• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

New 5900x crashing

devilboost said:
Right now I let it ran with stock values. Everything on default except SoC 1.1v
Been 32 mins and no errors.

I have a corsair h100i platinum which is a 240mm rad.

Which settings would you suggest to change in bios?
Click to expand...

maybe 240mm being the reason, I only messed about with PBO and curve optimizer on mine so cant really provide much help

having issues when everything is at stock nightmare
 
Sounds like the memory controller on the 5900X is borked to me. You could perhaps check the pins underneath to see if you have a broken or bent one.

I would RMA. If you have the 5600X to hand reinstall it and see if the same happens?

Ryzen chips do fail like Intel.

Alternatively, the RAM was not getting enough juice at the proper speeds, for whatever reason it is tripping.

But your errors point to that pesky 5900X memory controller.

I got the 5800x3D and replaced my 5600X, seems like it is fine thus far.
 
VonScar said:
Sounds like the memory controller on the 5900X is borked to me. You could perhaps check the pins underneath to see if you have a broken or bent one.

I would RMA. If you have the 5600X to hand reinstall it and see if the same happens?

Ryzen chips do fail like Intel.

Alternatively, the RAM was not getting enough juice at the proper speeds, for whatever reason it is tripping.

But your errors point to that pesky 5900X memory controller.

I got the 5800x3D and replaced my 5600X, seems like it is fine thus far.
Click to expand...

Im intending to do that. should I send it back to amazon or to amd?
 
I have fited back the old 5600x and ran a occt test with the ram memory oc to 3200mhz ( the one that failed in under 10 seconds on the 5900x)

here are the temps during the test print

Right now I am doing a prime95 test on it ( full blend, the one that usually fails in the first minute or in under 15 minutes )
 
Last edited:
Give your memory a bit more juice. I have 4 sticks of memory and on my old 3700X I needed 1.38v to keep them happy. All was well when I did that. ddr4 is fine up to 1.5v and 1.4v or below is more than acceptable.
 
Last edited:
devilboost said:
It failed on 5600x too after 12 mins

screen here print

right now I dont really know if it is faulty, I have another problem or just the prime95 is the issue
Click to expand...
If the RAM is failing at stock voltage and the advertised speeds are failing, then RMA the RAM.

When you put the 5900X CPU in, you reset the motherboard back to motherboard settings. So what is the recommended voltage for your specific RAM and what is it supposedly running at in BIOS? If running at correct voltage the RAM's gone bad.

What you can do is test RAM stick by stick and then identify the faulty stick. If Memtest runs merrily on one stick for an hour no issues, and the other shows faults you will have a culprit.
 
Make sure you got pbo turned off in mobo. Some mobos default it to on and can cause issues with idle and low load reboots ans crashes due to too high boost at not enough voltage.
Core cycler will catch those errors quick
 
Update: I got my replacement cpu. Same issues like with first.

I decided to set PBO on manual and fine tune each core using OCCT extrem test for 15 mins.

My final list of undervoltage is:
-10
+1 (this is core 1)
-15
-10
-15
-20
-15
-10
-10
-20
-15
-10

Now its seems stable. 30 minutes of extreme test and no errors. And my temps are somewhere in 80c. With some highs of 85c. Which is way better than the 90c I was getting last time.

But I still have the other issue to sort the other error I was getting.

Whea logger id 46.

From my experience with the old one. I had to raise SOC vdd voltage to 1.1v yet it doesn't fix it on this one.

This error appears only when idling or when the computer finished a task and goes to idle. (Last one happened after the 30 mins test ended. In the second when the pc finished the test boom. Reset)

Hence why I don't have a print to show you the temps during my test.

To mention that all the tests have been made with D.O.C.P enabled and memory at the rated 3200mhz. Disabling the D.O.C.P feels like it sorts the WHEA id46 error.


Any suggestions what can I do. I don't really want yo run my ram at 2133mhz.
 
devilboost said:
Update: I got my replacement cpu. Same issues like with first.

I decided to set PBO on manual and fine tune each core using OCCT extrem test for 15 mins.

My final list of undervoltage is:
-10
+1 (this is core 1)
-15
-10
-15
-20
-15
-10
-10
-20
-15
-10

Now its seems stable. 30 minutes of extreme test and no errors. And my temps are somewhere in 80c. With some highs of 85c. Which is way better than the 90c I was getting last time.

But I still have the other issue to sort the other error I was getting.

Whea logger id 46.

From my experience with the old one. I had to raise SOC vdd voltage to 1.1v yet it doesn't fix it on this one.

This error appears only when idling or when the computer finished a task and goes to idle. (Last one happened after the 30 mins test ended. In the second when the pc finished the test boom. Reset)

Hence why I don't have a print to show you the temps during my test.

To mention that all the tests have been made with D.O.C.P enabled and memory at the rated 3200mhz. Disabling the D.O.C.P feels like it sorts the WHEA id46 error.


Any suggestions what can I do. I don't really want yo run my ram at 2133mhz.
Click to expand...

Have you tried upping RAM voltage to 1.38V or somewhere near ?
 
As above, also, is your FCLK matched to the Ram speed? I have got WHEA errors when pushing my FCLK to 2000 when using 4000mhz RAM (should be half of the ram speed)
 
Experienced some whea crashes with D.O.C.P disabled (ram on 2133) while playing witcher 3.

went back to bios and switched PBO limit from disabled to manual and gave it these values:
PPT: 170w
TDC: 118A
EDC: 165A

No crashes since, been over 1 hr of playtime and no crash. Will try to put back the ram oc to see if the lack of limiter was the issue.
 
I had two faulty 5900X's back to back, the first crashed so often and so randomly I couldn't even tell if the build worked because it'd crap out before I could even get into the bios half the time. The second worked fine for a few months but got progressively worse to the point it was unusable and this was all at stock or less. I sent the second back to OcUK (first was from a competitor) but they returned it saying it was fine, had to RMA it directly with AMD, which was fun but they did send out a new replacement.

I know it's irrational but every crash had me worrying about the CPU again, that uncertainty annoyed the hell out of me and I couldn't say if I was especially lucky or the 5900X just has a higher failure rate but if yours can't handle stock settings I'd be looking for a different model entirely.
 
devilboost said:
Update: I got my replacement cpu. Same issues like with first.

I decided to set PBO on manual and fine tune each core using OCCT extrem test for 15 mins.

My final list of undervoltage is:
-10
+1 (this is core 1)
-15
-10
-15
-20
-15
-10
-10
-20
-15
-10

Now its seems stable. 30 minutes of extreme test and no errors. And my temps are somewhere in 80c. With some highs of 85c. Which is way better than the 90c I was getting last time.

But I still have the other issue to sort the other error I was getting.

Whea logger id 46.

From my experience with the old one. I had to raise SOC vdd voltage to 1.1v yet it doesn't fix it on this one.

This error appears only when idling or when the computer finished a task and goes to idle. (Last one happened after the 30 mins test ended. In the second when the pc finished the test boom. Reset)

Hence why I don't have a print to show you the temps during my test.

To mention that all the tests have been made with D.O.C.P enabled and memory at the rated 3200mhz. Disabling the D.O.C.P feels like it sorts the WHEA id46 error.


Any suggestions what can I do. I don't really want yo run my ram at 2133mhz.
Click to expand...
so you got a replacement cpu and your back at it again with the curve optimiser doing the undervolting. why not just turn pbo off and run pure stock and take things from there?
 
Cyber-Mav said:
so you got a replacement cpu and your back at it again with the curve optimiser doing the undervolting. why not just turn pbo off and run pure stock and take things from there?
Click to expand...
when I got the replacement ( which came from different seller , AMD amazon) I installed it, hit clear cmos and turned it on with everything default. Went for some browsing and mundane tasks and it crashed. I have not done any curve optimizer with the previous one, didnt knew how to or even what I am looking at. Was the first time when I tried to fix or modify a ryzen one, since then I had only athlons and intels.

Browsed through forums and found that some people undervolted their 5900x and it was more stable than stock. (felt more stable but still crashing like once every 40 mins with same errors whea 18 and 46)


I have replaced my ram sticks with another 2x8gb kit which I had in another pc. fortza 3000mhz cl15 (which I overclocked to 3000mhz and set the CPU fabric to 1500mhz. While installing them I decided to do some cable check to se maybe something is broken. saw that my CPU 8pin plug was not fully inserted. it had like a 1mm gap to click (that plug is quite hard to reach and even to spot, hidden by VRM heatsink and top-mounted fans)

Since then no errors, not sure if it was the ram or the plug.

Here are my stats during a 4 hour Warzone 2 gaming session print1 print2

Will come back in a few days with update.

Let me know if there should be other bios settings to address
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom