cancel
Showing results for 
Search instead for 
Did you mean: 

Drivers & Software

sbakic
Adept I

Re: gcc segmentation faults on Ryzen / Linux

I tested mprime for 8 hours on deafult UEFI setting (no OCed CPU or RAM) everything was ok. After that I tested kill-ryzen.sh on Ubuntu 17.04 it failed. I will test kill-ryzen.sh on Fedora 26 and check CPU production date and after that i will do RMA.

0 Kudos
l4c
Journeyman III

Re: gcc segmentation faults on Ryzen / Linux

I would like to report that the Ryzen 1600X I bought on december 20th shows the segmentation fault bug. It is labeled 1717SUT.

I'm thinking of selling it in a couple months to buy Ryzen 2, this looks simpler than going through the RMA process.

I'm surprised the stock of CPU produced in week 17 hasn't cleared. Or perhaps AMD is putting back on the market CPUs that it knows are defective?

0 Kudos
moralesfelipe
Journeyman III

Re: gcc segmentation faults on Ryzen / Linux

I can't believe how this issue is not fixed. I regret of buying this processor. I hope AMD ryzen sells stuck, and I will discourage every people that I know and in internet to NOT BUY AMD RYZEN.

0 Kudos
sbakic
Adept I

Re: gcc segmentation faults on Ryzen / Linux

I will update you here about my RMA request. I am from Serbia by the way:

27.12.2017 -  Filed AMD Service Request online, immediate auto response

29.12.2017 - Respone from AMD employee with question to confirm the country

02.01.2018 - Got RMA number for shipment

03.01.2018 - I sent CPU to the AMD with RMA number to Netherland (via EMS Serbia)

15.01.2018 - My CPU was delivered to AMD

18.01.2018 - Respone from AMD that they ordered CPU from Canada to Netherlands

25.01.2018 - I got new CPU

tomtor
Journeyman III

Re: gcc segmentation faults on Ryzen / Linux

I bought in the second week of december a 1700 batch 1722sus which had the problem.

In the same shop they had in addition 1800x cpus with 1718pgt code and a lot of 1700x batch 1743sus cpus. I replaced my 1700 with a 1700x for a small upgrade fee.

This 1743sus works flawlessly: clocked at 3.9ghz, core 1.35v, hyperx Fury memory rated 2666/cl16 running at 2933/cl16 on an MSI b350 gaming pro carbon. Very happy with this setup. Newer chips appear problem free.

kertp
Adept I

Re: gcc segmentation faults on Ryzen / Linux

No problem!

Also, who still hasn't RMA'd their CPU yet, if you plan to do so - it really speeds up the process if you have taken pictures of your BIOS settings in stock settings and overclocked, also pictures or screenshots of the faults in Linux. Photo of inside of your computer. Add as many information in the first e-mail as possible. I've read posts of people who say that tech support is making them do all these things anyway, so by doing it beforehand, you will save a good 2-3 days of waiting and e-mail exchange.

Overall i am very satisfied how the RMA process is being done.

pqyptixa
Adept I

Re: gcc segmentation faults on Ryzen / Linux

I started the RMA process on Jan 11 and sent my CPU some days ago (on Jan 15). I simply told AMD that my CPU had problems compiling, and that it was a known problem (gave them some links to phoronix and other website). They didn't even ask me any questions before approving the RMA.

Today, I received a couple of emails saying that the RMA had passed. I gues I'll receive my new CPU next week. So far, the process has been smooth. Now, I wish they'd send me a Zen+ CPU

mrs
Adept III

Re: gcc segmentation faults on Ryzen / Linux

We are slowly approaching a year since I had my first Ryzen segfault.

Still faulty Ryzens around... still no word from AMD...

0 Kudos
sbakic
Adept I

Re: gcc segmentation faults on Ryzen / Linux

I did my RMA and got new 1700x CPU (1733SUS). Previous CPU I could only OC to 3.8Ghz, this one I could do 4GHz  with Vcore 1.375V at idle and 1.344V at load (I will try to lower the voltage), 4.05 - 4.1 GHz is possible but need more testing with the same voltage.

Current results: 4Ghz with Vcore 1.369V at idle and 1.337V at load (I will try to lower that), 4.05 will need higher than 1.344V at load, 4.1GHz is not possible even with 1.38~1.39V at load, so  I will not test anymore 4.05 and 4.1 GHz.

I have tested new ryzen kill script for 48h, with command ./kill-ryzen 8 2 (because I have 32GB of RAM and new script needs 64GB). It passed 24 builds.

fedor_s
Journeyman III

Re: gcc segmentation faults on Ryzen / Linux

Returned my faulty 1800X (batch 1726SUS) on the 22nd of December. Just now received a replacement - 1711SUT! Of course the segfault is there, even earlier than on the previous unit. 41 seconds to fail, what a joke. I contacted the support again, we'll see what they can offer in this case. Simply unbelievable.