Ryzen CPU HyperThreads break if 100% busy and interrupted to top of Memory

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by paulmd on Friday August 04, 2017 @10:20PM (#14933881)

AMD's not doing a recall, it works decently enough for most applications. Their response is going to be "if it crashes your application, turn SMT off".

Consider they didn't even do a recall when Phenom had a showstopping TLB bug, they shipped a BIOS patch that disabled TLB entirely.

And remember, Epyc is on a new stepping of the silicon, it's possible this is already fixed on it. (Threadripper is not, however)

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by gruez on Friday August 04, 2017 @10:58PM (#14933990)

> Their response is going to be "if it crashes your application, turn SMT off".

it happens when SMT is disabled

>Epyc is on a new stepping of the silicon, it's possible this is already fixed on it. (Threadripper is not, however)

that's assuming they caught this bug, which i doubt is the case because it's only discovered now rather than being documented in the errata.

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by dom0 on Saturday August 05, 2017 @06:46AM (#14935133)

I dunno, there were several issues in the SMT implementation publicized earlier, it is entirely possible that the root cause is the same or related.

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by shizzlest on Saturday August 05, 2017 @07:51AM (#14935297)

its possible they overlooked mentioning it in the errata but still fixed it with their stepping. dont give up last hope :)

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by dis-sys on Saturday August 05, 2017 @05:11AM (#14934898)

Epyc crashes quite often, see segfault screen shot.

https://www.reddit.com/r/Amd/comments/6rmq6q/epyc_7551_minin...

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by netheril96 on Saturday August 05, 2017 @03:44AM (#14934651)

TLB, As Translation Lookaside Buffer? Won't memory access slow down dramatically if that is disabled?

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by sveiss on Saturday August 05, 2017 @04:50AM (#14934837)

Yup, and it did: http://techreport.com/review/13741/phenom-tlb-patch-benchmar...

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by dom0 on Saturday August 05, 2017 @06:44AM (#14935128)

The Phenom bug was about the L2 TLB; the L1 TLB worked just fine. It still decreased performance significantly — about 5-20 % depending on workload.

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by grwthckrmstr on Saturday August 05, 2017 @08:51AM (#14935457)

As of now, threadripper is epyc silicon with 4 ccx disabled

[ Parent | on HN ]

Re:Ryzen CPU HyperThreads break if 100% busy and interrupted... by gigatexal on Saturday August 05, 2017 @02:05AM (#14934431)

No! I was going to buy threadripper for this very reason.

[ Parent | on HN ]