PC Processors

mackbolan777 · ‎12-13-2020

Ok so weeks since fixing the IF running at 1866 to match my 3733, so called , "sweet spot" RAM on my 3600X build. I stopped having errors about Processor Core Failure/Hierarchy Error, and the usual complaints on here about the Ryzen CPU's in general. Switching to the Adrenaline 20.8.3 driver seemed to stop the error relating to that driver crashing.

Tonight after about 2 hours into BF4, the PC screams, black screens and reboots. The error: WHEA LOGGER 18 Processor Core Failure. So I'm done with AMD at least as far as CPU's go, just done. I looked for a single Intel with this error and found one in 2017 but it was error 19 and that resulted in CPU replacement after replacing everything else first.

Same song here but it's happening to every model CPU from the Ryzen 3000's to the new Ryzen 5000's and now the only thing "Ryzen" is my blood pressure! 7 full months of fighting this machine and I've had it. Trying to help others on here with similar issues when in fact there is no cure, not from AMD. All these "quirks" with RAM, board topology, the "uniqueness" is a cover for AMD's failure to make a working product. Hundreds of thousands of posts about the same failures and errors all over the internet and yet AMD still keeps pumping out this problematic hardware.

Sure I'll RMA the CPU but I'm done with AMD CPU's until I see they actually fixed something right. $1100 worth of a problem, bunch of you know what. This whole forum was designed to catch the flak for AMD. Like getting one CPU to work properly is finding a unicorn or a pot of gold at the end of a rainbow. For $1100+ No thanks. The RX 5600 XT is probably fine but we'll see after I switch CPU teams.

Yep, this will make a few of you guys happy that I'll not be in here very often if at all. At least I won't be wasting my time or money on trash that breaks every hour. AMD's RMA'ing Ryzen CPU's under a year old some less than a week, by the HUNDREDS, what's wrong here???? I'm glad I folded my business in 2012, because my customers would be banging on my door everyday since I built 99% AMD. I'd need to build a "forum" to hide behind too.

"It worked before you broke it!"

mackbolan777 · ‎02-25-2021

3 months late and a nope. It was a Windows Update that they fixed in a few days. But now I've moved on to a 5600X and an RX 6800. Same board, RAM, but an all core 4.7 OC at 1.27v and the card is killer. But thanks for reading and welcome aboard!

"It worked before you broke it!"

View solution in original post

colesdav · ‎12-13-2020

I have run BFV quite a lot, for up to 2 hours at a time, to test PC.

I have run it on Ryzen 2700X at fixed 4.3 GHZ and 64GB RAM at 3200MHz. Also at stock settings.

I have also run it on Ryzen 3000 series processors.

I have run it on AMD HD7970 6GB, R9 280x, R9 390X OC, R9 Fury X, R9 Fury, R9 Nano, RX 590, RX Vega 56, RX Vega 64 , and RX5700XT cards.

Also Nvidia GPUs from GTX780Ti to RTX2080OC.
Also on Intel Haswell i7-4770K and 4790K processors, and on much more recent Intel mobile gaming laptop processors.

General experience has been:

I have seen crashes on both Intel and AMD CPUs on BFV.

AMD and NVIDA GPUs have both crashed out, but Nvidia GPU much less frequently, rarely in fact, and with clean exit - i.e. BFV just shuts down, the GPU drivers, Nvidia Control Panel and GeForce Experience keep running.

AMD GPU Drivers have been the cause of the crashing the majority of the time. On Intel or AMD CPU.
Verified by repeating the crash on system with no overclocking at all, after Windows system checks, re-verifying game files, rerunning RAM Tests.

AMD Drivers on BFV have been pretty solid recently, even on RX5700XT.
I tested BFV frequently over 3 months August - November.
I did get a few black screens recently on Adrenalin 2020 20.11.1/2/3 drivers.
That was tested on Windows 10 Pro 2004.

RE: AMD's failure to make a working product.

AMD seem to have wiped out Intel CPU for new desktop builds so they must be working for most people.

Zen+ memory controller and ram compatibilty on top end Asus ROG Crosshair Hero VII (Wifi) X470 motherboard was an absolute pain to get working at motherboard rated DRAM speed of 3200MHz with fully populated 64GB of RAM. I did get it to work though.

Ryzen 3000 series Zen2 builds have been very easy. I watched & read motherboard reviews and make sure to select a good motherboard from MSI or ASUS so far. I purchase best Corsair PSU possible. Stick with what works on one build and repeat that build and components.

Which motherboard have you been running with please?
Did the crash happen when your PC was overclocked?
What CPU cooler were you using.

There have been lots of changes to Windows in 20H2, BFV updates recently etc.
Software is so complex, Microsoft do testing on their consumer user base "the herd" in the field., their OS code is not great at new update release. Windows 10 has generally been a nightmare at each major update.

RE: Yep, this will make a few of you guys happy that I'll not be in here very often if at all.
Not true.
You have been giving lots of useful support to AMD users on the AMD Community Forum for free.

tatane · ‎12-13-2020

@colesdav wrote:
Ryzen 3000 series Zen2 builds have been very easy. I watched & read motherboard reviews and make sure to select a good motherboard from MSI or ASUS so far. I purchase best Corsair PSU possible. Stick with what works on one build and repeat that build and components.
How is having to review every component for compatibility a "very easy" experience ? I understand having to do that kind of extensive review for a platform when you want optimal performance, but not for compatibility, standards are meant to ensure that (DDR4, PC-whatever rating of memory, pci-e, chipset/cpu support etc ...)
Having to do that for a motherboard when you go for what should be the latest and greatest supported chipset/cpu combo is in no way an easy experience, let alone a very easy one.
@colesdav wrote:
There have been lots of changes to Windows in 20H2, BFV updates recently etc.
Software is so complex, Microsoft do testing on their consumer user base "the herd" in the field., their OS code is not great at new update release. Windows 10 has generally been a nightmare at each major update.

This kinda goes in contradiction what you're saying about AMD "wiping out" Intel CPU for new desktop builds ... with AMD marketshare growing, Microsoft ought to test those patches on more and more amd cpus as well ...

colesdav · ‎12-13-2020

I have to check Intel Motherboard QVL list for tested and approved components as well.
Windows 10 major updates also cause problems for Intel CPUs.

Neither of those are specific to AMD Ryzen CPUs.

tatane · ‎12-13-2020

looking for QVL ain't what i describe as "doing extensive research", OP (and most topics about those kind of errors too, including issues I've had myself before having to RMA a 5950x that was less then a week old before it decided to just die and never boot again) actually have done their fair consumer duty in checking those, but it seems we now have to check what exact model from each brand might work correctly for each and every different AMD CPU ? Let me answer this straight : NO we don't, because AMD is telling us that an x570 chipset should support it, and that once we have an x570 mobo, an AMD Cpu and ram that's from the QVL, then it should just work, period, no black magic or rocket science should be needed from customer end as we have here ...

colesdav · ‎12-13-2020

No need to get emotional.

RE: "RMA a 5950x that was less then a week old before it decided to just die and never boot again".

Maybe you were just unlucky and got a bad chip?:
Maybe you messed with BIOS settings and ran the chip at higher than rated voltage?
Maybe your heatsink was not fitted properly?

Perhaps you had a defective or bad quality mortherboard or a bad PSU?
Could be many things that cause a CPU to fail within a week.
Are there widespread reports about Ryzen 5950X dying in a week?

tatane · ‎12-13-2020

Basically, before it decided to die, mine was also showing this kind of errors, hinting at L1 Cache issues and processor errors ... As OP is expressing, this is a trend you can see here and on a lot of other forums nowadays, on quite a number of both 3000 and 5000 series, as well as people struggling to get RMA's, people getting another lemon CPU after a first RMA taking a full month etc ...

"Widespread reports about 5950x" ? please don't make me laugh, how could anything about such a low spread cpu be widespread ?! got to place 2 orders on 5950x's on launch day here (one direct from AMD, the other from an official distributor), one arrived and died after 4 days, the other one still hasn't arrived, and AMD is out of stock to even provide me with a replacement for my rma, so how would one build statistics on that CPU when it's seemingly made of pure unobtainium ?

Messed with BIOS Settings ? welp, sure had to to try and keep it stable for more then 10 minutes, but most of the settings i've tinkered with were actually disabling PBO and any dynamic overclocking as I was looking for stability and not to break benchmark records ...

Heatsink ? I do watercooling and my CPU never went over 70° ...

Bad luck ? certainly, but worried that the WHEA errors were not all due to the cpu being a lemon when I see all those posts around, and since I was already seeing/encountering those issues before my cpu died I can't rule out those being possibly related somehow / not only due to one having a bad luck when it seems (as OP is pointing out righteously) that getting a CPU that just works means you're lucky and getting one that doesn't seems to be the new normal amd doesn't give a flying one about

Bad quality motherboard ? well, we're back to our initial point there ... x570 is a "premium" platform and AMD is the one providing the chipset and reference designs here, hence alledgedly keeping some level of control over the whole platform, also my motherboard was tested in store with another CPU and got pretty good reviews all over the web , RAM was in QVL and had a brand new, powerful enough and reputable PSU, meaning the "super easy" experience of buying a nearly $4k computer with every best component I could find ended up in a proper sh*t show that I spent the first 4 days with trying to fight against whea errors to then have a 1k CPU die on me ...

Many things that cause a CPU to fail within a week ? I've done some LN2 XOC back when I was a young guy, and appart from delidding gone wrong, forgetting to put back ln2 in the bucket or pushing really insane overvolts, killing a cpu in less then a week is pretty hard unless it's DOA (which did happen to me a couple times in datacenters too, so not saying these don't happen ofc, just that there's not so many things you can do while using a computer normally that can kill a CPU when it's well conceived...)

colesdav · ‎12-13-2020

RE: Widespread reports about 5950x" ? please don't make me laugh, how could anything about such a low spread cpu be widespread ?

Do you know how many 5950x have been produced and shipped to customers?
I would expect some noise in tech press from users complaining by now.
I would expect complaints Tech YouTubers who have completed builds using them if they failed.
Maybe you should get in touch with them and ask?

Which motherboard did you use with the 5950x, just out of interest?

Thanks.

tatane · ‎12-13-2020

I don't know how many were shipped, but with official resellers still struggling to fullfill day one orders (I'm speaking of an order passded at the same time as my amd.com order, so part of what should have been the reseller's first batch of stock that's still not arrived ...), AND AMD not being able to source one as an rma replacement for another 2 to 3 weeks from now, i'd say they really haven't produced or delivered much, and people who don't have another computer they can use to check the forums would be ... well ... stuck with no computer to tell people about their bad experience ...

noise in tech press ? where is the noise about all those WHEA errors when they are indeed very much present for quite a lot of people with 5000 series ? tech press is strugling to survive and they need to keep getting ads and product samples to do so, that means they're unlikely to say anything about it unless there's something as notable as a class action or a huge drop in AMD stock ... Youtubers are same as press, any reputable youtube won't attack amd because they want to keep receiving sample of their cpus and gpus in the future and not have to buy them with their own money to show people how nice they overclock when you have a golden engineering sample

get in touch ? welp, seeing at people reports on RMA, mine surprisingly went litteraly "no questions asked" ... this could be explained two ways :

- It's a very expensive CPU and therefore they decided to provide a decent customer service (well, if that was really the case, they'd offered cross ship rma and would have some spare reserved for that)

- It's a widespread problem and they kinda automate the procedure of rma'ing those (quite more likely, then any spare are indeed already sent and they're struggling to renew rma stock)

Motherboard was an MSI x570 Tomahawk, possibly not the best mobo around but known for having pretty good VRMs ...

colesdav · ‎12-13-2020

RE: and people who don't have another computer they can use to check the forums would be ... well ... stuck with no computer to tell people about their bad experience ...

I am pretty sure most people have laptops or smartphones or Work PC they could use to report problems.

RE: noise in tech press ? where is the noise about all those WHEA errors when they are indeed very much present for quite a lot of people with 5000 series ? tech press is strugling to survive and they need to keep getting ads and product samples to do so, that means they're unlikely to say anything about it unless there's something as notable as a class action or a huge drop in AMD stock ... Youtubers are same as press, any reputable youtube won't attack amd because they want to keep receiving sample of their cpus and gpus in the future and not have to buy them with their own money to show people how nice they overclock when you have a golden engineering sample

Tech YouTubers do speak out these days. The large ones at least. And they defend smaller YouTubers and Tech Press - Gamer's Nexus usually but just yesterday:

https://community.amd.com/t5/graphics/nvidia-told-to-sit-on-youtube-reviewers-quot-naughty-step-quot...

I do think they do get "golden engineering samples" of CPU and GPU for their day one reviews from AMD/Nvidia/Intel.
Some like Hardware Unboxed or Gamers Nexus do go out and buy GPUs/CPU's from WEB/Store to avoid that issue if they can.

RE: get in touch.
I meant you should get in touch with Gamers Nexus or other Tech Reviewer site and ask them if they have heard of problems with Failing AMD CPUs / WHEA error etc. Even if they have not - they could ask people watching theiur videos to report such issues.

Same thing happened about AMD Drivers on RX5700XT.
Reviewers claimed they were "not aware" of many driver crashes, but asked the Community.
That noise meant AMD took notice and improved their Drivers - after many months of Users complaining on here, and NOTHING being done to fix it.

They only seemed to respond to bad press coverage that may affect sales, not complaints from their own users.
Quelle suprise.

tatane · ‎12-13-2020

But why should all of that be needed for a company to acknowledge about an issue ? this whole discussion between us started on your allegation that building a Zen2/3 platform avoiding those WHEA errors was "very easy", sorry but you have 50 combined years of experience building PCs on this thread telling you it's not and you getting one that works OOB pretty much means you should buy a lottery ticket mate ;)

My aim here is not to try and destroy AMD sales but to get them to fix their sh*t as I ordered 2 of those 5950x i need for work purposes and would rather not have to spend the next 7 months debugging the platform as @mackbolan777 had to ...

kingfish · ‎12-13-2020

Keep in mind that this is a user-to-user forum. No engineers here. To report the issue where it can draw the attention of AMD and possible fix...... https://www.amd.com/en/support/contact-email-form

Everyone with this issue should report it. "A squeaking wheel gets the grease"

tatane · ‎12-13-2020

And if you read a couple posts up you'd see that my CPU is under RMA, so i obviously did contact amd support about it ;)

edit : also commented on how surprised I was by them not asking much/any questions on details about how it failed / errors I had before it failed etc ...

kingfish · ‎12-13-2020

@tatane wrote:
And if you read a couple posts up you'd see that my CPU is under RMA, so i obviously did contact amd support about it ;)
edit : also commented on how surprised I was by them not asking much/any questions on details about how it failed / errors I had before it failed etc ...
Well consider my post directed to those who have not reported it. Obviously.
One report is not going to get a lot of attention...you know....pebkac. Several reports of the same issue gets their attention.
So instead of talking to each other, involve AMD. It's the only way a issue, if there is one, will be resolved.

@

colesdav · ‎12-14-2020

But why should all of that be needed for a company to acknowledge about an issue ? this whole discussion between us started on your allegation that building a Zen2/3 platform avoiding those WHEA errors was "very easy",

What allegation? Building Zen2 platform is certainly easy. I have had no problems as long as I read reviews and select good Motherboard and PSU.

I have had no problems with Zen 2 CPU at all.

RE: sorry but you have 50 combined years of experience building PCs on this thread telling you it's not and you getting one that works OOB pretty much means you should buy a lottery ticket mate

You are both getting on like a pair of old drama queens if you ask me.
You have both been unlucky and have problems with Zen2 or Zen3 processors for some reason.
So work out what the problem is with your motherboards, or BIOS, or Windows installation or PSU or CPU or GPU.

RE: My aim here is not to try and destroy AMD sales but to get them to fix their **** as I ordered 2 of those 5950x i need for work purposes and would rather not have to spend the next 7 months debugging the platform as @mackbolan777 had to ...

I didn't say it was, but I am telling you that complaints to Gamers Nexus was the only thing that got listened to in the end to get AMD Radeon to attempt to fix GPU issues on RX5700XT, after about 8 months of problems, and also on RX Vega 64/56 GPUs as well which had GPU driver problems for even longer. Bugs were reported, emails were sent, posts were made on the forum. Nothing got fixed. Talking to Press or YouTubers should not be needed, but that is what happened.

tatane · ‎12-14-2020

RE: I have had no problems with Zen 2 CPU at all.
Good for you, why attacking people who are having problems then ? there's a 16 pages long topic on this very forum about Zen2 WHEA issues ...

RE : You are both getting on like a pair of old drama queens if you ask me.
oh, we're going with personal insults now ? great ...

RE: You have both been unlucky and have problems with Zen2 or Zen3 processors for some reason.
Look at the posts in this forum and all over the interwebz , seems you're the lucky one ...

RE : So work out what the problem is with your motherboards, or BIOS, or Windows installation or PSU or CPU or GPU.
sure, AMD acknowledging the problem being my CPU straight away, no question asked, means i now have to defend its not the motherboard, PSU or GPU to ~~fanboys~~ forum members ... absolutely wonderful
It would also be nice for AMD / people around here to realize all of us are not PC parts shops nor have 10K to invest in solving an issue just for the sake of staying with AMD ... We do test all we can with what we have at hand, but in the end of the day AMD is the one having the most control over the whole platform since they make both the CPU and the Chipset ... If it doesn't work as expected, how are we supposed to test a dozen different motherboard/ram/gpu combos from different brands especially when reports about WHEA errors come from all kind of brands/models anyway ? do YOU have the black voodoo magic formula to make it work ? if so i bet ya AMD would be keen on hearing it too rn ...
Now you're likely going to ask how AMD is supposed to test all those combinations ? I'm just asking you how are WE ? AMD makes billions out of selling us these components, we don't make any money by doing AMD's QE dept job ...

colesdav · ‎12-14-2020

RE: Good for you, why attacking people who are having problems then ?
I am not attacking anyone.

RE: oh, we're going with personal insults now ? great ...

No.
Just commenting on they way you are both getting on.

1. Claiming AMD can't make CPUs that work at all.
2. Telling me "my accusation that Zen2 works".
3. Wah wah Intel won't have this problem.

You know, its all a bit, well, dramatic.

RE: sure, AMD acknowledging the problem being my CPU straight away, no question asked, means i now have to defend its not the motherboard, PSU or GPU to ~~fanboys~~ forum members ... absolutely wonderful

OK, so AMD RMA process was quick and easy, isn't that a good thing?

How is suggesting you contact Tech YouTubers, like Gamers Nexus, and the Tech Press to ask about your problem, since you get no resolution or support from AMD on here quickly enough the actions of ~~fanboys~~.

@mackbolan777 has accused me of working and advertising for Nvidia for goodness sake.

You both seem to have "lost the plot" in your anger that you cannot get your PC Systems to run stable, for whatever reason.

There may well be something wrong with your CPUs.
I guess AMD will test them during RMA process to find out.

You never know, you might get your original CPUs back.

tatane · ‎12-14-2020

welp, moderators keep removing my answer (no rude language or anything tho ...)

Guess that's just about settles it for me and my future with team red ...

mackbolan777 · ‎12-14-2020

@tatane

I'm only responding to you because too many "cooks in the kitchen spoil the soup". To be clear, AMD has made good things prior like the Bulldozer/Piledriver CPU's were great in my opinion as far as setup, over clocking, not having random errors. These Ryzen's have been problematic more so with the Zen 2/3 than the Zen+ 2700X that colesdav insists on using to test an issue present on a Ryzen 3600X with an X570 board. He could lop off 3 cores, over clock, under clock, it's apples to oranges to try to compare or replicate this issue.

YouTuber's are helpful as far as reviews go, however they mostly get "cherry picked" or "binned" units to test. Gamers Nexus admits to that as well as when occasionally, Steve will buy one from a store but he makes that clear. Complaining to any of these does nothing, as AMD only cares what the Quarterly Report says. If it's down, they care. I looked up their 3rd quarter financial report and I can tell you they could care less about what any of us think or feel.

I could not find the recent article that said exactly how many 5000 series CPU's were sold but it was around the 1k mark. so not so many to have all these "error 18" type complaints and RMA's. AMD has always just RMA'd a CPU and unless the lid is blue from heat or has bent pins, they just send another. There is no fixing a CPU and little way to bench test it to trigger the same error. Nothing on die stores fault codes or data, so it's a moot effort. They send you a new CPU and call it a day.

@kingfish says to complain via the support form. That results in an RMA with pretty much no questions asked. AMD really doesn't look into the issue and just moves on. That behavior will eventually catch up to their corporate wallet and when Intel finally exceeds them once again, they will begin to care. Think possibly Q2, 2021.

Could it be a microcode issue? Sure, I found a few unstable BIOS's from AsRock and only this 3.20 seems "ok". This is totally an AMD issue as I found only one complaint that was recent about a code 19 (Cache Hierarchy Error) from 2017 on Intel's forum which is run by Intel employees. That resulted in CPU replacement after the OP replaced ever piece of hardware including the board first. An earlier complaint with error 18 was from 2015and the fix was to bypass the Spectre/Meltdown patch based on one's own risk assessment on a Dell forum.

The randomness of it indicates it could be a microcode issue or like I said a glitch in the way the memory controller works. Whatever it is can cause one to lose the OS or data so it's not "ok" to let it slide. Why does AMD need to come out with a new Agesa every 2 months? Most people never need to update their BIOS except for security flaws.

Everyone has their preferred RAM test, I like HCI Memtest64. It tests the CPU cache, the board's memory, and the RAM. I let that run all night more than once to well over 1000% coverage with zero errors. This other RAM test colesdav talks about does no good if the PC slams/reboots, it's not going to tell me a thing.

Yes, this is the AsRock board I complained about having a poor VRM and it does by design, however it's not at fault here. A brief address of the accusation of colesdav advertising for Nvidia was he posted a video link to a comparison that shows the 2070 beating the 5700XT on this forum, which is inappropriate. You can know in your heart that Nvidia may in fact beat AMD in some games or some tests using certain CPU and other hardware but don't publish it on the forum. I didn't place links on here showing Intel slamming AMD's Ryzen CPU in any test. I said Intel doesn't have this error 18 issue and they don't. They do have an issue using an AMD GPU that results in a screaming sound/green screen/reboot. It seems like that can be fixed using the 20.8.3 fix I posted before, as told on Ancient Gameplays.

Colesdav turned that into a heated trolling with a vengeance to prove AsRock boards are good. Calling me a "liar" and so I avoid that OP. Why he/she insists on testing for a failure issue using a Zen+ 2700X is beyond me. Apples to oranges technology. Also claiming to use muti-GPU's when he "reprimanded" me in post where I "hypothetically" addressed someone using 2 AMD GPU's in their build vs. using one Nvidia and one AMD at the same time. At any rate, OP indicates they are still angry over my quoting the action.

Back to the error 18 thing. AMD did offer people who bought a certain APU made at a certain date over the summer a new APU due to an SMM attack issue. The rest of us got an Agesa update. So memory controller issue is still high on my list of what causes the CPU specific errors. Otherwise, why was I able to play BF4 for 3.5 hours last night and it ran flawless? Why did the CPU pass Cinebench R 23? RAM is either good or bad, there's no "it might be". Silicon, that's different. If AMD went with hard soldered memory controller in the die vs. "Fabric" this might possibly end the issue. I understand the IF is for flexibility in performance but if it is unreliable or can degrade quickly, than it may not be a wise idea.

No matter how my issue gets resolved and @tatane 's, AMD needs to do better. I could not sell an AMD nor recommend any of their current products to a customer. I would add a disclaimer and offer no warranty if they insisted on me building an AMD unit. All one needs to do is a Google search and there's many forums about AMD, like it's a "cult" of people with problems all trying to figure out the enigmatic issues of AMD Ryzen, mainly Zen 2/3 CPU's. The rest are exploring the GPU's. The blind leading the blind, because there really is no definitive answer. Which makes me hesitant to even just save a buck and buy another 3600X while I RMA this one. Sure I have a lametop, but I really would prefer to have my rig running.

A week to failure, 7 months, even a year is too short of a time span for CPU life. With all the built in code to keep the CPU from drawing more than the allotted AMPs, making "all core oc" not worth it, the only way to surely kill a Ryzen is to change the Vcore too high. It's a "red flag" when they brag about PBO but if you turn it on it pops a warning up that it will void the warranty. Why sell it as "unlocked" then? Why advertise running 3733 RAM and 1866Mhz IF as the "sweet spot", then recant that to 3600 RAM and 1800Mhz IF because 90% of the CPU's can't do more than 1800. Even the 5000 series they tout as running 2000 IF and 4000 RAM, not a chance. Lucky to hit 1900 IF and 3800, most fall back to 3600 and 1800 IF. All that extra money for a CPU supposedly ahead of it's time. Perhaps these Ryzen's are ahead of Intel in many ways, but they need to be running to win. Would be pretty embarrassing to go head to head against an Intel only to crash mid test while live streaming the event. It probably has happened and is deleted so no one ever sees it.

I'm sure someone will point out the tech spec of the 3600X or non "x" is 3200Mhz RAM, so anything greater is an OC. Why then does AMD advertise the higher speeds or even offer an "X" model if the performance isn't there consistently? Sounds like "bait and switch". I buy the Ferrari and pop the hood to find it has a 4 cylinder Ford engine in it and they say "well in the fine print the engine may vary".

Hmm, looked up the competition and they allow any OC up to the spec of say 5Ghz as long as voltage is within spec AND if one wants to get "crazy", you can buy a PTPP for $20 that covers anything related to an OC. Stick it in a deep fryer, freeze it with liquid nitrogen, sky's the limit.

As far as the 5950X's being dead in a week or DOA, that's super bad because that's the flagship. However, it is par for the course with Ryzen's. When I built this, I nearly went Intel because of the complaints on Newegg about DOA's, bent pins out of box, not running right, sudden reboots, all this stuff. But I thought with my experience it must just be "noobs" slapping things together. Plus AMD had advertised all this performance above the Intel of the day. Sure it can run circles around some intel's but it needs to be running. Not noobs. I was wrong. No reliability in the Zen 2, now the Zen 3. The new video cards from the 5000's to the 6000's have issues. Why should we maintain confidence in AMD at this juncture?

I don't know about most people but personally I like to built it and forget it. Not build it, diagnose it, rebuild it, diagnose it, replace hardware to see if that's the cause, re install the OS 20 times, etc. Spend my life on this forum typing to the wall. Less time at board meetings more time in R&D, AMD.

"It worked before you broke it!"

colesdav · ‎12-14-2020

RE: These Ryzen's have been problematic more so with the Zen 2/3 than the Zen+ 2700X that colesdav insists on using to test an issue present on a Ryzen 3600X with an X570 board. He could lop off 3 cores, over clock, under clock, it's apples to oranges to try to compare or replicate this issue.

I tested BF4 on a Ryzen 2700X because that machine has an RX5700XT on it.
Zen+ memory controller is much worse than Zen2. Zen2 builds are easier because of that.
Now you know that it is not a BF4 crashing problem at least on Zen+

I have a Zen2 build with MSI Tomahawk Motherboard and Nvidia RTX2080OC.
That has been tested as well and no crashes on that machine either.

I can swap another RX5700XT into that machine and see what happens some time next week.
I doubt there will be a problem.

Like the way you buried more of your nonsense in that wall of text.
I do not care what you say any more, but I will not let you get away with it.

RE: Yes, this is the AsRock board I complained about having a poor VRM and it does by design, however it's not at fault here.
How do you know? You already said it was useless in your previous post.

RE: A brief address of the accusation of colesdav advertising for Nvidia was he posted a video link to a comparison that shows the 2070 beating the 5700XT on this forum, which is inappropriate.

Wrong again. OP already had an RTX2070 Super and you recommended they buy an RX5700XT which is a slower GPU overall.
I just sent a publically available benchmark video showing that.
There is nothing wrong with doing that .
You accused me of advertising for Nvidia which is a lie.

RE: Colesdav turned that into a heated trolling with a vengeance to prove AsRock boards are good.
More lies from you.
This is what happenend.

Someone asked for advice for a motherboard for an ITX build on a different thread. I pointed to one from Asrock.
You disagreed, told me ASRock Motherboards were useless, and told me that LevelOneTechs had said no one should use them.
The problem for you was, LevelOneTechs had actually done a mini ITX build, using the very ASRock board I recommended, and gave it a glowing review. I posted a link to the video on the thread.
Try telling the truth for once.

RE: Calling me a "liar"
You are a liar. You are lying again in the above post.
You do know that I can just provide a link to the post and people can see what was said?
What is wrong with you?

RE: Also claiming to use muti-GPU's when he "reprimanded" me in post where I "hypothetically" addressed someone using 2 AMD GPU's in their build vs. using one Nvidia and one AMD at the same time

You advised someone to buy two RX5700XT to use them in Crossfire for Gaming.
All I did was to respond to the post and say that DX11 Crossfire is no longer supported by AMD and even DX12 MultiGPU support is rare, and poorly implemented in about 50% of those Games.

RE: Also claiming to use muti-GPU's
Yes - I have MultiGPUs on the PC I tested. I use them for Blender and Compute, two of the reasons I stated you "might" want to buy more than one AMD GPU and fit it on your system. I did not buy them for DX11 Crossfire or DX12 MultiGPU.
Having so many GPUs on that PC apart from the Primary RX5700XT means the Driver should be having more to deal with, and more electrical noise on the system meaning a crash should be more likely. The PC did not crash during long tests of BFV and BF4.

RE: At any rate, OP indicates they are still angry over my quoting the action.
I am angry about being accused of promoting Nvidia, when all I have been doing is posting benchmark results from independent reviewers.

colesdav · ‎12-14-2020

Just to clarify.

No Crashing on machine with the following specs:

Specifications

CPU	AMD Ryzen 5 3600 3.6 GHz 6-Core Processor
Cooler	Wraith Prism RGB LED.
Motherboard	MSI B450 TOMAHAWK ATX AM4 Motherboard
Memory	Corsair Vengeance LPX 16 GB (2 x 8 GB) DDR4-3200 Memory
GPU 1	Palit RTX 2080 OC

Disc Drive 1	Team L5 LITE 3D 240 GB 2.5" Solid State Drive
Disc Drive 2	Seagate FireCuda 2 TB 3.5" 7200RPM Hybrid Internal Hard Drive - additional storage for "Main Drive".
Disc Drive 3	Seagate 5TB USB drive to store games.
PSU	Corsair RM 750X Gold.
Case	Kolink Horizon RGB
Monitor	Primary Monitor = BenQ 32 inch EW3270U 4K HDR monitor. Secondary Monitor = Acer 1080p 60Hz monitor.

I could replace the RTX2080OC with an RX5700XT some time next week maybe but I guess that would not be good enough.

This build has been absolutely rock solid on many other modern AAA games as well.

pokester · ‎12-15-2020

@colesdav wrote:
Just to clarify.
No Crashing on machine with the following specs:
Specifications
CPU AMD Ryzen 5 3600 3.6 GHz 6-Core Processor
Cooler Wraith Prism RGB LED.
Motherboard MSI B450 TOMAHAWK ATX AM4 Motherboard
Memory Corsair Vengeance LPX 16 GB (2 x 8 GB) DDR4-3200 Memory
GPU 1
Palit RTX 2080 OC

Disc Drive 1 Team L5 LITE 3D 240 GB 2.5" Solid State Drive
Disc Drive 2 Seagate FireCuda 2 TB 3.5" 7200RPM Hybrid Internal Hard Drive - additional storage for "Main Drive".
Disc Drive 3 Seagate 5TB USB drive to store games.
PSU

Corsair RM 750X Gold.

Case Kolink Horizon RGB
Monitor Primary Monitor = BenQ 32 inch EW3270U 4K HDR monitor.
Secondary Monitor = Acer 1080p 60Hz monitor.

I could replace the RTX2080OC with an RX5700XT some time next week maybe but I guess that would not be good enough.

This build has been absolutely rock solid on many other modern AAA games as well.

Did you load the latest bios? Mine has been rock solid but I did get a crash within 2 hours on the latest bios. I got home last night after reverting and still no crash.

colesdav · ‎12-14-2020

RE: Everyone has their preferred RAM test, I like HCI Memtest64. It tests the CPU cache, the board's memory, and the RAM. I let that run all night more than once to well over 1000% coverage with zero errors. This other RAM test colesdav talks about does no good if the PC slams/reboots, it's not going to tell me a thing.

RamTest provides a log file describing when test starts, when it ends, how much coverage and if an error is detected.

If you can use HCI MemTest 86 to do the following then fine, use it.

Did you bother to run FurMark or a game at the same time as running a memory test in Windows and check to see if electrical noise from the GPU activity was causing RAM errors?

I guess not.

"RE: Games are the only trigger.
Did you run Furmark Stress test on the GPU?
Does that crash?

Here is one test I would suggest.

Purchase Karhu Ram Test and run it in the background whilst you are running a game on the GPU.
You can adjust the amount of RAM Tested - make sure you leave enough Free RAM, apart from that used by the game and in the test.
See if it reports any Memory Errors whilst you are gaming with the GPU Active.
It could well be memory errors induced due to electrical noise from the GPU - either directly or because the peak power pulled by the GPU is causing noise on the PSU rails.

You can get Karhu RAM Test from here:
https://www.karhusoftware.com/ramtest/
"

colesdav · ‎12-15-2020

I have contacted the following reviewers.

Gamers Nexus.

OC3D.

Hardware Canucks.

TechPowerUp.

They all reviewed the 5950X.

I asked if they have problems with those or other RX5000 series CPU.

One of them might answer some time.

deepcuts · ‎12-13-2020

@colesdav

Try https://www.overclock.net/threads/replaced-3950x-with-5950x-whea-and-reboots.1774627/

I would expect some press coverage also, but I guess reviewers got cherry picked samples. Cannot find another explanation.

mackbolan777 · ‎12-13-2020

@tatane Thank you for that response! Nail on the head. We as customers, some of us professional builders, like myself, should not need to use "black magic", or anything other than properly assemble the correct parts and fire it up. If you were "unlucky" and got a bad chip, well so did the "I lost count" number on this forum complaining about the 5000 series right now and the "arm twisting", gut wrenching, effort to solve an enigma with no clear solution or even a foggy one. The 3000 series is just as plagued. We hear about a Zen+ that another OP managed to get working after what seems to be much angst, defending AMD.

Fortunately in my 20 some years of building mostly AMD PC's for customers, even through the AM3+ days of uncertainty, I had fairly good fortune. The old Bulldozer and Piledriver CPU's seemed to run great with an all core OC and 1866Mhz RAM. In fact video drivers were the big beef then. Now there are so many complaints regarding how to RMA both the 3000 and 5000 series CPU's due to error 18 and other related failure codes, it's epic. My CPU is only 7 months young, it's a baby. It should last the 3 years it's warranted for and longer with an AIO on it. Reports show AMD Ryzen's are expected to last 5 years before the silicon is done. Intel CPU's are rated for 10 years before a critical failure.

It's not software now. Windows is updated to version 20H2, no longer the 2004 Beta. The video driver has been sorted after 6 months of playing with that. This is a fresh install of Windows 10 x 64 Pro on a very new Seagate Compute 1TB SSD. Sure games can crash for a number of reasons but it does not normally cause a screaming audio noise for 5 seconds, with a black screen, then reboot. The whole process took about a solid 90 seconds, like it was a person dying a painful death. Normal game crashes cause the game to go to desktop and that's about it, one might lose the video driver or even Explorer making one restart. Then the error will show as "app crash" or "driver package failed, blah,blah". Not Processor Core failure or Processor Cache Hierarchy Error.

Every component on this has checked out down to the much hated AsRock board. I can eliminate that since there are so many others that have the same error coming up on Zen 2 or 3 on any board.

As far as OC goes, running PBO enabled, with SOC at 1.10v, board doesn't have LLC and the RAM is running underclocked at 3600Mhz. It passes HCI Memtest64 to 800% coverage, passes Cinebench R23, can run Prime 95 forever, no OCCT errors. Toss it in a game and after 1-2 hours this may or may not happen, it's random. Games are a mixed load, so that's the only thing that triggers this issue. I think it has a lot to do with how AMD designed the memory controller.

\I can and will RMA this CPU but whose to say it won't do it again? I might be stupid and buy a 3800X because that's the only one I haven't heard a complaint about, yet, just to see if it too does this. The "little voice in my head" tells me to run from AMD, spend $520 on an i9-9900K and Z390 board and be done. Because the 3800x is $400, a more than likely troublesome 3700x is $349 and no point on a $249 3600X while awaiting the RMA CPU to be shipped.

Spending another dime on AMD seems like burning money or the definition of insanity (trying the same thing over and over expecting a different result). However, funds for a new CPU while waiting is easier to come by and I'm half curious if it's really a bad CPU or the design itself.

I've been through 7 months of diagnostics on a new build is the point. Never was right, never will be it seems. Every post I read in any forum where the OP claims CPU replacement solved the problem, they write back in a few weeks to months saying there's a problem again. When I built my i7-8700K @5.0Ghz system in 2018 all new, $768 Nvidia 1080i, $1500+ spent, want to know how many times I had to visit a forum, contact Gigabyte, Team Group, Intel or diagnose anything? ZERO. This build ran me about $1100 and problems form day one.

I'm glad colesdav isn't mad anymore. We have differences that best be left alone. But although I'm sure he knows my specs, here they are again.

AsRock X570 Phantom Gaming 4S (BIOS 3.20), Ryzen R5 3600X, 16GB TeamForce Extreeme Gaming DDR4 3733 @3600Mhz CL16 1.35v., Seagate Compute 1TB SSD, Gigabyte Windforce RX 5600 XT, Corsair H110i GTX, Corsair RM1000, SATA DVD/RW

No overclock beyond PBO and the factory GPU OC. I can run at 2400Mhz RAM, disable PBO, underclock the GPU and it still happens. Games are the only trigger. All other hardware passes tests with flying colors, even the CPU. But yet it still fails while gaming randomly but will happen during any session of gaming at some point usually over an hour. Sometimes it won't "scream of death/reboot" and just goes to desktop but that's rare. no info in dumps, nothing in DXDiag, Nothing in MsInfo. HWiFO shows good voltage and temps until it crashes. Only errors are 18 and occasionally Processor Cache Hierarchy, that comes with a side of Bus Disconnect but recovered error.

If AMD can't build a CPU to run with PBO enabled, they should not sell it or take the "X" off the thing. The CPU is expected to run within spec. Besides, turning it off makes no difference whether the thing fails or not. This is very bad press for AMD, since this post came up number one in a blind Google search on this error code. They can't hide the problem anymore.

"It worked before you broke it!"

colesdav · ‎12-13-2020

RE: Sure games can crash for a number of reasons but it does not normally cause a screaming audio noise for 5 seconds, with a black screen, then reboot.

I have many videos and posts complaining about that very problem on both Intel and AMD CPU with Radeon GPU Drivers.

colesdav · ‎12-13-2020

RE: Every component on this has checked out down to the much hated AsRock board.

Is that the ASRock board you were complaining about that has the bad VRM?

colesdav · ‎12-13-2020

RE: I'm glad colesdav isn't mad anymore,

I am still angry at being accused of Promoting and Advertising Nvidia on an AMD Forum just because I post Tech Youtuber reviews. I disagree that the RX6800/XT 6900XT GPUs are better than Nvidia RTX3000. I pointed out , AMD GPU are way behing in compute, Blender, and productivity applications.

Are you now Adertising for Intel on an AMD forum because you complain about AMD CPUs?
I do not think you are but it is the same argument you used.

colesdav · ‎12-13-2020

RE: Games are the only trigger.
Did you run Furmark Sress test on the GPU?
Does that crash?

Here is one test I would suggest.

Purchase Karhu Ram Test and run it in the background whilst you are running a game on the GPU.
You can adjust the amount of RAM Tested - make sure you leave enough Free RAM, apart from that used by the game and in the test.
See if it reports any Memory Errors whilst you are gaming with the GPU Active.
It could well be memory errors induced due to electrical noise from the GPU - either directly or because the peak power pulled by the GPU is causing noise on the PSU rails.

You can get Karhu RAM Test from here:
https://www.karhusoftware.com/ramtest/

lakesideguy · ‎05-12-2021

It's not AMD GPU or drivers...sigh....

I get the WHEA processor cache hiearchy error with Nvidia 1070Ti,

So much for that theory.

MacNZ · ‎12-13-2020

@mackbolan777 wrote:
Ok so weeks since fixing the IF running at 1866 to match my 3733, so called , "sweet spot" RAM on my 3600X build. I stopped having errors about Processor Core Failure/Hierarchy Error, and the usual complaints on here about the Ryzen CPU's in general. Switching to the Adrenaline 20.8.3 driver seemed to stop the error relating to that driver crashing.
Tonight after about 2 hours into BF4, the PC screams, black screens and reboots. The error: WHEA LOGGER 18 Processor Core Failure. So I'm done with AMD at least as far as CPU's go, just done. I looked for a single Intel with this error and found one in 2017 but it was error 19 and that resulted in CPU replacement after replacing everything else first.
Same song here but it's happening to every model CPU from the Ryzen 3000's to the new Ryzen 5000's and now the only thing "Ryzen" is my blood pressure! 7 full months of fighting this machine and I've had it. Trying to help others on here with similar issues when in fact there is no cure, not from AMD. All these "quirks" with RAM, board topology, the "uniqueness" is a cover for AMD's failure to make a working product. Hundreds of thousands of posts about the same failures and errors all over the internet and yet AMD still keeps pumping out this problematic hardware.
Sure I'll RMA the CPU but I'm done with AMD CPU's until I see they actually fixed something right. $1100 worth of a problem, bunch of you know what. This whole forum was designed to catch the flak for AMD. Like getting one CPU to work properly is finding a unicorn or a pot of gold at the end of a rainbow. For $1100+ No thanks. The RX 5600 XT is probably fine but we'll see after I switch CPU teams.
Yep, this will make a few of you guys happy that I'll not be in here very often if at all. At least I won't be wasting my time or money on trash that breaks every hour. AMD's RMA'ing Ryzen CPU's under a year old some less than a week, by the HUNDREDS, what's wrong here???? I'm glad I folded my business in 2012, because my customers would be banging on my door everyday since I built 99% AMD. I'd need to build a "forum" to hide behind too.

Don't give up. Try the newest one: 20.12.1 (WHQL). I installed the latest BIOS, just out a few days, from Gigabyte (Aorus Elite V1). BIOS F31o, which is updated AMD AGESA 1.1.0.0D. Still a beta, but NOT a C revision. (I read your post on "C" with great interest and immediately jumped on this new BIOS when it became available.

This ended WHEAs for me. Been running for 3 days, with all types of use. Memtest, CPU Load Test with CPU-Z, plenty of idling, and ordinary uses (Office, browsing the Internet, etc.). Rock solid with XMP and PBO enabled.

IMHO, this is NOT a hardware issue, it's a microcode issue (BIOS). The APIC (Advanced Power Interrupt Controller) ID is the core where the Cache Hierarchy corruption occurs, and this has to do with processor C-states. That's why a lot of the reported issues are under light load or idle.

Hope that helps.

Another Mac!

colesdav · ‎12-14-2020

I just tested BFV on a Rzyen 2700X machine with 64GB of Corsair 3200MHz RAM and a primary RX5700XT attached.
I also have another 6 AMD GPU cores hanging of that machine, so in terms of GPU noise it is about as bad as you will get, apart from running a pair of RX Vega 64 Liquids in DX11 Crossfire or DX12/Vulkan MultiGPU.
I have a total of 3.1 KW of PSU.
The main PSU for the PC and GPUs in the PC case is a Corsair AX1600i.
There are another pair of Corsair SFX 750W PSUs for external GPUs.
All GPUs are active, deliberately, to give the AMD Adrenalin 2020 20.12.1 Driver "hell".
I am running Windows 10 Pro 2004 with latest patches.

I tested BFV for 4 hours at the following settings:

2 hours with CPU running at 3.7GHz andf RAM at 2133MHz.
4K Ultra 60FPs with Radeon Image Sharpening on and I also even tried turning on enhanced Sync. CPU Utilization ~ 42%
1080p low. VSync off. That gave me about 166 FPS with 76% CPU Utilization.
Not a single problem. No crashing, no anything.

2 hours with CPU running at 4.3GHz and RAM at 3200 MHz.
4K Ultra 60FPs with Radeon Image Sharpening on and I also even tried turning on enhanced Sync. CPU Utilization in similar range ~ 42-45%
1080p low. VSync off. That gave me a solid 200 FPS with 79% CPU Utilization.
Again not a single problem. No crashing, no anything.

I know you ran BF4, I did not have that installed.
It is downloading to test now.
There may be something wrong with that specific game.

I will soon find out.

colesdav · ‎12-14-2020

Battlefield 4 Premium now installed.

Please let me know if there was a particular area of the game you would like me to test.

I have some work to do and I will be testing in it about 8 hours.

Thanks.

pokester · ‎12-14-2020

@mackbolan777 wrote:
Ok so weeks since fixing the IF running at 1866 to match my 3733, so called , "sweet spot" RAM on my 3600X build. I stopped having errors about Processor Core Failure/Hierarchy Error, and the usual complaints on here about the Ryzen CPU's in general. Switching to the Adrenaline 20.8.3 driver seemed to stop the error relating to that driver crashing.
Tonight after about 2 hours into BF4, the PC screams, black screens and reboots. The error: WHEA LOGGER 18 Processor Core Failure. So I'm done with AMD at least as far as CPU's go, just done. I looked for a single Intel with this error and found one in 2017 but it was error 19 and that resulted in CPU replacement after replacing everything else first.
Same song here but it's happening to every model CPU from the Ryzen 3000's to the new Ryzen 5000's and now the only thing "Ryzen" is my blood pressure! 7 full months of fighting this machine and I've had it. Trying to help others on here with similar issues when in fact there is no cure, not from AMD. All these "quirks" with RAM, board topology, the "uniqueness" is a cover for AMD's failure to make a working product. Hundreds of thousands of posts about the same failures and errors all over the internet and yet AMD still keeps pumping out this problematic hardware.
Sure I'll RMA the CPU but I'm done with AMD CPU's until I see they actually fixed something right. $1100 worth of a problem, bunch of you know what. This whole forum was designed to catch the flak for AMD. Like getting one CPU to work properly is finding a unicorn or a pot of gold at the end of a rainbow. For $1100+ No thanks. The RX 5600 XT is probably fine but we'll see after I switch CPU teams.
Yep, this will make a few of you guys happy that I'll not be in here very often if at all. At least I won't be wasting my time or money on trash that breaks every hour. AMD's RMA'ing Ryzen CPU's under a year old some less than a week, by the HUNDREDS, what's wrong here???? I'm glad I folded my business in 2012, because my customers would be banging on my door everyday since I built 99% AMD. I'd need to build a "forum" to hide behind too.

I really am sorry for your issues and very much understand the frustration. I do think this is a very real issue and obviously so by the large number of complaints in these forums as well as others such as reddit etc...

For some reason the problem absolutely has several triggers that seem to make it worse. One being the combination of Navi GPU and Ryzen CPU seems to be most common. Many have this issue though even paired with Nvidia as well. This error does pop up on Intel systems but rarely, as it is a very generic error that can even be influenced by Malware. However the current situation really appears to be the AMD platform being way over the normal range of problems with this error, IMHO. I work with hundreds of computers at work and have only seen this error on them a couple times and they are all Intel machines. One case was a system with a bad ram chip the other was a highly infected work station.

Many also find relief as you have said yourself in finding more agreeable memory settings. I also think this issue is the least and 3200 or less on Zen2 and 3600 or less on Zen3. Based only really on what other users I have seen having said.

The one thing however that proves to me it is something software/firmware related is that many with Zen 2 processors that can regress bios have found that regressing to bios prior to late summer that predates the microcode released to recognize Zen3 seems to make the issue disappear for many. Unfortunately not all boards allow this regression or is that even an option for those now having this problem on Zen3.

While I agree with @colesdav that likely a better choice of motherboard may help alleviate the issues, I can hardly blame the end user. AMD needs to do a better job of validation of board partners then. Motherboards are very much not equal between the makers the layering how the address the ram slots the vrms all these things are very erratic on the AMD side of things and not something I ever find to the same degree on the Intel side. Regardless I don't ever like to lay the blame on the end user. Most users want to install their parts, run at defaults and have a pleasant experience. Unfortunately these days with AMD on the GPU side it is nearly impossible and for many they have to have special settings on the motherboard to be stable too.

I know you do service for a living and I am sure you fully realize how few end users know or want to know how to adjust under the hood of their computers. They just want to use them and have them work right.

I want to point out that after reading this yesterday and I had long been curious if this would happen to me if I upgraded to the lates bios on my board as I have 3 B450 boards and am sitting 2 bios revisions back. It tried the latest bios on one as a test. Within 2 hours of running a looping test I got the WHEA18 error. I then reloaded the old bios and began the test again. It is still running at home remotely as I am now at work, and I have checked on the machine a couple times now and it is still fine. I definitely did my homework in buying my board a MSI Tomahawk B450 and it is probably the best rated B450 board made. I now know that this board too, will get this error with the right or more accurately wrong bios. It also tells me that until this bios issue is fixed I will not upgrade to a Zen 3 chip on this platform. I highly suggest if you can to regress your bios and see if you have better luck. I know with some boards you can't do this.

I do agree though that it looks like the only way to get traction on these issues is for someone like Linus or Steve Burke to spotlight issues. This is a topic I have written Steve about several times now. He is a great guy and usually responds to the emails too. I hope that eventually they get this fixed for you and the MANY other users with this problem. None of you deserve this, regardless of how much research you did on your hardware purchase. What you deserve is for your hardware to work right.

colesdav · ‎12-14-2020

I tested Battlefield 4 Premium for 2 hours straight, and then some more.

I tested at DX11 4K Ultra and 1080p Low.
Same machine Ryzen 2700X, 64GB Corsair Vengeance RGB Pro 3200MHz DRAM, CPU run at fixed frequency of 4.3GHz. ASUS ROG Crosshair Hero WIFI. BIOS version 3103. GPU = a PowerColor RX5700XT Red Dragon no overclock but fan profile set to 100% rectangle and Power Slider maxed out to +50%. Another 6 AMD GPUs on the system, but left at default driver settings.

Windows 10 Pro 2004 updated with latest patches.
Adrenalin 2020 20.12.1 freshly installed following DDU.

I tested the following. With Enhanced Sync on/off, RIS on/off, Chill on/off, Ingame Vsync on/off and combinations thereof.

Campaign mode.
No problems at all.

Multiplayer mode on various servers.
No problems at all apart from one server with slow ping where the game just closed once.
No AMD GPU driver crashes at all.
No driver crash notifications.

I will keep testing and let you know if anything bad happens over next few days.

There is a new BIOS for that motherboard to support new AMD Ryzen processors, but I cannot afford any time to test that at the moment.

======================
Version 4007

2020/12/10 15.37 MBytes

ROG CROSSHAIR VII HERO (WI-FI) BIOS 4007
1. New CPU support
2. Offer a Re-size BAR Support option to enhance GPU performance.
3. Remove AMD 7th Gen A-series/ Athlon X4 Processors support
Before running the USB BIOS Flashback tool, please rename the BIOS file (C7HWIFI.CAP) using BIOSRenamer.
DOWNLOAD
======================

I might try that BIOS later this week, but just on the Ryzen 2700X.
Not sure if that will tell you anything though?
I also need to make sure with Asus support it is not a 1 way flash from 3103 to 4007.

pokester · ‎12-14-2020

Yes make sure you can flash back.

cmtdfritz · ‎02-25-2021

Ryzen 3800X running all Core Overclock at 4.4GHz 1.28 Volts, never had an Issue.

I suspect it is your Board and or the BIOS. I run a Crosshair Hero VIII.

mackbolan777 · ‎02-25-2021

3 months late and a nope. It was a Windows Update that they fixed in a few days. But now I've moved on to a 5600X and an RX 6800. Same board, RAM, but an all core 4.7 OC at 1.27v and the card is killer. But thanks for reading and welcome aboard!

"It worked before you broke it!"

PC Processors

WHEA 18 Again!!! I'm done!