cancel
Showing results for 
Search instead for 
Did you mean: 

Processors

Jazzinplayer
Adept II

System unstable: Ryzen 5 5600x and Aorus x570 crashing on games at default settings

Sorry for the long post....
My System:
Ryzen 5 5600x
x570 Aorus Elite Bios F 31h
Gskill TridentZ RGB 2x8 3600 mhz CL18
MSI 2080 Super Ventus
1 M.2 disc for Windows and 1 M.2 for games and data. 
Seasonic Focus 850W 80+ gold PSU

What a nightmare this "upgrade" has been. Many hours spent and many windows reinstall. Just having the PC to post was a big deal, Q flash never worked had to get a r5 2600 CPU from a friend to actually post and update bios. 

Everything looked fine in windows, after updating BIOS Installed chipset, audio and LAN drivers from Gigabyte website. I was running the system with all settings as default/auto. Turned XMP on and ran benchmarks, all good. Cinebench r 20 and r23 running good. Started a game (Rogue Company) and system crashed (Bios reset) within 5 min. Repeated and crashed. Opened Horizon Zero Dawn to run the benchmark and crashed within 20 sec.

Event viewer showed WHEA errors, kernel erros and more. Decided reinstall windows fresh (again) guessing something went wrong with it. New windows, installed same drivers, only downloaded steam and HWinfo, SAME problems of course.

Found out that if I disable Core Boost Performance from bios and enable XMP everything runs smooth and crisp (benchmarks and games). But my chip is running at 3.7 ghz. On the other hand, if I run my DDR4 at default 2133 mhz and Disable CBP I can All core boost 4.6 ghz 1.3 volts and be stable. Both scenarios give me similar scores in HZD benchmark.

Any sort of CBP and PBO values I do will end up crashing on HZD benchmark, regardless the memory speed. Running ryzen master for PBO and Auto clock delivers same results. After many hours of memory timings and overclocking managed to stabilize the system with: all core boost 3.9 ghz 1.2 V, all settings auto (LLC, SoC volts, etc.) and XMP on 3600 mhz CL 18 1.35v. 
This run smooth and crisp in games and benchmarks, allowing for best "stable" scores.

Obviously I'm not happy, besides the hard work and many hours invested to stabilize, this is not the performance I payed for. 
Should I get a new mobo, should I issue a RMA on the CPU, wait for Bios updates/better drivers?

0 Likes
1 Solution
Jazzinplayer
Adept II

Guys!!! I SOLVED it!!!!!!!!

Omg, this is actually upsetting but relaxing at the same time since I finally found the problem.

It was PSU related but not because of the total Watts, it was the cables. I've been using this Phanteks Cable extensions for over 2 years now, never had an issue. Since I was monitoring the watts usage and noticing very weird behavior,  like huge spikes and dumps, in one of many reboots I noticed once cable was pretty hot, that is not normal so I removed the extensions (CPU, Motherboard and PSU) and PROBLEM SOLVED. Running smooth like butter, ran all the benchmarks, ran prime 95, gamed for 2 hours and not a single problem. Giving me the expected performance now.

It makes a lot of sense by the way the PC was working. Still going for that 850 W PSU for more room on that department. Gonna keep testing but so far you can TELL is working as it should, it looks and feels fine. Thanks to everyone that helped. 

cables phanteks.jpg

View solution in original post

25 Replies
Jazzinplayer
Adept II

I just checked and there is a new bios F31j, it might solve the issues. Also drivers from AMD site are different (newer it looks) than Gigabyte site. Gonna try them out as well. Hope it works. 



 

Jazzinplayer, please always get your drivers from the individual vendors , never the MB vendor ( https://www.amd.com/en/support ).  They are almost always down level.  Please set all to default do a Clear CMOS and run Cinebench R23.  Please post a screenshot of Ryzen Master (RM) but not CB - just tell us the score.  Tell us how much power your video card uses.  I would let others test the new BIOS.  Thanks and enjoy, John.

This is the screenshot, everything default new Drivers installed. Score in CB R23 is 10583. It crashed on HZD benchmark just the same. 

Updated Bios to F31j and it crashed as well, same way. 

default everyhing no xmp cinebench 10873.PNG

0 Likes

Not good, Jazzinplayer.  What is HZD benchmark?  Google does not seem to want to say.  Why are all the cores at 4335 MHz?  Looks like something is forcing the frequency.  Can you borrow a power supply with more power?  What is the wattage of your video card?  If you have a dump file in your minidump folder, please compress the folder and post it here.  Please look at the Apps under Settings and see if you find any OC applications.  Thanks and enjoy, John.

0 Likes

Windows is fresh install, there is nothing in the PC other than Ryzen Master, Cinebench, HwInfo, Steam with Horizon Zero Dawn (the Benchmark is from that game) and drivers. 

Don't know why cinebench stayed at 4.3 ghz. When I ran the game benchmark it went up to 4.6 ghz. I can set the cpu to run 4.6 ghz all core at 1.3 and it works fine. No crashes on games but only if the ram stays at 2133 mhz, any other value it crashes on games. 

My power supply is fine, is top tier quality.  1 week ago I was running my i7 9700k at 4.9 Ghz with same components, that cpu under that load will consume twice as much as this one. Gpu consume around 300 W and was the same. 

I can run benchmarks like cinebench for +10 min with PBO and XMP and it'll pass. Ryzen master stress test easy pass. Open up 1 game say Rogue Company which is limited to 150 fps and it'll crash in minutes. 

0 Likes

Jazzinplayer, I am not criticizing your power supply, but if your Video card pulls 300 Watts, then the CPU may be getting shorted.  I would like you to try about 100 Watts more.  You should be able to run with Ryzen doing its own thing.  I know nothing about Intel chips, so cannot comment.  I have no other suggestions at this time - sorry.  Please do try AMD Online Support .    If your MB has an additional 12V power connections, please make sure they are connected,  Enjoy, John.

0 Likes

Thanks John appreciate your help! Going to contact AMD support as well.

Everything is connected properly and I can assure you the PSU is good enough for this build since the intel cpu was consuming much more power. I can Overclock this CPU to 4.6 ghz even 1.4 v and run stable, it won't crash if I keep my DDR4 at 2133 mhz. If setting something else then it crashes. 

0 Likes
spot_ys
Adept II

Hi!

As some members already mentioned, always get the drivers from AMD, this is really important and adds a great deal of performance and stability.

 

Another thing that might sound silly, but have a look at your PCIe x16 lane set speed on Motherboard, resolved many issues with my mates PC's by locking manualy the lane speed to what the either Board/CPU/GPU combo would allow.

 

In your case locked it to x16 Gen3

 

Keeps us posted on your findings ;)

Jazzinplayer
Adept II

I've spent many many hours now and nothing really works (besides base clock), I can stabilize the system even with 4.5 ghz all core and xmp. But will crash in heavy games and Bios reset. Only thing now is changing to another DDR4 but at this point I don't think will work. 

A friend of mine just got his and is running without any issues on a b550 Aorus board (I have a x570 aorus). The store where I bought the product does not have "Amazon-like" return policy so I might to issue a RMA and that is going to be a huge pain. If it was amazon I would've return it and buy a new one already. 

I could try to replace the motherboard but don't think that's the issue. Pretty sad right now. I was performing well on my i7 9700k and with all this hassle I regret this "upgrade" so much

 

0 Likes
mrsense
Adept II

Your power supply is fine.

Find and flash the bios with AGESA ComboV2 1.1.0.0 C but without SAM.

The SAM is causing a lot of headaches for many X570 & B550 owners.  SAM is useless without all 3 components; a Ryzen 5000, X570/B550, and RX6000.

0 Likes

Hey thanks, yeah I saw you mentioned in other thread, the only other bios I can get from the mobo webpage is F30 which is 

  1. Update AMD AGESA ComboV2 1.0.8.1 for AMD Ryzen 5000 processors support

gonna flash that one and see what it does. 

The first bios I tried was actually f31 with SAM support, so maybe it can be a bit more stable at least. Hopefully can work like intended. My only other option is change to another DDR4 but will have to ask a friend since I don't have a spare kit at hand.  

0 Likes


@Jazzinplayer wrote:

Hey thanks, yeah I saw you mentioned in other thread, the only other bios I can get from the mobo webpage is F30 which is 

  1. Update AMD AGESA ComboV2 1.0.8.1 for AMD Ryzen 5000 processors support

gonna flash that one and see what it does. 

The first bios I tried was actually f31 with SAM support, so maybe it can be a bit more stable at least. Hopefully can work like intended. My only other option is change to another DDR4 but will have to ask a friend since I don't have a spare kit at hand.  


Flashed the bios f30 and it's the same basically. 

0 Likes

UPDATE:
So old bios didn't worked. New bios didn't work also (F31j and k). Different memory kit no. 
Playing with the clock I got a stable 4.1 ghz and xmp on with all other values on AUTO. This settings gave pretty good and stable results on Horizon Zero Dawn benchmark. Managed to run it even in ultimate quality pretty smooth. Kept this settings and played Rogue Company for 2 hours, exactly at that time the PC crashed again with bios reset. 

Ordered a motherboard replacement from amazon, gonna build the whole system again. If it keeps doing same thing it might be the cpu. For a fast test I might need to BUY another one and try it on the same platform, while the RMA process might take months to solve for the actual cpu. This is the worst thing ever happened to me. 

0 Likes
Jazzinplayer
Adept II

Another update for anyone having issues. Sorry for the long post.

I got another x570 motherboard, same type just new. Situation was exactly the same or even worse so thought the CPU was faulty and needed to RMA. Decided to move my components (2080 super and 3600 mhz DDR4) to my streaming PC (Ryzen 5 2600, ax-370, gtx 1060, 520W PSU!) and play there while the RMA runs. It will also work as a test for those components making sure they are working fine. Everything ran smooth no problems whatsoever. 

So I actually thought it was a good idea to try the gtx 1060 on the 5600x build, maybe there is something there to see. Well, the system run smooth like butter with that gpu. Did all the benchmarks and tests possible, never crashed. CBP on XMP on GPU 100% usage, cpu reaching 4.65 ghz and running smooth.

First I thought it was a bios problem with the 2080 super, somehow it was bugged or crashing due to PCie or something. I did set it to gen 3 in Bios and problem was the same. So, my 2080 super and DDR4 are fine since they are running in the other PC., swapped memory and both DDR4 kits are fine no issues.

I then put the 2080s back to my 5600x build and crashed again. So 1060 is good, 2080 super insta crash... There you would think is the 2080 super the problem.. BUT, I do have a 1080ti that could test, IF the gtx 1060 worked fine, then I can play with my 1080ti until there is a Bios fix, right? Well, plugged the 1080ti ran Cinebench and PRime all good. Heaven Benchmark for the gaming/graphics and insta crash. Just like the 2080. 

So then I understood there must be a power issue, checked my PSU and it is a 750W Focus Plus (I thought it was 850w). So it has to be the power spikes are too high and system is shutting down. I ran one more test: 
With the 1080ti installed, CBP and XMP on, ran the Heaven benchmark at lowest detail posible, lowest res posible. It ran perfectly fine for 10 min, I was monitoring watts on HwInfo.  Then, up the res just a notch to demand more resources from the gpu. IT instantly crashed. So the power spike was too high and it shut down. 

So, why this same system was running perfectly fine on my i7 9700k build before? It was 4.9 ghz OC (huge energy consuming), same GPU, m.2 drives, HDD, DDR4 and PSU. Well, I think it was running well but on the very limit of its capacity, remember playing some DOOM eternal on ULTRA and crashing on the heaviest parts once or twice. So, this 5600x although is not consuming more watts than the other cpu (maybe at max OC it does) it is indeed making the whole system demand more energy, probably due to the increased IPC and performance. So is taking my GPU to its max limit and the power spikes are just out of boundaries.  That's what I think at least now. 

I already ordered a new 850W PSU to validate the theory can't confirm it's the solution just yet.
@mrsense@misterj what you think?  @CrispyCrunch check this out

0 Likes

Mhm ... I think a 750 watt PSU should be enough though... I have one too but I also use a 3080. But my friend , when put in his ryzen 5 3600 , everything worked although the difference between the 3600 and the 5600x are not that big , wattage wise. That's why I don't think it has anything to do with how many watt you have , it is all about the CPU i think.

0 Likes

it matters, the 5600x on max boost can go 4.7 ghz at 1.45 volts or more, on multithreaded load that rises the wattage a lot +150w. Where the 3600 won't go that high. Thing is I'm playing on the very edge of the limit here. Check this numbers: 

according to outervision.com power calculation, under a max clock and voltage both cpu and gpu, the system (including Cpu,GPu, Aio, Fans, Leds, Storage,etc) can use up to 645 W.
My PSU is a 750W Gold standard. Now a gold standard can give you 85% to 90% of that wattage. Considering the 85% we talking about 638 W, that's why Im at the very edge. And the system calculation do not account for power peaks where GPUs can hit even hit 50 W more. 

This means I'm playing on the very edge of my PSU power limit, that's why I can be "stable" with CBO turned off or with low res low detail. I just up the res and details and it crash, because GPU draws 100% power hitting the limits. There is also a bios setting that has like a protection feature as dummy load or something, disabling it might get me through some scenarios but I'm still playing dangerous ground. 

This chip draws 100% of gpus capacity and it's important to consider having enough power to run stable. This theory makes a lot of sense with all my tests so far. Gonna run one more test meanwhile wait for the new PSU: MSI afterburner under volt and under clock my GPU and run like that, it should be stable under low settings like how I ran with the 1060. 

0 Likes

wait for agesa 1.2.0.0 in january

0 Likes

Hi! Well I have found out that most of the gigabyte x570 boards have a problem with ryzen 5000 even after updating bios. If possible can you test and try the same on MSI x570 tomahawk with wifi mobo. 

0 Likes
Jazzinplayer
Adept II

Guys!!! I SOLVED it!!!!!!!!

Omg, this is actually upsetting but relaxing at the same time since I finally found the problem.

It was PSU related but not because of the total Watts, it was the cables. I've been using this Phanteks Cable extensions for over 2 years now, never had an issue. Since I was monitoring the watts usage and noticing very weird behavior,  like huge spikes and dumps, in one of many reboots I noticed once cable was pretty hot, that is not normal so I removed the extensions (CPU, Motherboard and PSU) and PROBLEM SOLVED. Running smooth like butter, ran all the benchmarks, ran prime 95, gamed for 2 hours and not a single problem. Giving me the expected performance now.

It makes a lot of sense by the way the PC was working. Still going for that 850 W PSU for more room on that department. Gonna keep testing but so far you can TELL is working as it should, it looks and feels fine. Thanks to everyone that helped. 

cables phanteks.jpg

Amazing news for you!

I returned mine to my vendor today so I could get a refund. I tested the standard cables of my PSU and of another PSU but the problem was still there so it wasn't the same problem we had :)

iv lost count of how many times I have seen and had to fix similar issues and the OP never does what they should. remove all components from the case put it on an open test bench ie the motherboard box or any non-conductive surface. unplug everything and start again using the very bare minimum. CPU, GPU, 1 stick of ram, and a basic SSD, not nvme and a directly connected PSU. start from there and keep adding stuff until it crashes and save yourself hours of frustration and a headache.

 

you live and learn bro. shame others keep doing the same mistakes and expecting different results lmao.

at least you fixed it, so all good in the end. don't ever think you know better and do it right way 1st time. like they say pride comes before a fall

Have the same probelm, i remove the extensors and nothing, some solution for this? I test 10000 times with diferent solutions and nothing...

0 Likes

Was it just the cpu power supply or also the 24pin?

0 Likes

great that you solved your problem and i actula have the excact same issue.
But i had a 750w and 1000w power supply and both crashes like your cpu did even though both psu's use different cables.
I know that both power supplys work 100%..
By the way only way i could make my pc actual run games was to manualy find cpuVcore and change auto to 1,4v
But this is in no way ideal and i loose some performance by doing so BUT it perform really well though.
Just saying a cpu should be able to run out of the box and control the power itself but my 5900x cpu cannot no matter which power supply or graphics card i use. ( i own a 1080rtx  and a 6900xt )

But i will go and check IF i by some mistake have changed the motherboard power cable so im not using the original one that came with the power supply..  ( pro tip NEVER reuse old power supply cables with other produts, and many power supply the connector on the cable is actula a little bit different so you cant even plug them in another product     )

0 Likes