Some Ryzen Linux Users Are Facing Issues With Heavy Compilation Loads

Who the hell gives a shit if someone is in denial or not over a stupid shitty internet post. How about damaging your cpu because it pulled out of the socket, damaged some pins and now you have to buy a new one, when you could have left well enough alone. :D Shit, how about we get real and realize, no one cares about stupid internet posts. (Yes, I posted here, no, that does not mean I care because, after all, this is not a 3rd grade school playground. :D)
 
Who the hell gives a shit if someone is in denial or not over a stupid shitty internet post. How about damaging your cpu because it pulled out of the socket, damaged some pins and now you have to buy a new one, when you could have left well enough alone. :D Shit, how about we get real and realize, no one cares about stupid internet posts. (Yes, I posted here, no, that does not mean I care because, after all, this is not a 3rd grade school playground. :D)

Let me explain it clearly:

My head is on fire , must be AMD micro op cache bug ...
 
Last edited:
Let me explain it clearly:

My head is on fire , must be AMD micro op cache bug ...

Depends on if it is on fire for what you took or if it was lit on fire from without. :D I am not particularly concerned one way or the other. If there is a problem, it needs to be fixed. But if there is not, then there is nothing here to see. Notice the ones chiming in with the fanboy tag almost immediately? You have to wonder why.
 
I don't see what the big deal is.
IF there's a problem, it will get fixed. Bugs and problems are found all the time, both in hardware and software. A new driver, UEFI/AGESA update or similar and the problem is gone. I doubt it's so serious that AMD will have to recall every Ryzen sold so far.
 
IF there's a problem, it will get fixed.

However as I said earlier in this thread I can't consider purchasing a Ryzen 7 based workstation for home or work until this is clearly fixed. For home the workstation I want to build will replace a machine that is approaching 11 years old and has been stable for its entire lifetime of 24/7 usage. At work I could get fired if I purchase a 1/2 dozen of these and they cause random application crashes at high load doing our medical imaging research. And yes we do a lot of our medical imaging research under linux.

Edit: With this said if it was currently winter (where I would have lots of free time to tinker) I could be convinced to try anyways (at least at home).
 
Last edited:
However as I said I can't consider purchasing for home or work until it is clearly fixed. For home the workstation will replace a machine that is approaching 11 years old and has been stable for its entire lifetime of 24/7 usage. At work I could get fired if I purchase a 1/2 dozen of these and they cause random application crashes at high load (doing our medical imaging research) under linux.
I'm 100% in the same boat. I have a desktop machine I use to work from home that is 4 years old & it's @ stock speeds. I really want to buy Ryzen to replace home linux box (which I use for virtualization, gaming, compiling, etc) but I really need it to be 100% rock solid.
 
It is not what the OP stated a micro op cache bug that can be fixed by load line calibration. It would be a very dubious problem if that was needed for a stock platform and also much more apparent in stress testing not exclusively show up under Linux only.

Load line calibration is in general only used for when a cpu comes out of idle or sleep and gets a stupid amount of workload to deal with not being able to switch to a higher voltage fast enough to handle it.
It makes little to no sense that this would actually "fix" a cache bug Given with the current state of AM4 bios I wonder if that has really got an impact overall, it shows that some systems prolly don't get enough power under load rather then a bug with a micro op cache (because that is all that LLC is doing really).

You could say that a workload as rendering blender or cinebench would nearly as much show the same symptoms as compiling source under GCC in Linux. That is why I am wondering if GCC under Windows would show the same problem(s).

Whatever, if I lower my own LLC from 6 to any lower value my OC will crash in p95 and IBT. That much to "it only affects ramp up voltage". What you say is true, just not the whole truth to what LLC affects.
 
Whatever, if I lower my own LLC from 6 to any lower value my OC will crash in p95 and IBT. That much to "it only affects ramp up voltage". What you say is true, just not the whole truth to what LLC aff
ects.


Odd that the person from Asus says the same thing
 
No official answer still from AMD. They continue investigating the issue, and their last recommendation was to disable uop or SMT

https://community.amd.com/thread/215773

DragonFly BSD has a partial workaround for the RyZen bug. The bug is not solved but its frequency of reproduction is reduced

http://gitweb.dragonflybsd.org/dragonfly.git/commitdiff/b48dd28447fc8ef62fbc963accd301557fd9ac20

from your link:

alfonsor 13-Jun-2017 10:42 (in response to atipro)


gcc rarely segfaults; it is bash that crashes 99% of the times; so, or it is a combination of various bugs in various places or it is ryzen's fault; again: the real question is: why many users have no problem at all?

You keep slamming Ryzen you give advice on the issue by disabling SMT while plenty of people in the thread say it does not work , you claim AMD give advice to disable this ....
In all of your links you never show anywhere how people can solve it.

Other people don't have the bug. So what is it? link the parts where people can solve this ....
 
Last edited:
from your link:



You keep slamming Ryzen you give advice on the issue by disabling SMT while plenty of people in the thread say it does not work , you claim AMD give advice to disable this ....
In all of your links you never show anywhere how people can solve it.

Other people don't have the bug. So what is it? link the parts where people can solve this ....
seems more like a software conflict or as you stated an unstable OC or inability to maintain stock stability. The part that is bothersome is it isn't everybody, therefore not likely an inherent issue with Ryzen.
 
And the funny story here is that Ryzen holds much lower value if i have to fucking overpay for my motherboard for it work properly.

But hey, in the end marketing stunt from AMD worked, unlocked Ryzens are unlocked, and we don't care if it won't work once overclocked if you paid less than $130 for a mobo.

You buy a shit board, you get what you pay for. This was always the case regardless of platform.
 
Some users reported that they were not overclocking and still had the random crashes under high load.
 
Some users reported that they were not overclocking and still had the random crashes under high load.

Running the ram past stock speeds is still overclocking. It can seem stable but its not, got bit by that myself running at 3200, took a ton of messing with settings to get it to be actually stable.
 
Running the ram past stock speeds is still overclocking. It can seem stable but its not, got bit by that myself running at 3200, took a ton of messing with settings to get it to be actually stable.
None of that makes me feel like Ryzen is production ready then. So hard pass for now.
 
None of that makes me feel like Ryzen is production ready then. So hard pass for now.

I am using double rank dimms which is not guaranteed to work at 3200 speeds. No different on Intel when you choose to go above recommended ram speeds. It ran flawlessly at recommended ram speeds on the Ryzen so I dont see the issue.
 
Looks like XMP is gaming and not mission critical.
Edit- If it is memory management, that points to the chipset.
 
Last edited:
You keep slamming Ryzen you give advice on the issue by disabling SMT while plenty of people in the thread say it does not work , you claim AMD give advice to disable this ....
In all of your links you never show anywhere how people can solve it.

It is AMD who is giving the advice to disable uop or SMT to try to solve the issue. I linked twice to the amdcomunity thread where AMD gives this advice to RyZen users. AMD claims is investigating the issue and they will return to the thread to give more information which this was available.

I also gave a link to a patch is being introduced on BSD. The trick doesn't solve the RyZen bug, but it alleviates the problem by reducing the frequency of reproduction of the fault.
 
seems more like a software conflict or as you stated an unstable OC or inability to maintain stock stability. The part that is bothersome is it isn't everybody, therefore not likely an inherent issue with Ryzen.

The board can't handle the load and somehow it's the CPU's fault?

We're sure it's not an overclocking or stability issue?

The bug was identified. It affects to people is not overclocking and it affects people with X370 mobos. All of this has been stated in this thread before.

Last unofficial word I have is a microcode update is in the works.
 
The bug was identified. It affects to people is not overclocking and it affects people with X370 mobos. All of this has been stated in this thread before.

Last unofficial word I have is a microcode update is in the works.

Funny you cant link all of the sudden to this conformation of a bug by AMD?
 
About B350 having unlocked multiplier being a mockery. Besides, there is always the bug from OP, no matter how much you would like to be in denial.

B350 is a chipset not a processor so it has nothing to do with Ryzen being unlocked. There are good B350 boards but most are cheap and should not be used with a overclocked 8 core chip running 24/7. Most that bought a B350 likely did so cause they could not find a X370 at the time as they were in high demand, patience is better then regret. So if your gonna chummy up with the J guy then show me where AMD has said this is a bug and they will fix it?
 
There are good B350 boards but most are cheap and should not be used with a overclocked 8 core chip running 24/7
Most can't even handle stock 1800X, however, let alone any meaningful OC.
Most that bought a B350 likely did so
Don't kid yourself, most bought B350 because it thought it would suffice. Bought 65W tdp, you know.
 
It is AMD who is giving the advice to disable uop or SMT to try to solve the issue. I linked twice to the amdcomunity thread where AMD gives this advice to RyZen users. AMD claims is investigating the issue and they will return to the thread to give more information which this was available.

I also gave a link to a patch is being introduced on BSD. The trick doesn't solve the RyZen bug, but it alleviates the problem by reducing the frequency of reproduction of the fault.

The link you posted holds so many comments by so many different people even the ones that don't have the bug you fail to produce _anything_ cept that some people have it. You call it in bug in Ryzen yet AMD has trouble reproducing it or there would be a fix.
Most can't even handle stock 1800X, however, let alone any meaningful OC.

Don't kid yourself, most bought B350 because it thought it would suffice. Bought 65W tdp, you know.

High work load will cause problems for the current B350 based hardware. You can see from the linked video that you can overclock and run 4 cores on a B350 pretty well..
 
Don't kid yourself, most bought B350 because it thought it would suffice.
Or because they got a mATX case. There is precisely one X370 board in that form factor and beside the chipset it has identical feature set to its B350 counterpart. If there was a good X370 mATX board, it would sell like hotcakes.
 
Most can't even handle stock 1800X, however, let alone any meaningful OC.

Don't kid yourself, most bought B350 because it thought it would suffice. Bought 65W tdp, you know.

Since I actually involve myself with AMD not just complain about it, most that went to a B350 was due to the fact X370 was impossible to find but a B350 was in stock. Dont overclock that 8 core and stress it 24/7 and it's fine, overclocking stresses the components and you know that. Overclock a Intel and let me know if it's in TDP, oh wait it's not. B350 is not a server chipset and should not be used that way. Hell the X370 boards are far cheaper then the crappiest Intel boards, so saving a buck on something as vital as the power delivery to your chip and expecting 24/7 use out of it was silly.
 
Since I actually involve myself with AMD not just complain about it, most that went to a B350 was due to the fact X370 was impossible to find but a B350 was in stock. Dont overclock that 8 core and stress it 24/7 and it's fine, overclocking stresses the components and you know that. Overclock a Intel and let me know if it's in TDP, oh wait it's not. B350 is not a server chipset and should not be used that way. Hell the X370 boards are far cheaper then the crappiest Intel boards, so saving a buck on something as vital as the power delivery to your chip and expecting 24/7 use out of it was silly.
Ask most anyone here why they are here? Over clocking. You could take a 300A to 1.5x it's retail clock speeds all day long. Almost any Intel chip can be OCd with no issues. And yet we have to baby the AMD part? Fuuuuuuck that.
 
Ask most anyone here why they are here? Over clocking. You could take a 300A to 1.5x it's retail clock speeds all day long. Almost any Intel chip can be OCd with no issues. And yet we have to baby the AMD part? Fuuuuuuck that.

300A is a long time ago and a silly thing to even open a statement on modern processors. Ryzen 8 core can be overclocked just fine, you just need to use the proper board for it. If that is to hard for you to understand then Dell can make you a nice computer to use. Tho having a choice is a hard thing for a Intel fan to understand.
 
Un
300A is a long time ago and a silly thing to even open a statement on modern processors. Ryzen 8 core can be overclocked just fine, you just need to use the proper board for it. If that is to hard for you to understand then Dell can make you a nice computer to use. Tho having a choice is a hard thing for a Intel fan to understand.
unless motherboard manufacturers became inept overnight a cheap motherboard doesn't break a CPU. I've over clocked Intel chips to be prime95 stable for days on lowend MSI boards and unlocked K CPUs.
 
Back
Top