Asus Prime X370-Pro Owner's thread

[SumpNut]

Well I just update my BIOS. No change in my ram.

Still cant do 3200Mhz at 1.4v with the Corsair LPX.

Is it because I am running a 1700?

At least my previous OC settings are still running just fine.

But disappointed that this new AGESA has not brought me closer to the magical 3200Mhz.

I can on occasion get to Windows at 3200Mhz but it took 1.45v and it failed a HCI mem test. (Although it was fine during two rounds of BF1)

I run Corsair LPX 3200 (Hynix) @ 3200 with my 1700 on the x370 prime pro.

If I simply applied the rams DOCP profile it would not POST, however, if I adjusted the following voltages then the ram DOCP profile could be used

It took a lot of voltage tweaking to get it stable, of the top of my head the key voltages required are:
An SOC is around 1.05 at the moment (has been as high as 1.1 on older bios's)
VDIMM is at 1.3950
VTTDDR is at 0.69750 (must be exactly half of VDIMM)

Other power related settings are:
SOC power phase control set to extreme and the SOC switching frequency bumped up to 600

 

[opethdisciple]

I run Corsair LPX 3200 (Hynix) @ 3200 with my 1700 on the x370 prime pro.

If I simply applied the rams DOCP profile it would not POST, however, if I adjusted the following voltages then the ram DOCP profile could be used

It took a lot of voltage tweaking to get it stable, of the top of my head the key voltages required are:
An SOC is around 1.05 at the moment (has been as high as 1.1 on older bios's)
VDIMM is at 1.3950
VTTDDR is at 0.69750 (must be exactly half of VDIMM)

Other power related settings are:
SOC power phase control set to extreme and the SOC switching frequency bumped up to 600

Thank you. I'll give this a try when I get home.

Currently my ram is at 3066Mhz using SOC 1.1v and DRAM Volt at 1.4v and ProcODT at 60ohms.

Increasing DRAM volts to 1.45v I can get in to Windows and even play BF1 but it fails before 100% in HCI mem test.

----

So based on what you said above, I will leave my settings as they are but try VTTDDR at 0.7 and SOC power phase control set to extreme plus SOC switching frequency bumped up to 600.

Cheers

Unfortunately that didn't work for me.

DOCP profile on (3200Mhz)
SOC at 1.1v
DramV at 1.4v
ProcODC at 60ohms
SOC power phase set to extreme
SOC Switching at 600.

I can get to Windows with no fuss, but it didn't take more than 30secs for HCI to error.

Any clues where I should go next?

Does @8 Pack have any clues how I can get my Corsair LPX ram stable at 3200Mhz on this board?
https://forums.overclockers.co.uk/members/8-pack.129102/

 

[AndreiD]

Most people don't properly test their RAM, so their "3200 stable" isn't actually that stable in the end.
Issue with Ryzen is that there seems to be some variation in the IMC, so it's up to the silicon lottery to get rated speed.

I'm currently stuck at 2933Mhz with B-die because my 1700 can't keep the ram at 3200Mhz without any errors.

 

[SumpNut]

Thank you. I'll give this a try when I get home.


I can get to Windows with no fuss, but it didn't take more than 30secs for HCI to error.

Any clues where I should go next?

Does @8 Pack have any clues how I can get my Corsair LPX ram stable at 3200Mhz on this board?

I'm sorry it didn't help. Have you tried higher SOC voltage than 1.1v to try and rule out the memory controller?

I wonder if there are a range of qualities within the various families, could I simply have been lucky with my Hynix based LPX modules?

 

[SumpNut]

Most people don't properly test their RAM, so their "3200 stable" isn't actually that stable in the end.
Issue with Ryzen is that there seems to be some variation in the IMC, so it's up to the silicon lottery to get rated speed.

I'm currently stuck at 2933Mhz with B-die because my 1700 can't keep the ram at 3200Mhz without any errors.

I'm not sure if this is aimed at me?

Either way, I'm currently stable in memtest86, HCL and prime95 at the above settings, I've even managed to apply a little bit of additional memory timing tightening as per the work by the Stilt...

I don't put this down to some super human ability I have to punch numbers that I got off the internet into a bios. I suspect, as stated above, that my Hynix brand sticks are simply not the worst ones out there...

 

[AndreiD]

@SumpNut I just meant in general

I've seen a lot of people thinking if Battlefield 1 runs for a few maps then their memory is "stable". Not that many seem to go the memtest86/HCI/GSAT route.

 

[opethdisciple]

Noticed today the latest BIOS has been pulled from the website.

 

[AndreiD]

It's probably because it could randomly brick the board (some people on other forums were reporting it) or due to the reports of increased memory latency, seemed like it had some issues.

 

[Spyhop]

What a farce this stuff is

 

[DarkHorizon472]

I swapped to the 8 pack 3200 ram as the Corsair LPX was not BF1 stable beyond 3066 and even then kept resetting to 2133 randomly on bootup. Now completely BF1 stable at 3466 c15, I find being stable in BF1 for over an hour as a quick test means it is stable, BF1 is very good at finding unstable overclocks.

 

[opethdisciple]

I swapped to the 8 pack 3200 ram as the Corsair LPX was not BF1 stable beyond 3066 and even then kept resetting to 2133 randomly on bootup. Now completely BF1 stable at 3466 c15, I find being stable in BF1 for over an hour as a quick test means it is stable, BF1 is very good at finding unstable overclocks.

Hmmm. You may want to double check that with something like HCI.

I was able to play two rounds of BF1 (40mins) at 3200mhz but then it fell over pretty quick when loaded up with HCI.

 

[DarkHorizon472]

Mine does similar at 3600 c15.

 

[AndreiD]

BF1 is a pretty bad way to determine if RAM is stable, use memtest86 or HCI memtest if you want to be sure.
Also that 8 Pack Dark Pro seems to be really good, too bad I have the regular Dark Pro

 

[opethdisciple]

New BIOS has just been posted. Version 3401.

I guess it's just a reissue of the previous one.

 

[AndreiD]

Guess they fixed the bricking issue the previous bios had. New agesa doesn't seem to do anything apart from bring APU support.

 

[Spyhop]

3402 out. They can't be bothered with a changelog.

 

[Revel&Quaff]

Just thought I would post regarding the PRIME X370 PRO.

I was running BIOS 0807, this was upgraded from 0805 (out of the box) about 6 months back but I had not noticed that several new version where available since.
So I upgraded, to 3402 via the windows AI Suite (EZ Updater)... Big mistake... Huge Mistake
It bricked the board.
With the board, I get the two beeps and Crashfree display regarding X370PRO.CAP.
Both CD and USB (every USB port was tried.).

Luckly (long story), I had another brand new board.
This time though I updated from 0805 straight to 3402 via the updater in the BIOS, no problems. System booted up with no complaints, even W10. It appears very stable but I've been only running this for a few hours (at the time of posting)
Spec.... Nothing OC'd yet... Windows 10 pro, Ryzen 1700, 32GB 2x16GB Corsair LPX (yes ...problems with speed (so 2166 at the mo), however posts in this thread are promising.. I must try again), EVO 960 500GB, Several other HDs, EVGA Geforce GTX 1050 Ti 4 GB.

If someone has got Crashfree to work, I would be interested. However I wouldn't rely on it.

Apart from this issue it has been a great board for me.

 

[AndreiD]

Never use the windows utility to update the bios, always do it via the EZ Flash utility within the BIOS itself, it's the safest way to do it.
Just updated to 3402 too, haven't noticed any improvements.

 

[Davanius]

This board has had a lot of updates and still can't run most ram kits at 3200 Mhz.

With ram prices being so high I would want to get the rated speed for what I paid.

Looks like first gen Ryzen is still in beta, 2nd gen better get this right.

 

[AndreiD]

I hope 2nd gen is better too, since my current 1700 can't run my B-die kit at 3200 on this board. Part of it is probably due to motherboard firmware and some trace optimization, but at this point I'm pretty sure most of the issue stems from Ryzen's IMC.