Hi,

I've been running the project on both Windows and FreeBSD, and I've noticed that FreeBSD only seems to have the v100 binary compared to vectorized SSE2/3/AVX binaries on Windows...the result being that work units are completing roughly 3x slower on FreeBSD:

FreeBSD:

53674280	23156476	21 Nov 2014, 0:34:10 UTC	21 Nov 2014, 8:31:04 UTC	Completed and validated	28,539.17	28,439.79	480.00	Period Search Application v100.00
53674318	23156485	21 Nov 2014, 0:34:10 UTC	21 Nov 2014, 8:31:04 UTC	Completed and validated	28,487.03	28,389.78	480.00	Period Search Application v100.00
53674367	23156507	21 Nov 2014, 0:34:10 UTC	21 Nov 2014, 8:31:04 UTC	Completed and validated	28,413.68	28,315.14	480.00	Period Search Application v100.00

Windows:

53826331	23232313	22 Nov 2014, 5:17:54 UTC	22 Nov 2014, 8:38:20 UTC	Completed and validated	10,078.31	9,896.56	480.00	Period Search Application v102.10 (sse3)
53826333	23232315	22 Nov 2014, 5:17:54 UTC	22 Nov 2014, 8:38:20 UTC	Completed and validated	10,127.30	9,905.70	480.00	Period Search Application v102.10 (sse3)
53826335	23232317	22 Nov 2014, 5:17:54 UTC	22 Nov 2014, 8:38:20 UTC	Completed and validated	10,036.28	9,849.80	480.00	Period Search Application v102.10 (sse3)

If someone can point me to the source, I'd be happy to provide optimized binaries for FreeBSD.

Thanks!

-Brian

Crunch3r has graciously shared his FMA4 code, so I am working on porting to FreeBSD now!