Skylake and Kaby Lake HyperThreading bug:

If you own such a CPU, read THIS:

http://www.guru3d.com/news-story/debian-project-warns-turn-off-hyperthreading-with-skylake-and-kaby-lake.html

https://www.hardocp.com/news/2017/06/25/skylake_kaby_lake_processors_allegedly_have_broken_hyperthreading

Affects all OS and all CPU's of those to families, from Pentium to Xeon.

Good thing I went with the i5-7600K.

it doesn't say what the actual problem is, or how it effects the cpu, or that it even is a problem... there's no info anywhere from any reliable source...

seems to me this is someone just trolling...

it's like if someone on this forum just stated AMD cpus have a defect that could possibly maybe cause hats to start coming out of rabbits...

Let's wait a bit to see if Intel accept this defect, otherwise is like a pseudo techincal description. Not too much, not to less... just confusing :) .. or trolling.

I would not say that can also be an marketing attempt from AMD against Intel... oops i just did it :D

And this is just a mastermind command:

grep -E 'model|stepping' /proc/cpuinfo | sort -u

...c'mon ... !!! :)

Let's wait a bit to see if Intel accept this defect, otherwise is like a pseudo techincal description. Not too much, not to less... just confusing :) .. or trolling.

 

I would not say that can also be an marketing attempt from AMD against Intel... oops i just did it :D

 

And this is just a mastermind command:

 

grep -E 'model|stepping' /proc/cpuinfo | sort -u

 

...c'mon ... !!! :)

;)

LoL, what a foolish way to look at this.

Intel has already issued microcode updates to mobo vendors.

Just join the discussion on HardOCP and read it up, you will see how "acknowledged" that already is.

I am not gonna take part in DCS Hardware discussions anymore, this is such a foolish behaviour that I am way too old for.

adios amigos

Let's wait a bit to see if Intel accept this defect, otherwise is like a pseudo techincal description. Not too much, not to less... just confusing :) .. or trolling.

 

I would not say that can also be an marketing attempt from AMD against Intel... oops i just did it :D

 

And this is just a mastermind command:

 

grep -E 'model|stepping' /proc/cpuinfo | sort -u

 

...c'mon ... !!! :)

I guess you have no idea what this grep command means !?

Let's wait a bit to see if Intel accept this defect, otherwise is like a pseudo techincal description. Not too much, not to less... just confusing :) .. or trolling.

 

I would not say that can also be an marketing attempt from AMD against Intel... oops i just did it :D

 

And this is just a mastermind command:

 

grep -E 'model|stepping' /proc/cpuinfo | sort -u

 

...c'mon ... !!! :)

Intel SKL150, known, accepted and published

I guess you have no idea what this grep command means !?

Assuming too much ;)

Intel accepted and document it indeed... and this a a valuable information for sure.

Don't need to be rude or such dramatic BitMaster. I really appreciate this info as it might impact myself too, but I learned in time to not eat anything it flies. And I am not connected to everything in this world.

Assuming too much ;)

 

Intel accepted and document it indeed... and this a a valuable information for sure.

 

Don't need to be rude or such dramatic BitMaster. I really appreciate this info as it might impact myself too, but I learned in time to not eat anything it flies. And I am not connected to everything in this world.

Indded, I had a bad day as well, too much going on.

If Debian devs send this out in their mailing list, its 99.8% valid. They think twice before they hit the SEND button on such a matter as they for sure know the impact such news have

Indded, I had a bad day as well, too much going on.

 

If Debian devs send this out in their mailing list, its 99.8% valid. They think twice before they hit the SEND button on such a matter as they for sure know the impact such news have

It would be interesting to see the asm code that reproduces the bug given it appears to be short short code segments.

Under complex micro-architectural conditions, short loops of less than 64 instructions that use AH, BH, CH or DH

Cheers

For me, the questions are :

What is the probability that it happens.

If it's happens... just a crash ?

:smartass:

For me, the questions are :

What is the probability that it happens.

If it's happens... just a crash ?

 

:smartass:

Read up the links from Post #1, it can cause data loss, crashes, etc.

One guy on HardOCP can reproduce it with bittorrent and a card game running same time, his PC locks up with HT, w/o HT it works. It is sporadic and not very often but there is a chance it happens out of the blue if the apps use the functions described by Intel in SKL150.

I have my HT still on, only had a few lockups an BSOD's, which I blame to this bug now.

One guy on HardOCP can reproduce it with bittorrent and a card game running same time, his PC locks up with HT, w/o HT it works.

So only one guy can reproduce it with no data loss.

I will keep HT on. :D

It worth what it worth but I had system lock up (windows freeze, mouse moving at 0.2-0.3 fps) and it happened with and without HT....

https://lists.debian.org/debian-devel/2017/06/msg00308.html

Its a real issue.

Intel has advised for fix.