Comment 222 for bug 1690085

(In reply to fin4478 from comment #136)
> I did not check what phoronix wrote. You can have a new Ryzen cpu anyway if
> you are compiling Mesa and having the segfault problem.

When has phoronix become a respectful place for info? When they published the "performance marginality problem exclusive to to certain workloads in Linux"? phoronix wrote whatever AMD told them to get some Threadripper and Epyc systems for review. I don't trust phoronix. They first escalate the segfault issue, then they buy the "marginality" excuse from AMD.

It has been reported in the original thread in AMD forums (https://community.amd.com/thread/215773) that CPUs from week 28 bought from stores still excibit segfault issues. CPUs from week 33 and after seem to have this fixed, but maybe still not 100%.

> I am Msc software engineer and when you say win virus hoover does not have
> random lockups, then it is a software problem made by wintel kernel
> developers. Intel develops Linux core and test only with their latest
> hardware.

This "Intel develops Linux core" is a statement of no value. Unless you present some specific commits from Intel that specifically affect Ryzen (AMD's Opteron for example is not affected by this idle freeze issue), you are just talking nonsense. We should better stay with facts. Fact is that even rcuo threads enabled don't fix the issue for everyone, some CPUs require C6 disabled completely. Fact is that some CPUs don't idle freeze even with rcuo threads disabled and C6 enabled. To me this is a clear indication about a hardware and not a software issue.