removed atomic adds as far as possible

[erikzenker]

• Improved cpu performance
• Still a lot of time is spend in importance sampling in the distribute rays randomly kernel

[erikzenker]

PUSH !

[slizzered]

very well done, thanks. I especially like the move towards a consistent whitespace use in our code :) Commits are nice so far, I'll merge once you give the signal

[erikzenker]

I will run our test and give my ok if the result is correct.

[erikzenker]

Test crashed after 105 steps (infiniband cluster foo), but results look good!

Alpaka
0.7992
0.8297
0.8607
0.8921
0.9240
0.9563
0.9889
1.0217
1.0548
1.0880
1.1212
1.1544
1.1876
1.2206
1.2534
1.2860
1.3182
1.3501
1.3815
1.4124
1.4428
1.4726
1.5017
1.5303
1.5581
1.5852
1.6115
1.6370
1.6618
1.6857
1.7089
1.7312
1.7528
1.7735
1.7934
1.8124
1.8306
1.8481
1.8648
1.8808
1.8960
1.9105
1.9244
1.9375
1.9500
1.9619
1.9731
1.9838
1.9939
2.0035
2.0125
1.9505
1.8935
1.8410
1.7924
1.7475
1.7057
1.6669
1.6306
1.5967
1.5649
1.5350
1.5069
1.4804
1.4554
1.4318
1.4094
1.3882
1.3681
1.3490
1.3308
1.3135
1.2970
1.2812
1.2662
1.2518
1.2381
1.2249
1.2123
1.2002
1.1886
1.1775
1.1668
1.1565
1.1466
1.1371
1.1279
1.1191
1.1105
1.1023
1.0944
1.0868
1.0794
1.0722
1.0653
1.0586
1.0522
1.0459
1.0399
1.0340

[slizzered]

good work finding those slow sections!