Ryzen Owners Zen Garden

[Klemc]

Doesn't look normal, but also doesn't look completely trashed either. Pins don't usually bend themselves, judging from lack of damage on CPU I'd lean more towards user accident

Not all pads/pins are absolutely necessary to CPU functionality. You can check the pinout to see what the bent pins are, and what the pads correspond to. The problem with only some of the DIMM slots working sounds like a bent pin problem.

If your CPU exploded like the others there would probably be visible pin/pad damage in a larger area, not a single pin. And in the vicinity of a main rail like Vcore that actually delivers substantial power.

Wait, the CPU has been altered ?.. i see something but what could it be ?

About pins, why the PC worked and after no more ?.. btw RMA is in Amazon window time for motherboard.

 

[tabascosauz]

Wait, the CPU has been altered ?.. i see something but what could it be ?

About pins, why the PC worked and after no more ?.. btw RMA is in Amazon window time for motherboard.

You ask a lot of questions but you don't have any additional info so no one can answer any of your questions. Were the pins bent when the product was new? Have you ever removed and reinstalled the CPU at any point? If so, did you inspect the socket each time? Has your CPU install technique been 100% perfect? How did the pads look when the CPU was new? etc etc...

There's also this - I have no idea how credible this reddit post is:

https://www.reddit.com/r/pcmasterrace/comments/10f1rc6

Sounds like you should just RMA it then. I'm not sure what AMD's policy is for LGA CPUs, for PGA CPUs they state that they will not warranty bent pins, but some people still get their RMAs approved.

 

[Klemc]

I can provide this last images and please tell me it the bent pins contact can have scratched the CPU :


Thanks a lot.

 

[tabascosauz]

I can provide this last images and please tell me it the bent pins contact can have scratched the CPU :


Thanks a lot.

It's kind of impossible to tell with low picture quality.......camera isn't even focused on the bent pins, and the pads picture is not even in focus. Looks like dust on some of the pads.

Anyway, at this point it's pretty clear you didn't suffer the same catastrophic 00 code failure as other people have - not sure what happened in your case, but the CPUs blowing up isn't subtle at all.

The Strix-E has a POST readout, so unless you came home to your PC off but burning up with a 00 code and bulging/burns on pads, it's probably not the same problem.

 

[Klemc]

I've not bended pins here when i removed it first time !!

AMD Ryzen 7000X3D Processors Prone to Physical Damage with Voltage-assisted Overclocking, Motherboard Vendors Rush BIOS Updates with Voltage Limiters

This issue is recall worthy, cuz good luck to the average joe flashing a BIOS.

www.techpowerup.com

 

[3x0]

Then you've probably bent them while reinstalling it

 

[Klemc]

So when i receive my replacement mobo, how can i fully test CPU integrity at all levels ? Preferably from boot disk !.. please, if it's perfect i can keep it, if i find problems i can arg for RMA.

 

[3x0]

So when i receive my replacement mobo, how can i fully test CPU integrity at all levels, preferably from boot disk ?.. please, if it's perfect i can keep it, if i find problems i can arg for RMA.

Release v0.9.3.0 · sp00n/CoreCycler

Added support for setting the tests for y-Cruncher Added support for setting the allocated memory for y-Cruncher Added a disableCpuUtilizationCheck setting to disable the periodic CPU utilization c...

github.com

y-cruncher - A Multi-Threaded Pi Program

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

Create a Windows to go USB stick and copy the test programs on it if you want to test from boot usb.

 

[Klemc]

Release v0.9.3.0 · sp00n/CoreCycler

Added support for setting the tests for y-Cruncher Added support for setting the allocated memory for y-Cruncher Added a disableCpuUtilizationCheck setting to disable the periodic CPU utilization c...

github.com

y-cruncher - A Multi-Threaded Pi Program

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

Create a Windows to go USB stick and copy the test programs on it if you want to test from boot usb.

Can i boot from WinPE Sergei Strelec and add this tools it will lauch them ?

thanks

 

[3x0]

Yes

 

[TheoneandonlyMrK]

So when i receive my replacement mobo, how can i fully test CPU integrity at all levels ? Preferably from boot disk !.. please, if it's perfect i can keep it, if i find problems i can arg for RMA.

Stop doing it yourself, take it to a shop, hold them to account for their work, sometimes easiest IS best, And cheaper in time for everyone around you.

I can't get that life back and I'm but one.

 

[Klemc]

@TheoneandonlyMrK @everybody

I'm sorry.

/

My last is somebody poiting at that, i don't know why i can't see it with eyes, only camera can spot it.

 

[tabascosauz]

@TheoneandonlyMrK @everybody

I'm sorry.

/

My last is somebody poiting at that, i don't know why i can't see it with eyes, only camera can spot it.

No, you didn't do anything wrong, it's completely understandable. It's a crappy situation. All I'm saying is that it doesn't look like your hardware issue is the same as the hot button exploding issue that's been in the news lately, and that it's really hard to tell from the pictures what's going on aside from that you should RMA if possible.

CPU and RAM should not have risk of being damaged by the board unless there has been some sort of significant hardware failure - blowout of a VRM component, another component blowing up, a short developing on the board, etc. From my 5700G debacle I have only a little bit of experience as to what the knock-on hardware damage looks like, which you can test when you receive your new RMA'd board:

• Cores: not too sure, but maybe inability to sustain curve optimizer offsets like it could before, bad or weird clocks and performance, random crashing or reboots that don't look memory related
• SOC: iGPU instability, inconsistent memory training, memory-related BSODs, WHEAs, memory instability, poor UMC performance unable to reach previously stable UCLK, UMC throwing a fit over reasonably loose timings, etc.

The apparently damaged(?) pads are in an area that has a lot of memory-related pins, so it's not inconceivable that it could be related to one of your DIMM slots going dark. If AMD won't stop you from RMAing, I would just take the opportunity to RMA the CPU at this time, instead of waiting for the board to come back.

Usually the customer service ticket agent only asks for a picture of the heatspreader before processing your RMA, not the underside of the CPU. I don't think you will have to show them the pads unless they ask (for some reason). AMD CPU RMA is pretty fast at least in North America, and you can't use the CPU while waiting on a board anyways.

 

[Klemc]

No, you didn't do anything wrong, it's completely understandable. It's a crappy situation. All I'm saying is that it doesn't look like your hardware issue is the same as the hot button exploding issue that's been in the news lately, and that it's really hard to tell from the pictures what's going on aside from that you should RMA if possible.

CPU and RAM should not have risk of being damaged by the board unless there has been some sort of significant hardware failure - blowout of a VRM component, another component blowing up, a short developing on the board, etc. From my 5700G debacle I have only a little bit of experience as to what the knock-on hardware damage looks like, which you can test when you receive your new RMA'd board:

• Cores: not too sure, but maybe inability to sustain curve optimizer offsets like it could before, bad or weird clocks and performance, random crashing or reboots that don't look memory related
• SOC: iGPU instability, inconsistent memory training, memory-related BSODs, WHEAs, memory instability, poor UMC performance unable to reach previously stable UCLK, UMC throwing a fit over reasonably loose timings, etc.

The apparently damaged(?) pads are in an area that has a lot of memory-related pins, so it's not inconceivable that it could be related to one of your DIMM slots going dark. If AMD won't stop you from RMAing, I would just take the opportunity to RMA the CPU at this time, instead of waiting for the board to come back.

Usually the customer service ticket agent only asks for a picture of the heatspreader before processing your RMA, not the underside of the CPU. I don't think you will have to show them the pads unless they ask (for some reason). AMD CPU RMA is pretty fast at least in North America, and you can't use the CPU while waiting on a board anyways.

Thanks for all this explanations, i'm on AMD now, i think really the 7800X3D is my CPU like i wanted i never had such hardware in my life (always on 60 series, now 70ti...).

I also never had problem before, i return mobo today morning and yesterday buyed the same on Amazon.

Now for CPU it's with another reseller, they will test the CPU... i had them by phone... should i return it to them ?

There apparently not had hard hardware failure, the PC boot with B2 (the failed) dimm only, it just freezes PC at 6-8%.

 

[Mussels]

BURNout !!

one of those is too blurry to see, the others better - are they burned or bent?

You need ambient lighting reflected off something - like a lamp shining at a white wall, so you can light up the CPU without reflections in the way

Then you need a camera that's in focus and stabilised - rest it on something so shaking hands dont blur it. As long as it's in focus you can crop the phone later to make the CPU more visible, so don't get too close up to it.

I can provide this last images and please tell me it the bent pins contact can have scratched the CPU :


Thanks a lot.

scratched pads arent an issue - you can wipe them down with isopropyl on a lint free cloth to remove any debris or fingerprints, as long as they make contact they're good to go
the bent pins are quite a problem, and that's usually caused when installing or removing the CPU

 

[Klemc]

I finally found AMD RMA form and submitted, my MBoard beeing Asus i think it can help, i also told us mobo is RMA and replaced.

My smartphone (i call GSM) is verry good, i like it a lot, it's from 2019, has 8 core and is tiny, but hell the camera can't keep focus (all light setups tested), when it reaches a perfect focus it looses it immediately. A2 Core.
So, the camera can focus to a good quality image but i can't save it, too fast/quick. Video record didn't help.

Well, also i will have to wait the Agesa 1.0.0.7 and flashback without CPU/RAM/GPU.

For pads i can use air gaz compressed bombs i have.

 

[AusWolf]

So... the "hang up at cold boot, have to restart the PC once to properly boot up" error is back with this latest BIOS. It was like this with the factory BIOS, then everything was fine with 1.50, and now it's happening again! I'll never update to a beta BIOS again, I don't care about the mitigations, either!

At least the "memory context restore" function seems to finally work now, that's a plus.

 

[Mussels]

So... the "hang up at cold boot, have to restart the PC once to properly boot up" error is back with this latest BIOS. It was like this with the factory BIOS, then everything was fine with 1.50, and now it's happening again! I'll never update to a beta BIOS again, I don't care about the mitigations, either!

At least the "memory context restore" function seems to finally work now, that's a plus.

What DDR5 speed are you at?
Makes me wonder if they cranked up the SoC to help bypass the cold boot issues

Oh, and did you manually set the SoC in the past/this time? I'm surprised by so many people who seem unaware of how to even do it, in the other threads on the topic

 

[AusWolf]

What DDR5 speed are you at?
Makes me wonder if they cranked up the SoC to help bypass the cold boot issues

Oh, and did you manually set the SoC in the past/this time? I'm surprised by so many people who seem unaware of how to even do it, in the other threads on the topic

It's at 6000 MHz, and my SoC was 1.2 V, and still is with the new BIOS. Exact same settings.

 

[Mussels]

It's at 6000 MHz, and my SoC was 1.2 V, and still is with the new BIOS. Exact same settings.

1.2v on the old BIOS may have ran higher than it showed, is my guess. you can try 1.25v now, or lower DRAM and see how it behaves - if the setting is simply more accurate now, that might change it's behavior a lot.
hibernate and sleep are implicated in the SoC voltage burnouts so everyone on AM5 should avoid using that for now

 

[trparky]

1.2v on the old BIOS may have ran higher than it showed, is my guess. you can try 1.25v now, or lower DRAM and see how it behaves - if the setting is simply more accurate now, that might change it's behavior a lot.
hibernate and sleep are implicated in the SoC voltage burnouts so everyone on AM5 should avoid using that for now

I don't get the hibernate part though. Hibernate is simply a full shut down with an added full system memory dump to disk.

 

[Mussels]

I don't get the hibernate part though. Hibernate is simply a full shut down with an added full system memory dump to disk.

Because some boards hibernate, but then remain in sleep instead of fully powering down to wake up faster. They only use the hibernate if power was lost, otherwise it's still using sleep.

Something to do with that is breaking voltage monitoring, and killing AM5 CPUs - until people know more, its your hardware to risk.

Got some help and tuned in my ram a bit more
For 2x32GB, i'm impressed it runs even close to this fast - best i ever got with 2x8 was just a hair under 60ns, for comparison.

 

[trparky]

Something to do with that is breaking voltage monitoring, and killing AM5 CPUs - until people know more, its your hardware to risk.

You know, up until a recent AGESA update, sleep was completely broken on Zen 4. You'd put your computer to sleep and then seconds later it would wake back up. I wonder if that's somehow connected.

 

[tabascosauz]

Because some boards hibernate, but then remain in sleep instead of fully powering down to wake up faster. They only use the hibernate if power was lost, otherwise it's still using sleep.

Something to do with that is breaking voltage monitoring, and killing AM5 CPUs - until people know more, its your hardware to risk.

Got some help and tuned in my ram a bit more
For 2x32GB, i'm impressed it runs even close to this fast - best i ever got with 2x8 was just a hair under 60ns, for comparison.

The fact that it's 2x32GB is not more relevant than the fact that it's dual rank 16Gb hardware, so it's not a whole lot different to the UMC than a 2x16GB CJR or Bdie kit.

Quad rank 3800 would certainly raise some eyebrows, however. Would be pretty crazy, actually.

You know, up until a recent AGESA update, sleep was completely broken on Zen 4. You'd put your computer to sleep and then seconds later it would wake back up. I wonder if that's somehow connected.

Have seen that on certain AM4 hardware combinations as well, just never thought much of it I guess.

 

[AusWolf]

1.2v on the old BIOS may have ran higher than it showed, is my guess. you can try 1.25v now, or lower DRAM and see how it behaves - if the setting is simply more accurate now, that might change it's behavior a lot.
hibernate and sleep are implicated in the SoC voltage burnouts so everyone on AM5 should avoid using that for now

But if the actual SoC voltage was higher with the older BIOS, then why is my idle power consumption and temperature 5 W and 5 °C higher now? Something doesn't add up.

Quick question: What is CPU_VDDIO? I've noticed that it's 1.35 V, same as my DRAM voltage. Is that normal?