Question I have lost 2 x 5600g in a span of 2 months.

[nicalandia]

Likely a faulty MB

 

[DAPUNISHER]

Likely a faulty MB

of course... I have tried different boards, i have tried different ram, threw on different nVME's incase PCI-E 4.0 was messing something up, i even threw on a different PSU. I think i pretty much debug'd the hell out of the chip to validate its a dead chip.

can not load windows setup on a B450 or even load past post for that matter.
System freezes after post, and is stuck infinitely on the motherboard logo in uefi boot.
I dont get the windows loading rings in the bottom center when i debug the 5600G.

Again if i swap back to 3600 everything is peachy.

The system leaves me after a good 8 hour stress test for validation.
I even call back 48 hours later and ask if it ever randomly rebooted.

I can understand 1, but both cpu's really has me stumped.
If it was a AGESA issue, i would assume i could not validate the machines before it left my house.

I will try the beta 1.2.0.6b beta bios ASUS has and see if they fixes anything.
But i am not holding my breathe...

Whatever happened, did so on 2 systems. And given the OP's skill level, any normal TS advice is captain obvious territory. It is rare for a CPU that is not being abused to fail. For it to happen to 2, in different systems, and in such a short time frame, under normal usage, is much rarer still.

 

[PingSpike]

Certainly seems like the motherboard is the most likely culprit. If you put another video card in and disable the iGPU does it boot?

Another possibility I suppose is that both CPUs were from a batch of bad ones, but given I've never actually had a CPU fail on me I always blame the motherboard first.

 

[aigomorla]

Likely a faulty MB

This was my inital thought, however when i put in a 3600, the board boots up fine.
Its currently in use on a loaner 3600 + W4100 running like a champ.
Infact my cousin says it feels faster then when he had it running the 5600G for some weird reason.
I have ruled out the board being broken, unless it is a BIOS issue, but that would mean ASUS has no BIOS available that does not have an issue for the APU, which i find highly unlikely, or both boards and my B450 in my possession are faulty and have issues with APU, which again is not as rare but still rare for 3 / 3 working boards to fail, unless they were like some uncommon Chinese brand boards like the x79 / x99 rebrands.

Certainly seems like the motherboard is the most likely culprit. If you put another video card in and disable the iGPU does it boot?

Another possibility I suppose is that both CPUs were from a batch of bad ones, but given I've never actually had a CPU fail on me I always blame the motherboard first.

You can't really disable IGP on APU's.
You can tell the board to either boot from PCI-E or IGP.
I did not see any feature to "disable" IGP.

The 5600G posts fine.
It goes into bios fine.
It makes changes fine.
Its what happens after.

After post, it gets stuck in the motherboard logo boot screen. aka, the UEFI boot screen.
After many fail bootups windows tries to repair... it freezes on repair loadup, wont even get into repair.
It fails any and all attempt to flash bios with bad cpu.
It fails any and all attempts to reinstall windows.

When the cpu was new, i did not have any of the issues above, it booted / installed / flashed fine... it was after some time passed, these problems occurred.

I am leading closer to the suspicion that its a bad batch.
Both cpu's were bought at same vendor.
Both cpu's were bought within 1 week of each other.
I am assuming that both cpu's were probably in the same shipment, which highly like means they were both in the same batch at production.

 

[maddie]

This was my inital thought, however when i put in a 3600, the board boots up fine.
Its currently in use on a loaner 3600 + W4100 running like a champ.
Infact my cousin says it feels faster then when he had it running the 5600G for some weird reason.
I have ruled out the board being broken, unless it is a BIOS issue, but that would mean ASUS has no BIOS available that does not have an issue for the APU, which i find highly unlikely, or both boards and my B450 in my possession are faulty and have issues with APU, which again is not as rare but still rare for 3 / 3 working boards to fail, unless they were like some uncommon Chinese brand boards like the x79 / x99 rebrands.


You can't really disable IGP on APU's.
You can tell the board to either boot from PCI-E or IGP.
I did not see any feature to "disable" IGP.

The 5600G posts fine.
It goes into bios fine.
It makes changes fine.
Its what happens after.

After post, it gets stuck in the motherboard logo boot screen. aka, the UEFI boot screen.
After many fail bootups windows tries to repair... it freezes on repair loadup, wont even get into repair.
It fails any and all attempt to flash bios with bad cpu.
It fails any and all attempts to reinstall windows.

When the cpu was new, i did not have any of the issues above, it booted / installed / flashed fine... it was after some time passed, these problems occurred.

I am leading closer to the suspicion that its a bad batch.
Both cpu's were bought at same vendor.
Both cpu's were bought within 1 week of each other.
I am assuming that both cpu's were probably in the same shipment, which highly like means they were both in the same batch at production.

When the cpu was new, i did not have any of the issues above, it booted / installed / flashed fine... it was after some time passed, these problems occurred.

This important line should be paid attention to, by those making suggestions.

 

[CuriousMike]

As a useless data point, I have been running a 5700G non-stop since September as my work machine.
This thing is loaded at least 8 hours a day.
I have had my machine crash once in that time period.

Running on an Asus B550 with two 32-GB modules = 64GB ram.
I have run it with a 5700XT graphics card and built in graphics.

 

[coercitiv]

You can tell the board to either boot from PCI-E or IGP.

I think PingSpike was referring to PCI-E boot, the idea being you'd try to emulate the same "path" that works with 3600. Problem is you already have basic graphics adapter functionality, and that suggests you'd be able to get into Safe Mode or boot from a Windows install stick at least, which isn't happening.

I would also attempt to boot from an USB port or a SSD connected via the chipset (or at least check whether USB stick or disk drives are visible from all ports). Something is definitely broken, and yet not broken enough to prevent UEFI from working, so avoiding the iGPU or avoiding different I/O paths may give you hints about what's happening. I find it fascinating that UEFI is seemingly working and stable.

All of that if you're still curious about what it could be ofc, as I imagine you've already spent quite a lot of time debugging something with no sense of progress.

PS: You can always take a break and watch a 5700g literally blow a fuse and become unusable on anything but a Lenovo system. (no they're not able to get into UEFI on other systems)

 

[igor_kavinski]

I am leading closer to the suspicion that its a bad batch.
Both cpu's were bought at same vendor.
Both cpu's were bought within 1 week of each other.
I am assuming that both cpu's were probably in the same shipment, which highly like means they were both in the same batch at production.

Agree.

 

[aigomorla]

Also my nieces PC which amazon replaced the CPU with a new one booted fine once it got the replacement 5600G.
I am 100% sure her's was dead, because the replacement boots fine.
I am just waiting to see how long it will last, if it was from the same batch as the other two which i hope not.

I got a RMA approved from AMD.
I'll fill you guys in on the new chip once i receive it.
But i bet you it will boot up and work just fine like my nieces replacement did.

Running on an Asus B550 with two 32-GB modules = 64GB ram.
I have run it with a 5700XT graphics card and built in graphics.

i am wondering if having a dedicated GPU will somehow buffer it.
Well, if these replacements do not fail, then i am pretty sure it was a bad batch.
If they do fail again, then its most definitely the ASUS STRIX boards killing them somehow, or the IGP being fragile to the point of it being a deck of cards stacked in a castle.

PS: You can always take a break and watch a 5700g literally blow a fuse and become unusable on anything but a Lenovo system. (no they're not able to get into UEFI on other systems)

WTF man... why did they add this?
I understand TPM being important and all, but seriously... they making it so second hand is impossible.
Did they pick up some evil tricks from apple?

But i highly doubt the ASUS boards had anything to do with PSB, or did they?
Or did the boards trip that fuse unknowingly?

 

[moinmoin]

You can always take a break and watch a 5700g literally blow a fuse and become unusable on anything but a Lenovo system. (no they're not able to get into UEFI on other systems)

Ugh, I hate this PSB feature so much. Dreading seeing such gimped e-waste CPUs appear on the used market. Lenovo needs to be dead as a brand in the desktop market for this alone.

 

[igor_kavinski]

If they do fail again, then its most definitely the ASUS STRIX boards killing them somehow,

iGPU getting overvolted by mobo? Perhaps BIOS optimized defaults are not optimized for APUs?

 

[aigomorla]

iGPU getting overvolted by mobo? Perhaps BIOS optimized defaults are not optimized for APUs?

Hmmm i think i'll hardlock the voltages for this APU this time and not leave it on auto.
Does anyone know where i can find the official factory stock voltage settings for the 5600G?
Googling it does not give me any official technical documents.

 

[igor_kavinski]

Hmmm i think i'll hardlock the voltages for this APU this time and not leave it on auto.
Does anyone know where i can find the official factory stock voltage settings for the 5600G?
Googling it does not give me any official technical documents.

Nothing official but this guy talks about the voltages here: https://www.reddit.com/r/overclocking/comments/qhg2qf

You can push the RAM much, much further. The IMC on Cezanne is unbelievebly strong as Mine is currently dailying at 4533C18-18-18-38 at FCLK 2266 (albeit mine's b-die rather than CJR, but CJR is pretty good chips too so hitting these clocks should not be difficult). Mine was running at 1.25V SOC and 1.45V VDDR, iGPU 2200mhz.
The key is to raise the SOC voltage. Vermeer has a 14nm IO die so 1.2V+ lead to degradation, and that's fair. In contrast Cezanne's IMC is on the same die as 7nm CPU, so ~1.25V is safe.
In fact, in stock config vSOC is at a even higher 1.3V+, as iGPU is also pulling power from vSOC, and anything less than 1.2V vSOC would lead to an unstable iGPU.

 

[igor_kavinski]

AMD Ryzen 7 5700G 8 Core Cezanne APU Performance Leaks Out, Overclocked To 4.75 GHz & Faster Than Ryzen 7 5800X

The AMD Ryzen 7 5700G 8 Core Cezanne APU featuring the Zen 3 core architecture has been overclocked and benchmarked over at Chiphell.

wccftech.com

The leaker also states that enabling SAM directly affects the SOC voltage. Without SAM, the voltage is set to 1.15V by default while enabling SAM raises that to 1.2V.

So stock is supposed to be 1.15V

 

[igor_kavinski]

[Official] AMD 5700G 5600G 5300G APU APU

I want to test but it takes too much time in the end for all variables. Might test 1333-1600FCLK and see if I can maximize CO to -20 & -30 something for the cores later today. I don't think we need to go too low on FLCK before we maximize the available settings for CO. Would be good to find a...

www.overclock.net

iGPU voltage is derived and is a offset from vSoC. Doesn't need to be that high. Actaully you don't want the iGPU to guzzle voltage for unnecessary wattage if it for some reason defaults to vSoC which I found it out to do at times. (check it's not using more than ~1.150V @ stock)
It should remain @ stock even if you fiddle with the vSoC voltage.
iGPU doesn't need that much voltage at stock speeds even if you OC MEM & FCLK.

Seems 1.15V it is for the iGPU.

 

[DAPUNISHER]

So stock is supposed to be 1.15V

Interesting, I thought 1.2V was the recommended for the SoC for 5 series APUs.

I am posting from the 5700G at the moment. Moved it to a White CoolerMaster NR200P with ASRock B450 Gaming ITX/ac. Really like this build.

In the UEFI under AMD CBS - NBIO - GFX, it includes the setting to disable the iGPU. I don't think Asus has that though, at least not under CBS. And I would not baby the 5600g either. I have mine in the OG B450 Tomahawk, with CPU and iGPU overclocked, and it is great, even with a less effective cooler than either the OP is using i.e. Noctua Redux 12. These APUs can take it when they are not defective or somehow damaged. Maybe they were in that Amazon warehouse hit by the tornado in early December.

 

[PingSpike]

iGPU getting overvolted by mobo? Perhaps BIOS optimized defaults are not optimized for APUs?

That was my thought (not even a theory at this point) when I suggested trying to take it out of the equation by disabling it. It could be a bad batch but maybe the APUs are fragile enough that crappy overvolting by the board burns them out, or there is a hardware issue in general. And that's why they worked at first, sort of like Intel's self destructing SATA sandybridge SATA controllers that worked for awhile until they didn't. But if you really can't disable the iGPU (I've never used one of these chips) then even with a PCI-e card I suspect it will continue to cause problems.

 

[A///]

I saw someone mention theirs failed a few months ago but shrugged it off. It's going to be interesting going forward if the igpu failure cause issues within the Ryzen 7000 lineup, whereas Intel has had over a decade to perfect their integrations.

My sons 1700x is still going strong (he doesn't even want me to upgrade it to a 3900x for free)

I can't comprehend this!

 

[aigomorla]

I can't comprehend this!

my cousin doesn't want to give me back the 3600 even tho i told him i got RMA approved for his 5600. lol...

 

[ElFenix]

most likely just bad luck


Be sure to use to much non-conductive heatsink paste. The more the better:

“Too Much Thermal Paste” – Benchmark of Thermal Paste Application & Quantity

The “correct” method for applying thermal paste is still the subject of arguments, despite plenty of articles with testing and hard numbers to back them up. As we mentioned in our Threadripper paste application comparison, the general consensus for smaller desktop CPUs is that, as long as enough...

www.gamersnexus.net

didn't test my fav method: cross application, clamp HS, remove HS, wipe paste off HS, reclamp HS.

 

[moinmoin]

my cousin doesn't want to give me back the 3600 even tho i told him i got RMA approved for his 5600. lol...

If you want him to switch to the 5600G dare him to some benchmarks runs with the other 5600G system so he can see the performance difference.

 

[igor_kavinski]

my cousin doesn't want to give me back the 3600 even tho i told him i got RMA approved for his 5600. lol...

Your cousin just needing to watch movies means you get the 5600G. Congrats!

 

[DAPUNISHER]

I saw someone mention theirs failed a few months ago but shrugged it off. It's going to be interesting going forward if the igpu failure cause issues within the Ryzen 7000 lineup, whereas Intel has had over a decade to perfect their integrations.

Fun fact: AMD has been doing APUs since 2011, that's over a decade too. I have used quite a few of them. Before that boards with IGP from AMD and nVidia. Have always enjoyed testing which games can be made playable on iGPUs. From the Zen gen APUs I have used Athlon 3000g, Ryzen 3200, 3400, 5600, 5700, and no stability or hardware issues. The 3000g responds the best of the bunch, in terms of overall gaming performance gains, from overclocking both CPU and Vega.

The 5600g&5700g can play almost anything well enough to enjoy, and most at 1080p. These were selling in OEM systems well before the retail release in N.A. With so many in use for so long, it is a safe bet they are fine. Like every other AMD APU I have ever used, going back to Trinity I think.

 

[DAPUNISHER]

t maybe the APUs are fragile enough that crappy overvolting by the board burns them out, or there is a hardware issue in general.

His cuz had it die from watching movies. I let whatever board I have them in at the time, go full HAM on them. CPU and Vega overclocked. Done it through UEFI, Ryzen Master, and the Radeon software suite. Then played hours and hours of Master Chef collection, Fallout series, Skyrim, Old Star Wars games, Tomb Raider 2013, OG Crysis, really old games in my GOG library. Still no issues. The whole "fragile" hypothesis sounds silly based on my anecdotal experiences thus far. If one of them goes belly up, I will report back, but I don't anticipate that being more than a remote possibility.

If this came off confrontational, I do not intend it that way. I am a little irritated one person's 2 sample experience even fosters such a conversation, where people are beginning to extend it to the entire product line. The idea is quite honestly. laughable IMO.

 

[igor_kavinski]

Hmmm i think i'll hardlock the voltages for this APU this time and not leave it on auto.
Does anyone know where i can find the official factory stock voltage settings for the 5600G?
Googling it does not give me any official technical documents.

Maybe you could get the stock voltage information from the BIOS itself if there is a "PC Health" tab there showing the various voltages or through Ryzen Master?