New Ryzen Is Running Solid Under Linux, No Compiler Segmentation Fault Issue

[duby229]

I'm waiting for my RMA chips back (both my 1700 and 1700X are affected with the SEGV BUG). Both those machines also fail to run the Ashes of the Singularity - Escalation built in benchmark in Vulkan or DX12 mode (DX11 is OK) crashing to desktop (disabling SMT gives it stability) .

Couple of people reported this issue as well.

I'm wondering if it's possible to determine whether the Ashes of a Singularity crash to desktop is caused by the same fault? Or is it a separate problem that happens to be a coincidence?

[attaboy]

Does this occurs in a virtual machine also? I'm running the script on a ubuntu vm on a windows 10 host but it doesn't seem to segfaulsegfault.

[attaboy]

Does the problem occurs on vm? I'm using ubuntu on a windows 10 host via virtualbox. Running the script doesn't seem to segfault (1h in). Btw I've experienced some freeze on idle in win10 but only with tight ram timings. 1700 built in june.

[Funks]

I'm wondering if it's possible to determine whether the Ashes of a Singularity crash to desktop is caused by the same fault? Or is it a separate problem that happens to be a coincidence?

I guess i'll found out if they are related or not. I'm supposed to be getting one of my RMA chips back monday (1700X) , if it starts magically working with AOTS:Singularity DX12 or Vulkan Benchmark without crashing to desktop (nothing else changed aside from the CPU on my system) then I'll go hmmmm (my main system).

When I get the second RMA in (1700) and it also starts magically working with AOTS on my son's system with no change anything aside from the CPU - then all hell gonna break lose as it means Wintel Gamers are also affected.

Main System: Asrock Taichi X370 (v3 BIOS), RM650i Power Supply, H110i AIO, 32-Gigs of RAM, R7-1700X (not overclocked, BIOS defaults)
Son's System: Gigabyte AB350 Gaming 3 (vF7 BIOS), CX550M Power Supply, AMD Wraith Spire, R7-1700 (not overclocked, BIOS defaults)

These two machines have the Linux SEGV issue as well.

[Melcar]

Well crap. Guess my planned Ryzen build will have to wait a bit longer.

[AndyChow]

Well crap. Guess my planned Ryzen build will have to wait a bit longer.

Why? It's fixed now.

[Funks]

There is a very good chance that all the pre-week-25 Ryzens are sold out already in most places. There is pretty high demand for Ryzen CPUs worldwide. If you buy one now, I would assume it would very likely be a newer one, especially if you buy from Amazon or other big vendor that goes through a high volume of product. If you buy in person / in a physical shop, you can have a look at the CPU through the window in the box and see the number for yourself, to be certain. Even if you buy online and get one of the older CPUs, you can send it to AMD and they will replace it with a new one.

I live in the Bay Area, my local Fry's and the Central Computers have all old processors. Good luck finding the 1725+ anywhere - the shops are still sitting on a whole bunch of old stock. You'll probably see the new ones at the end of the year on the shelves (unless you buy a threadripper that is) or if AMD decides to proactively replace old stock.

[scorpio810]

I'm wondering if it's possible to determine whether the Ashes of a Singularity crash to desktop is caused by the same fault? Or is it a separate problem that happens to be a coincidence?

I think you have tried prime95 blend test and memtest86+ multithread mode before?
Disable OpCache control in your BIOS and try again ..

[chithanh]

Why? It's fixed now.

It is fixed for CPUs that were produced after week 28 (or so?). But you can't order specifically those anywhere. Unless you show up in person in a shop, you risk getting a broken Ryzen and have to go through RMA, which can take several weeks.

Until potentially affected CPUs have been purged from the channel, beware of buying Ryzens.

[Ricky@1]

not a big deal, even intel skylake and kabylake also got segfault, and intel release a microcode update to workaround the issue.

http://gallium.inria.fr/blog/intel-skylake-bug/