There is an already known floating point bug when doing FFTs discovered by GIMPS on Skylake processors. The last I heard, the reason is unknown. I think @PhillyB and I have stumbled on a similar issue for Skylake and Broadwell when using Floating point variables, but this is a more generalized arithmetic in our benchmark. Intel announced that it should not affect less complex algorithms - but now that may not be the case.
Now before I go crying wolf, I need both AMD & Intel users to test so I can submit a complete picture to both Microsoft and Intel.
Latest Version of our benchmark:
http://www.bytemedev.com/BishopPublish/Bishop/publish.htm
Excuse the mess - it's not fully integrated with HWBOT
Test Setup:

If you do not have an issue you will get a score like so:

If your floating points are not working correctly, the application will freeze in an endless loop at this stage:

I have traced the issue all the way down to the JIT layer - Microsoft will need more data and some time to eliminate .NET frameworks JIT as the variable.
Now, as far as I know, the Skylake floating point bug maybe the same thing, but it is was "isolated" to just that family - that might be true, and this is something else.
If You Get The Freeze - Post The Following
CPU Model @ speed @ voltage (and temps if possible).
Users Who Have Found Issues
RagingCain - i7 4790K @ stock & overclock.
PhillyB - i7 6700K @ overclock
Nick the Slick - i7 4770K @ 4.6GHz @ 1.38v @ 66c
SuperZan - i7 3930K @ 4.6 GHz @ 1.344 @ 62c
Pointy - i5 4400S @ 3.3 GHz
rx7Racer - i7 6700K @ 4.8GHz @ 70c
Users Who Don't Have Issues
drmrlordx - AMD A10-7870K
ManofGod1000 - AMD FX 8350
Now before I go crying wolf, I need both AMD & Intel users to test so I can submit a complete picture to both Microsoft and Intel.
Latest Version of our benchmark:
http://www.bytemedev.com/BishopPublish/Bishop/publish.htm
Excuse the mess - it's not fully integrated with HWBOT
Test Setup:
If you do not have an issue you will get a score like so:
If your floating points are not working correctly, the application will freeze in an endless loop at this stage:
I have traced the issue all the way down to the JIT layer - Microsoft will need more data and some time to eliminate .NET frameworks JIT as the variable.
Now, as far as I know, the Skylake floating point bug maybe the same thing, but it is was "isolated" to just that family - that might be true, and this is something else.
If You Get The Freeze - Post The Following
CPU Model @ speed @ voltage (and temps if possible).
Users Who Have Found Issues
RagingCain - i7 4790K @ stock & overclock.
PhillyB - i7 6700K @ overclock
Nick the Slick - i7 4770K @ 4.6GHz @ 1.38v @ 66c
SuperZan - i7 3930K @ 4.6 GHz @ 1.344 @ 62c
Pointy - i5 4400S @ 3.3 GHz
rx7Racer - i7 6700K @ 4.8GHz @ 70c
Users Who Don't Have Issues
drmrlordx - AMD A10-7870K
ManofGod1000 - AMD FX 8350