AM4 B350/X370 black screen crash thread

[eyetrip]

Ergh, no, not having any luck.

Since changing to this air cooler getting code 8 under load. Time seems to vary wildly on how long before I hit it.

3.95ghz used to be a doddle, and 4ghz not particularly difficult either.

Now am struggling with 3.925ghz, and always getting code 8 hard black screen lock up.

Changing the ******* cooler shouldn't affect oc to this degree. Temps are still really reasonable - while it's running.

I just don't ******* get it.

Google searching, code 8 seems to affect most people at boot, not under load.

Wanted to move back to air for simplicity and because water just seemed like overkill considering temps I was getting.

 

[eyetrip]

Currently have re-seated ram, and testing that before I continue with cpu.

 

[eyetrip]

I know you've already passed stress tests before at 3.9. Try backing down to 3.8.
Reason I say this is mine was at 3.9 for weeks, then for no apparent reason didn't like 3.9 anymore and I had to dial it back.

Yeah, I'm reading about some kind of thermal wall on Ryzen over on oc.net currently. Not sure it makes sense in my head, but I take nothing for granted with this system now.

I may get an AIO - one thing I don't like about this cryorig r1 is that it pulls the cpu from the socket and there's no way to avoid it.

Already bent couple pins once!

 

[eyetrip]

Apparently, according to a few (and one who seems fairly clued up), once you hit a certain thermal limit (they reckon around 68 or so degrees), the voltage requirements shoot up.

So, e.g. If you on water like I was and you stuck 1.42/1.44v in for 4ghz, I could get away with it cause my temps didn't go over 65.

Now that I am on air, I pass that limit earlier - currently testing 3.9ghz at 1.33v, and that is getting me to the same temps roughly as my 4ghz above.

Going over that, takes me over that thermal limit, and I start requiring more voltage than I did under water.

So 3.925ghz would be easily doable at 1.37v before, now it's only very nearly stable at 1.395v.

3.95ghz will probably take me more voltage than 4ghz did under water.

Not sure how correct all this is. Could explain my current woes though. I thought I would simply experience slightly higher temps. Didn't think the voltage requirements would shoot up.

Pain in the ass system :/

 

[eyetrip]

Be interested to know if you can achieve better again if you enable higher fan speed again. If you can be bothered

 

[eyetrip]

Well, I passed 2 hours of occt linpack avx with 3.925ghz at 1.42V, but had 2 whea errors.

Several times when have tried a 2 hour test it would fail in the last 5 minutes, always with couple of whea errors.

Had that when trying for 3.9ghz with 1.35V.

This has been such a ball ache since putting this air cooler in.

 

[eyetrip]

This last run had a peak of 73c, though mostly around 70c.

So it's not like it's miles above where I was getting before.

I'm just getting tired of testing this ******* system now.

If I had known I would have to start all over again moving to air, wouldn't have bothered.

 

[eyetrip]

I don't know why occt linpack insists in failing in the last 10 mins... always in the last 10 mins - when running a two hour test.

Obviously can pass an hour test, but two hour - fail in last ten minutes.

I need to find something that fails sooner.

But, have jsut done my final test on this air Cryoprig r1. Am putting my WC back in - though just for CPU to keep it simple as possible.

Just can't help but think it's applying pressure that the board doens't like. will find out soon enough.

 

[eyetrip]

Installed EVGA precision and no crashes so far...

Ah, I forgot about that.

Does it work with non evga products?

Please let me know how it goes.

Have put my loop back together for cpu only.

Currently testing with occt linpack avx at 4ghz with 1.42v and peak temp has been 62c and is generally holding at 59/60c

I don't expect it will actually pass. Think for 4ghz will require bit more juice. Am mostly just keen to see the behaviour vs the air cooler.

I'm not entirely convinced by the temp thing that it really has as much impact ad I was seeing on stability and oc ability.

Still suspect it was just putting pressure on the board didn't like.

Who knows...

 

[eyetrip]

Had to bump up vcore two notches, but an hour in now without whea errors and no black screen.

Fingers crossed the next hour goes as well.

 

[eyetrip]

2 notches from before on WC or from the air cooler?

Oh, I'm ignoring everything I did on air.

Not getting any of the code 8 black screens hard crashes I was getting all the time on the air cooler.

Still getting whea errors though, and occt linpack still failing just under two hours - that two hour thing is rather specific. What is it that's making it fail then?

Worth reinstalling?

Still playing with voltages to get whea errors sorted -. Though again, only happens after first hour.

May just back off to 3.975.

You cant use it without an evga serial number.

Waiting for my gskill flare x (3200MHZ) ram now. Those corsairs won't be stable i think (even with agesa 1.0.0.6)

At the moment i won't oc...xfr pushes my cores up to 3.9ghz (last hwinfo check) and i think oc for ryzen is still early beta. Waitimg for post 1.0.0.6 agesa.

Pity won't work on my palit :/

 

[eyetrip]

Passed two hours 3.975 with 1.395/1.417/1.439v under load - 1.395v being heavy load.

No black screen crashes, no whea errors, max temp 61c.

Though I did reinstall occt as well, and altered voltages to use llc2 instead of 3.

With that having passed, bumped to multiplier to 40 again and running that. If it fails will simply roll back to 3.975ghz and call it done.

 

[eyetrip]

Leaving at 3.975. Can get 4ghz by upping vcore about 4 notches, but not bothered. It barely breaks 60c, and appears stable.

Have my ram running 3066 and that appears stable too.

I still don't think my issues were temp related in any way, but rather something to do with the mounting of that air cooler. Behaviour was just too off with it.

Oh no I get PTSD just from hearing about black screens due to my infuriating experience with amd r2xx gpus. Right when I was thinking about investing into ryzen I come across this thread. If I were I would just change platform to save you from a lot of anger.

Sooner or later black screens will come back no matter what.

These black screens are somewhat different. I am familiar with the ones you mention, but these are particular to:
mounting pressure - I suspect mine and definitely some EK blocks caused this
bad ram - have had some myself
monitoring software - typically under heavy load, though just odd that this happens
And can also happen with too much OR too little voltage
Has happened with certain nvidia drivers and certain stress apps like realbench

 

[eyetrip]

Given the luck I have had with black screen crashes, I will be encountering that before long too

 

[eyetrip]

Got my new CPU from RMA. Unfortunately (all stock settings) black screen issue is not a plug and play fix. Also swapped to new RAM G.Skill Samsung so my setup should work together as confirmed by many.

Will sit down and tweak tomorrow and this weekend when I have more time. Hope I don't have to replace my motherboard, but it's the only thing I haven't replaced yet....

What are you typically doing when you get black screen?

 

[eyetrip]

Had a black screen issue the other day whilst sat on the desktop. Well it was a code 8 hard crash. Only thing I changed was the performance bias from auto/none to Aida.
Disabled that again andhas been on for 48 hours without a hiccup.

It's like a game of some kind

Makes me think, I tend to leave it on auto. Maybe should just leave it properly disabled... I can still replicate the display cut-out under occt linpack avx, but as that apppears to be the only thing that causes it, am not too bothered, but will try disabling this and see what happens.

 

[eyetrip]

AVX stress is the fastest way for it to occur (Linpack or ycruncher).

On the new chip/ram, the only setting tested was 3.6 CPU and 2133 ram with all voltages/other settings on auto, so I have some tinkering to do, but it's ridiculous to have to fiddle with it this much out of the box. Won't have time to really sit down with it until Saturday.

Is it just the display cutting out, or a hard crash?

If you are not sure, when you get it, if you unplug display connector and reconnect it, does it come back?

Or, if you can see your board when this happens, do you get a q code error?

If it's just display cutting out, what monitoring software for you have installed?

 

[eyetrip]

Yeah, have found overvoltage can be just as much an issue as undervoltage.

If it's hard crash code 8, it's not likely monitoring.

But one thing that oddly caused the most occurrences of code 8 crash for me was moving to a cryorig r1 air cooler.

Went back to my water cooling in the end for my cpu.

 

[eyetrip]

Yea, it's kind of silly.

I've kept up on your adventures it's been a crazy ride for you as well.

I'm on the C6H monoblock from EK. Made sure not to tighten too hard this time around either.

I know there were issues with EK and backplates - have you investigated that? Something to do with the gasket. They caused this exact issue.

 

[eyetrip]

Think I'm just going back to stock for a while.

Find settings that pass stress testing, runs fine for a few days, then starts giving crap all over again.

Boards not worth th PCB it's printed on currently in my opinion.