[QUOTE=pegnose;423011]Is it a problem to cut the internet connection while Prime95 is doing the hard work?[/QUOTE]

You should let the client contact the server every few days, but other than that it's fine to not have an Internet connection. Note that the communications only consume a hundred bytes or so; nominally once a day.

[QUOTE=chalsall;423013]You should let the client contact the server every few days, but other than that it's fine to not have an Internet connection. Note that the communications only consume a hundred bytes or so; nominally once a day.[/QUOTE]

Ok, thx! The traffic is totally ok, I just tend to switch of my wifi router on going to work.

Nick Shih posted a [URL="http://picx.xfastest.com/nickshih/asrock/Z17OCF202.zip"]2.02 BIOS[/URL] for the Z170 OC Formula today. The new BIOS contains the newer Microcode 74.

Too bad it's not told what's the difference between 6A and 74, as 6A already fixed the 768k problem.

[QUOTE=ralleh;423026]Nick Shih posted a [URL="http://picx.xfastest.com/nickshih/asrock/Z17OCF202.zip"]2.02 BIOS[/URL] for the Z170 OC Formula today. The new BIOS contains the newer Microcode 74.

Too bad it's not told what's the difference between 6A and 74, as 6A already fixed the 768k problem.[/QUOTE]

Hopefully memory compatibility and power management issues. Many people out there are going bonkers because of randomly freezing machines (particularly idle state).

Can someone with MersenneWiki privileges please edit the stub on the bug?
[url]http://mersennewiki.org/index.php/Skylake_Bug[/url]

I have received today a beta bios from MSI, I'll give it a try and keep you informed.
But before I'd like to do some benchmarks with prime95 to see if it has performance impact.

Probably a stupid question as I haven't had the time to read through the thread yet, but does the bug affect LL results, or does it simply cause computers to freeze? If it's the former case, should we prioritize double-checking on results returned by Skylake machines?

[QUOTE=ixfd64;423140]Probably a stupid question as I haven't had the time to read through the thread yet, but does the bug affect LL results, or does it simply cause computers to freeze? If it's the former case, should we prioritize double-checking on results returned by Skylake machines?[/QUOTE]

Yes, we should double-check all skylake results run at 768k FFT size. Luckily, I did that in the time it took me to write this sentence, so we're good.

[QUOTE=VBCurtis;423141]Yes, we should double-check all skylake results run at 768k FFT size. Luckily, I did that in the time it took me to write this sentence, so we're good.[/QUOTE]

LOL...

The good (?) news is that even if it affects other FFT sizes (which seemed like a possibility, maybe remote), it seemed to trigger round-off errors.

Regardless, they'll be double-checked eventually. :smile:

[QUOTE=ixfd64;423140]Probably a stupid question as I haven't had the time to read through the thread yet, but does the bug affect LL results, or does it simply cause computers to freeze?[/QUOTE]


The (sadly) misleading thing in the news (web article headlines explicitly stating this) is that while it might lead to computers freezing (who knows; although there are so many Skylake problems out there with memory, power management etc. that actually! freeze PCs), I learned here (and by testing) that Prime95 (at least 27.x) will just drop workers.

Unfortunately, by now many people with Skylakes freezing believe that the 768k bug accounts for their problems. And the first already are disappointed learning that a bios update with the new micro code isn't helping their PCs.

I am also curious. I had solved memory issues for my machine until I finally could run memtest85 and Prime95 for whole days without a freeze and idle states wouldn't bother me as well (as opposed to other people). At that point I decided it would be worth to finally resync my software raid (which usually takes ~1d too).

And what can I say? MY PC FROZE!!! I am so p***ed off, excuse my french.

Ok, I had Prime95 running during that on all 8 cores, but that shouldn't be an issue for a 1.5k Euro machine. Seriously not.

[QUOTE=Madpoo;423153]LOL...

The good (?) news is that even if it affects other FFT sizes (which seemed like a possibility, maybe remote), it seemed to trigger round-off errors.

Regardless, they'll be double-checked eventually. :smile:[/QUOTE]

The majority of Skylake systems will probably be using v28 as well which means FMA avoiding this bug.