Comparing GCC 6.4.0 vs ICC 17.0.4, on xeon E5-2643 v3 @ 3.40GHz
3/28/18

spreadtestnd d 1e7 1e7 1e-6 2

Timings in sec:

single-thread:

	t1		t2
		
	GCC	ICC	GCC	ICC

1d	2.7	1.7	3.3	2.7
2d	4.5	2.7	5.0	4.5
3d	8.6	5.7	10.9	8.4

24-thread:

1d	.49	.38	.48	.36
2d	.5	.34	.48	.37
3d	.82	.62	.76	.58

Note: sorting a bit slower under ICC, but spreading is faster.

