People are following the discussion. There is interesting information. Please carry on! :smile:

[QUOTE=kladner;424247]People are following the discussion. There is interesting information. Please carry on! :smile:[/QUOTE]

That was my impression. I also thought: the 768k thing is over, we are in the after show party here. So I thought it didn't matter much. If I am wrong I am happy to leave the field. We can leave a link to the new thread, so hopefully, interested readers can follow us over. Maybe you mersenne people can give us your opinion?

Although I am not sure what good it will do. I came here by accident, like others. I think that I stayed here because of the valuable expertise of people dealing with CPU torture the whole day. What we achieve otherwise is what you can see over at toms: mostly turning in circles. Nothing against the people there, they are just greatly in need of help, just like me. But the fluctuation is high it seems, and without any experts we do not really know what we are talking about. ;)

Interesting thing, on-topic for a change:

Several news articles on the web claim (that this or that brand patched the 768k bug first and) that it actually is MC rev. 74 that does the trick. Does anybody have more insight into this?


EDIT: I know that 6a was already stable for many people, but maybe there is more to it.

[QUOTE=pegnose;424264]Interesting thing, on-topic for a change:

Several news articles on the web claim (that this or that brand patched the 768k bug first and) that it actually is MC rev. 74 that does the trick. Does anybody have more insight into this?


EDIT: I know that 6a was already stable for many people, but maybe there is more to it.[/QUOTE]

I have a feeling their earlier microcode patch, that they apparently had already worked on before any of this Prime95 / 768K FFT stuff came to light, was a happy coincidence. They had already done some microcode update that seemingly also corrected this issue.

But we don't know what exactly the bug was, and I'm sure when Intel rolled back to their previous microcode and saw the issue, they probably realized there was a little bit more to the fix than the "accidental" solution they already had. Or maybe they realized they could optimize things somewhat and do a proper fix.

That's all speculation on my part and we'd have to wait for the Intel errata to know for sure what's included.

EDIT: Here's where I'm looking for the latest errata for Skylake. Right now it's just a Sept 2015 edition.
[URL="http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/desktop-6th-gen-core-family-spec-update.pdf"]http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/desktop-6th-gen-core-family-spec-update.pdf[/URL]

I'm just assuming at some point they'll update it with the latest.

[QUOTE=kladner;424226]Don't take sarcasm from chalsall [U]too[/U] seriously. :razz:[/QUOTE]

Sorry guys... I try to be informative, eclectic, and funny. I often fail...

Please continue the discussion. It is I who should shut up.

[QUOTE=Madpoo;424322]I have a feeling their earlier microcode patch, that they apparently had already worked on before any of this Prime95 / 768K FFT stuff came to light, was a happy coincidence. They had already done some microcode update that seemingly also corrected this issue.

But we don't know what exactly the bug was, and I'm sure when Intel rolled back to their previous microcode and saw the issue, they probably realized there was a little bit more to the fix than the "accidental" solution they already had. Or maybe they realized they could optimize things somewhat and do a proper fix.

That's all speculation on my part and we'd have to wait for the Intel errata to know for sure what's included.

EDIT: Here's where I'm looking for the latest errata for Skylake. Right now it's just a Sept 2015 edition.
[URL]http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/desktop-6th-gen-core-family-spec-update.pdf[/URL]

I'm just assuming at some point they'll update it with the latest.[/QUOTE]

Thanks for shedding a bit more light on this!

[QUOTE=chalsall;424329]Sorry guys... I try to be informative, eclectic, and funny. I often fail...

Please continue the discussion. It is I who should shut up.[/QUOTE]

C'mon now. I really thing we should drink a beer together sometime. :)

[QUOTE=pegnose;424339]C'mon now. I really thing we should drink a beer together sometime. :)[/QUOTE]

I look forward to it. But, please, not here in Bim. Our beer is *terrible*; a bit like making love in a canoe.... :smile:

I can give you a tour of the more interesting German beers...

Reply from Intel; no details...
 

This [URL="https://communities.intel.com/message/370294#370294"]post appeared on the Intel forum with relation to this issue[/URL] on 2016.01.29:

"I work for Intel Customer Support and just wanted to let you know that Intel will be publishing the next specification update in February following an update to our publication process.

This specification update will include information on this issue. No other additional information for now."

This was in response to a customer complaining (on 2016.01.26) that their MB supplier (ASUS) appeared to not even be aware about the issue....

[QUOTE=chalsall;424890]This [URL="https://communities.intel.com/message/370294#370294"]post appeared on the Intel forum with relation to this issue[/URL] on 2016.01.29:

"I work for Intel Customer Support and just wanted to let you know that Intel will be publishing the next specification update in February following an update to our publication process.

This specification update will include information on this issue. No other additional information for now."

This was in response to a customer complaining (on 2016.01.26) that their MB supplier (ASUS) appeared to not even be aware about the issue....[/QUOTE]

Sigh... I guess not every MB manufacturer is as hip and cool. They just twiddle their thumbs and wait for the official Intel errata to even start anything.

It does give me food for thought when I buy my next home system... I'll have to remember the motherboard makers who were actually on top of this and paid close attention.