So it turns out that on certain intel CPU's sandsifter will terminate prematurely (segfault in injector) when the injector is not compiled as static.
#14
This has been resolved and tested. It turns out that a majority of the test results where thus broken as they miss a set of data after either of these instructions:
660f8fff000000000000000000000000
660f1fff000000000000000000000000
The instruction where it should end on is:
f3dfc7 or dfc7 for short.
CPU's that were unaffected by this include:
* All AMD Ryzen CPU's
* Intel Pentium B970 (Sandy Bridge)
* Intel i3-3120m (Ivy Bridge) - Probably a Sandy Bridge part actually
All other tests terminated prematurely result in a square 200Mb Log file size.
This includes:
@jotebe
- core2duo P8400
- Intel i3-4130T
- Intel Xeon E3-1225 v3
- Intel i5-2540M
@killerkalamari
@Lewiscowles1986
I sincerely apologize for not catching this earlier.
I would like to ask you to rerun the test and contribute the full result.
You are of course free to do so at your own will & time.
Sincerely
Rigo
So it turns out that on certain intel CPU's sandsifter will terminate prematurely (segfault in injector) when the injector is not compiled as static.
#14
This has been resolved and tested. It turns out that a majority of the test results where thus broken as they miss a set of data after either of these instructions:
660f8fff000000000000000000000000660f1fff000000000000000000000000The instruction where it should end on is:
f3dfc7ordfc7for short.CPU's that were unaffected by this include:
* All AMD Ryzen CPU's
* Intel Pentium B970 (Sandy Bridge)
* Intel i3-3120m (Ivy Bridge) - Probably a Sandy Bridge part actually
All other tests terminated prematurely result in a square 200Mb Log file size.
This includes:
@jotebe
@killerkalamari
@Lewiscowles1986
I sincerely apologize for not catching this earlier.
I would like to ask you to rerun the test and contribute the full result.
You are of course free to do so at your own will & time.
Sincerely
Rigo