0.5 - 6% of Ryzen 5000 Processors Faulty

[philo-sofa]

https://www.techradar.com/uk/news/amd-ryzen-5000-processor-failures-are-they-really-that-bad

1. "According to PowerGPU, of the 50 5950X CPUs it received, eight were DOA. Out of 50 5900X CPUs, four were DOA.

PowerGPU’s 5800X and 5600X stock also reportedly had high failure rates as well."

2. However PowerGPUs initial tweet on this was deleted, and other retailers have failure rates of c2%:

"Hardware Unboxed asked some ‘major’ PC parts retailers in Australia, and were told the AMD Ryzen 5000 failure rate was below 2% – which is around normal."

Is even 2% normal? I've not encountered too many examples of DOA CPUs on forums before.

3. (edit) Recent Mindfactory.de data show an even lower fault rate for the first 15 days of c0.5% (source via @Zeeflyboy)

4. (edit) PC World have suggested an aggregate failure rate of 2.9% for Ryzen 5000 (source via @humbug)

 

[philo-sofa]

Theres already a thread about this mate

There's a thread about crashes we're having, but not one about alleged DOA chips etc is there? Apologies if I'm doubling up.

powerGPU.com are skint, probably because of the lockdown and the fact that the money in the industry right now is office PC’s not overpriced RGB gaming systems. Clearly they are desperate to drive traffic to the site and what better way than to drive a bit of controversy against AMD.

It’s baseless nonsense.

Hmm, that would make a fair bit of sense.

 

[philo-sofa]

Think its in the crashes thread mate but could be mistaken though i know i read it somewhere i apologise if im mistaken no doubt dave will be along soon

Haha true I did actually think about how there's be gloating from one side of the Aisle (Dave in particular) but numbers are numbers.

I think its more to be honest. Ill give them a big side eye before buying one..

For what it's worth, I've had no problems outside of memory and even that was just an overclocking issue; I'd classify myself as a very happy Ryzen 5000 owner.

Still even 2% seems high for DOA chips.

 

[philo-sofa]

We would have heard more about this by now. Seems like a hot load of nothing.

It's pretty bold making a thread titled "~2-6% of Ryzen 5000 Processors Faulty" based on one company's experience...

The upper estimate is based on one company’s estimate, the lower one is based on several companies’ estimates.

it could be a bad batch of CPU as many online pointed out, the fabrication node is the same as Zen 2. X570 boards have been succesfully used with Zen 2 with no reported issues such as this.

A bad batch (or technically two) that makes a lot a sense as a cause of the higher failure rate.

Unfortunately don't know anyone who has done big purchases of 5000 series CPUs.

From previous experiences of people building big render farms, etc. Intel CPUs have very low failure rate - to the point some setups didn't have a single obviously faulty CPU, some of the first Ryzen lines weren't so good through with a fair few failures especially with the 1700(X) but the 2000 series has been pretty much the same as Intel.

That definitely seems more in line with what I've seen than 2%. In this article @smilingcrow provided, PowerGPU suggest they've only had a single properly DOA Intel chip.

 

[philo-sofa]

I did wonder about that but then it didn't explain the constant BSODs and reboots at default settings and the disparity between core counts in the bios. Ryzen Master was just another thing and I knew AMD were releasing an update to sort out the telemetry in that but I wasn't happy with the instability of the system. Even with a CCD disabled it should work at stock settings and it simply didn't.

I had a choice of waiting to see if a mobo update would sort the problem and putting up with being dumped out of raids etc, sometimes up to every 20 mins, or returning and trying a new chip.

Personally I think that if AMD are putting chips out that behave like this then the BIOS code they supply to mobo manufacturers should make sure it's up to scratch to prevent problems at stock settings.

I wouldn't be happy if that had been my experience either TBH, very understandable.

 

[philo-sofa]

Op and thread title updated with mindfactory.de estimates (thanks @Zeeflyboy).

Please stop attacking one-another (and more importantly me) if you don't like the number. No one has the final RMA/failure datasets yet so we're all trying to interpolate and predict reality; everyone I've read here's doing it in good faith.

 

[philo-sofa]

This has been debunked. PowerGPU deleted the tweet, PC World gathered data from other vendors.

Ryzen 5000 series fails at 2.9 percent.
Ryzen 3000 series fails at 3 percent.
ThreadRipper 3000 series fails at 2.5 percent.

Intel 9th-gen fails at 0.9 percent.
Intel 10th-gen fails at 1.2 percent.

3% which is higher than Intel but their 14nm is far more mature.
These CPU's are also NOT DEAD, as PowerGPU reported, they are rejected for failing their torture test requirements, that being stress tested with high frequency RAM with all RAM slots occupied. "They didn't pass muster for frequency or letency"

Other tech journalists had spoken with their countries retailers, Hardware Unboxed reported under 2% failure rates for Ryzen 5000 in Australia, Mindfactory reported under 1% for Ryzen 5000 in the EU.

https://www.pcworld.com/article/3608349/ryzen-5000-failure-rates-we-reality-check-the-claims.html

With respect this isn't quite debunking, 2.9% is smack in the middle of the 0.5-6% range.

Thanks for the additional info though, have added to the OP.

 

[philo-sofa]

The 2.9% is not a failure rate, its the rate the CPU's failed "to meet this particular system builders standards" a bit like 8Pack might bin CPU's for his systems, if they don't meet his standard they don't go in, that's what that is. PowerGPU said their CPU's were DOA.

The exact phrase seems to be 'internal quality screening' which is a bit more ambiguous, but fair point it's not necessarily borked.

 

[philo-sofa]

I think scale will have an impact on these rates.

PowerGPU are a system builder, in the short time they have been building Ryzen 5000 PC's how many have they sold? a few hundred? so they order several CPU batches of 50 units, the first two batches they may have had one dead CPU, so "0.5%" they order another two batches and there are 3 dead CPU's "3%" and another two batch has 4 dead in one 50X batch and two in the other "6%"...


The thing here is if you're only shifting a few 100 units you don't really have a sample size that's big enough for a conclusion, if you're shifting tens of thousands you probably do and they ain't necessarily going to agree with eachother.

We do actually have enough data to start doing this now. Am a statistician by trade so did a 99% confidence interval for the PowerGPU 5950X and 5900X on the back of a napkin and it gives a range 3.6% to 20.4% rejection rate for their CPUs. However combining PowerGPU with Mindfactory.de's dataset you get a much narrower failure rate: a 99% chance it's between 0.44% to 0.67%.

The trouble is we also need PC world's data, given they suggest a large manufacturer has a rejection rate of 2.9%.

Looking at it overall I'd suggest there's both low volumes for some outfits leading to unusual findings and also different failure rates for different batches (e.g. whomever PC world's large 2.9% source is, vs mindfactory). On top of that there are obviously differing definitions of failure.

My sense is that the true 'fail rate' is about of 0.44% to ~3.2% most likely, but I really need all the raw data.

 

[philo-sofa]

What are you doing? no we don't.

I did a confidence interval for a binomial probability function, based on the Bernoulli distribution. Jargon aside it’s a dead simple, robust calculation.

Adding in the two system builders one of which isn't claiming to have 2.9% of "dead" CPU's is not going to make the blindest bit of difference to large mainstream retailers selling tens of thousands of CPU's each, adding PowerGPU to mindfactory isn't going to shift their 0.6% failure rate by so little as a fraction.

The confidence interval for the combined data set suggests that 99% of the time, the true failure rate of all Ryzens will be within 0.44% and 0.67%.

Most of the variability of the range of my estimate comes from the sample size. The point of adding PowerGPU to mindfactory data is just to use all available data.

There’s a separate estimate from a ‘major manufacturer’ of 2.9% (as per your link) that can’t be ruled out either yet if we trust that journalist.

 

[philo-sofa]

You keep including that 2.9% figure knowing this is not a failure rate.

I’ve already noted the specifics of this a few posts up man.

To reiterate there are differing definitions of failure here which complicates things. I’d add that when a CPU can’t run high frequency or low latency memory as PCWorld reported that’s not exactly a pass.