second to press DEL to enter. My b550m stands there checking the drives and USB ports (according to the debug codes) for something like 10 seconds. And now the X670 will take 100 seconds per RAM stick to boot.

edit: both are asrock boards with 2x8gb ram.
Posted on Reply
#27
Nanochip
ymdhisThere's something seriously wrong there, boot times keep increasing between chipsets. My b450m boots so fast that I sometimes have to reset to enter the bios because there's less than 1 second to press DEL to enter. My b550m stands there checking the drives and USB ports (according to the debug codes) for something like 10 seconds. And now the X670 will take 100 seconds per RAM stick to boot.

edit: both are asrock boards with 2x8gb ram.
Z170 booted fast too (normal post was less than 6 seconds, trained new memory in under 15 seconds) and woke from s3 sleep extremely fast (within 2 seconds).

These newer motherboards have added more functionality more layers more ports and also more latency.
Posted on Reply
#28
evernessince
ymdhisThere's something seriously wrong there, boot times keep increasing between chipsets. My b450m boots so fast that I sometimes have to reset to enter the bios because there's less than 1 second to press DEL to enter. My b550m stands there checking the drives and USB ports (according to the debug codes) for something like 10 seconds. And now the X670 will take 100 seconds per RAM stick to boot.

edit: both are asrock boards with 2x8gb ram.
You misunderstood, first boot or cmos clear. Not every boot.
Posted on Reply
#29
InVasMani
Z170 boots plenty fast even when you turn down some of the boot options to speed up boot time. I forget what they label ufast or some silly name. I don't use it because if I need to enter bios otherwise I have to use a app to boot into it or clear cmos. It will take slightly longer the first time you train the RTL and IO/L however, but not 100 seconds or more.
evernessinceYou misunderstood, first boot or cmos clear. Not every boot.
Important distinction to keep in mind. Still a 100 seconds or longer is quite extreme to train memory.
Posted on Reply
#30
Aldain
For people that leave the EXPO setting on AUTO it will be fine , for fine tuners and OC people it will be a fucked up thing
Posted on Reply
#31
[XC] Oj101
I find it hard to believe that ASRock can be so inept, and after seeing the trainwreck of early DDR4 on Ryzen 1000 series I'm inclined to believe this is an issue on AMD's side? Hopefully it can be fixed with an AGESA update.
Posted on Reply
#32
ARF
evernessinceTo either set timings and test them or ensure that user set timings function. Without memory training the system would simply fail to boot in instances where the user set bad timings or an XMP / EXPO profile doesn't work as you pointed out. That is not an acceptable option as people who aren't computer whizes will think the PC is broken and aren't aware that they could remove the CMOS battery. Mind you it's inconvenient to have to remove the battery in the first place. Memory training mostly gets rid of having to do that.
So, this needs 100 seconds to just override the user preferences and make possible that the boot never fails no matter the user preferences?
Posted on Reply
#33
MikeMurphy
evernessinceIt's inconvenient but given that it only happens on the first boot it's really a non-issue.

If the longer training time is being used to better tune the memory that's a superior approach IMO. First boot being longer in exchange for superior memory performance is more than a worthwhile trade-off.
I wish my x570 had this feature. I have no problem giving it some time to tune itself rather than me having to do it manually.
Posted on Reply
#34
evernessince
ARFSo, this needs 100 seconds to just override the user preferences and make possible that the boot never fails no matter the user preferences?
In the case where a user set RAM values the motherboard would test the RAM to ensure those settings work. It would only do this once when the values are first set.
MikeMurphyI wish my x570 had this feature. I have no problem giving it some time to tune itself rather than me having to do it manually.
Yeah, a lot of performance is left on the table due to loose timings, particularly subtimings.
Posted on Reply
#35
TheLostSwede
News Editor
NanochipI hope so because a production BIOS should not take 100-400 seconds to train memory. That’s unacceptable.
Indeed. This is very much a bug, but it might be present when the IGP is being used on shipping hardware later this month.
Posted on Reply
#36
A Computer Guy
evernessinceIt's inconvenient but given that it only happens on the first boot it's really a non-issue.
Unless you have thread-ripper and 256GB of ram. Expect to wait a very long time manually tuning I guess.
Posted on Reply
#37
evernessince
A Computer GuyUnless you have thread-ripper and 256GB of ram. Expect to wait a very long time manually tuning I guess.
Really can't say in either case as we only have a sample size of 1. X670 won't support threadripper class CPUs anyways.
Posted on Reply
#38
efikkan
This sound very finicky.

Long memory training isn't unheard of. I've seen a Kaby Lake machine taking >1 min.
ARFSorry but what does "train" memory mean?
The BIOS is fine-tuning the memory timings to "ensure" stability.
If the current settings fail, it will try a different set of timings, repeatedly until it either finds a working setting or stalls. Retraining is usually not needed until the PC fails to POST (or the BIOS is reset). As the memory controller and RAM eventually degrades, it will after months or years try to retrain, resulting in gradually slower speeds.

If you want to avoid this, just stick with a memory kit matching the highest JEDEC speeds of the platform and it should just work reliably. Be aware than many expensive "OC" memory kits are not certified for higher JEDEC speeds, so the fallbacks are often very slow. If/when your memory OC fails, it will likely revert to this speed.
Posted on Reply
#39
InVasMani
It's hard to tell if it's just thermal paste dry out or slow hardware degradation at times unless you replace the paste semi-periodically. I think there is definitely cases of assumed hardware degredation when it's more the paste itself losing thermal conductivity slowly and putting additional stress on the IMC with a OC that's heavily pushing thermal limitations.
Posted on Reply
#40
natr0n
Servers do this already my old dual 1366 xeons do this too. Loong boot times for 6 channel memory.
I havent played with them in a long while.
Posted on Reply
#41
efikkan
InVasManiIt's hard to tell if it's just thermal paste dry out or slow hardware degradation at times unless you replace the paste semi-periodically. I think there is definitely cases of assumed hardware degredation when it's more the paste itself losing thermal conductivity slowly and putting additional stress on the IMC with a OC that's heavily pushing thermal limitations.
If the thermal paste is getting bad, you'll see it by checking the temperatures. I believe most quality pastes should last 3 years, but if it's older than that it's wise to replace it anyways.
natr0nServers do this already my old dual 1366 xeons do this too. Loong boot times for 6 channel memory.
I havent played with them in a long while.
Server and workstation boards have extensive POSTs, even mainstream boards get longer POSTs if you turn off quick boot (I usually do, I want it to fail if there is something wrong.). But this isn't memory training though, more like PCIe device validation etc.
Posted on Reply
#42
Dirt Chip
And so the magic of 'new platform fetures' begin :)

I'm going with a 64gb ddr5 in the coming build, waiting 800 sec (more than 10 min..) will be a deal breaker.
Posted on Reply
#43
DeeJay1001
This is why we skip the first generation of a new platform.
Posted on Reply
#44
ARF
DeeJay1001This is why we skip the first generation of a new platform.
I agree but AMD won't be happy to hear this :roll:
AMD has left the chat..
Posted on Reply
#45
FaintCarpet
So high platform costs, reported high cpu temps, and now this for AM5 platform? Yeah, I think it would be best to wait until the bugs are ironed out.
Posted on Reply
#46
mama
This is outrageous! Clearly AM5 is broken. As people have already recommended, everyone should stay with their Intel systems and existing AMD system. I need product availability so please don't go AM5. Really an outrageous problem which I am sure will never get fixed.
Posted on Reply
#48
usiname
Dirt ChipAnd so the magic of 'new platform fetures' begin :)

I'm going with a 64gb ddr5 in the coming build, waiting 800 sec (more than 10 min..) will be a deal breaker.
I don't know where you get this 800s, for 64gb is listed as 200s.
Also this is more like related only to Asrock not with all vendors and it could be fixed by the release.
If this is happening only on CMOS clear, I don't see big problem. For normal overclocking you won't need more than 2-3 times to reach close to the sweet spot and even if your system is unstable, it would be enought to go in the BIOS and change the settigns. I learned it after switched to 3070 that make the removing of the battery impossible and after my mobo does not have button for this I must to short the pins that reset the CMOS manually which is very big pain with the postion of my PC.
This could be problem only for extreame overclockers that tunning the RAM to the best possible settings, but as I said, this most likely is asrock's problem and will be fixed fast
Posted on Reply
#49
Mussels
Freshwater Moderator
400 seconds?


I'd assume it was dead and have it packed away for an RMA by then
mamaThis is outrageous! Clearly AM5 is broken. As people have already recommended, everyone should stay with their Intel systems and existing AMD system. I need product availability so please don't go AM5. Really an outrageous problem which I am sure will never get fixed.
And you're already proven wrong, it's an asrock only bug that's already been fixed in a BIOS update
Posted on Reply
#50
Max(IT)
BloaxThe Fury X will be an overclocker's dream, and so will AM5!
Based on what ? AMD historically never has been an overclocker’s dream (if overclocking still makes any sense at all…).
Posted on Reply
Add your own comment
Dec 22nd, 2024 09:03 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts