[QUOTE=ewmayer;418313]Anyone with access to a Skylake system of the problematic kind running Linux is welcome to try it out. The auto-build setup included with the latest Mlucas release (the one which recently entered the Debian 'unstable' branch for testing) will invoke all distinct build modes (scalar-double, sse2, avx, avx2+fma) supported by the target hardware, and create a binary for each. You want the avx2+fma binary.

No idea if Mlucas will hit the same issue, as its self-test setup is different and it is still somewhat less efficient than Prime95 (i.e. may not push the hardware quite as severely, if that is the cause of the issue in question). But worth a shot - here is my testing suggestion for would-be Skylake builders: Assuming you get a working avx2+fma binary, run the standard small/medium/large self-tests like so (this assumes the avx2+fma binary is called Mlucas_avx2)

Mlucas_avx2 -s s -iters 1000
Mlucas_avx2 -s m -iters 1000
Mlucas_avx2 -s l -iters 1000

Once we see what happens with those, we can take it from there - closest thing to George's torture test is running an actual LL-test at the desired FFT length.

George, please confirm or deny: The skylake 768K torture-test failures are using single-threaded mode? (And if so, running on just 1 core or 1 job per physical core?)[/QUOTE]

In this case though, they'd want to compile it with the AVX only compilation options (AVX2/FMA disabled).

It probably doesn't matter that it's not quite as efficient or stresses it as much. It sounds like it's not heat/voltage related anyway. So the algorithm itself would be a good enough test.

I don't know just how similar the FFT algos are between mlucas and mprime, but the differences would be a good thing... a 768K FFT on AVX that works okay on mlucas but not mprime would be significant (or if both fail, also significant).

[QUOTE=Madpoo;418402]I don't know just how similar the FFT algos are between mlucas and mprime, but the differences would be a good thing... a 768K FFT on AVX that works okay on mlucas but not mprime would be significant (or if both fail, also significant).[/QUOTE]

Indeed.

Empirically we still have a problem.

[QUOTE=Prime95;418389]Best now is to either (or both) rattle someone's cage at Intel or (and) get the ASRock engineer to reproduce it and rattle their Intel contact's cage. Alas, neither is likely to happen until after the weekend.[/QUOTE]

"Living for the Workday!"

Hmmm... Doesn't seem to have the same resonance.... :smile:

[QUOTE=chalsall;418472]"Living for the Workday!"

Hmmm... Doesn't seem to have the same resonance.... :smile:[/QUOTE]

So, did someone contact Intel today? It looks to me as this may end up requiring a new stepping or worse.

[QUOTE=tha;418545]So, did someone contact Intel today? It looks to me as this may end up requiring a new stepping or worse.[/QUOTE]

Don't know. I didn't.

But it's suddenly gotten really quiet around here. Perhaps this is a good thing?

[QUOTE=tha;418545]So, did someone contact Intel today? It looks to me as this may end up requiring a new stepping or worse.[/QUOTE]

I did not hear anything from ASRock today.

Who was it here that may have an Intel contact?

[QUOTE=Prime95;418553]I did not hear anything from ASRock today.

Who was it here that may have an Intel contact?[/QUOTE]

Dubslow's room mate.

I will try myself, but my contacts left Intel about ten years ago (on good terms).

Asrock told me that the pressure that a CPU cooler may apply on the socket is usually way beyond spec. Seems as if Noctua is one of the few companies which stay within Intels limits. Fortunately I'm using the Noctua NH-D15 ... so pressure shouldn't be an issue.

[QUOTE=Aurum;418561]Asrock told me that the pressure that a CPU cooler may apply on the socket is usually way beyond spec. Seems as if Noctua is one of the few companies which stay within Intels limits. Fortunately I'm using the Noctua NH-D15 ... so pressure shouldn't be an issue.[/QUOTE]

To support you, it doesn't explain why only the 768K test fails.

That's what they told me. We need some intel guys who are willing to support us ...

[QUOTE=Aurum;418563]That's what they told me. We need some intel guys who are willing to support us ...[/QUOTE]

Dubslow?