Thomas Dinges
dde740bcd7
Cycles / CUDA: Change inline rules for BVH intersection functions.
...
* On sm_30 and above there is no change (was not inlined already before), this just fixes a speed regression from yesterday. 6359c36ba4
* On sm_2x (tested with sm_21), I get a nice 8% speedup in the bmw scene with this. As a bonus, cubin compilation time and memory usage is significantly reduced. Regular cubin size went from 2.5MB to 2.0MB, Experimental one from 3.8MB to 2.5MB.
2014-10-05 03:53:51 +02:00
..
2014-10-03 16:03:49 +02:00
2014-10-05 03:53:51 +02:00
2014-10-03 16:03:49 +02:00
2014-10-03 17:51:21 +02:00
2014-10-03 18:00:45 +02:00
2014-10-04 19:00:26 +06:00
2014-09-06 15:10:17 +06:00
2014-08-19 11:40:33 +02:00
2014-08-19 11:40:33 +02:00
2014-09-25 23:28:01 +06:00
2014-08-27 10:51:50 +02:00
2014-06-13 21:59:12 +02:00
2014-04-30 16:07:27 +02:00
2014-06-15 01:09:59 +02:00
2014-10-04 19:00:26 +06:00
2013-11-18 08:48:15 +01:00
2014-09-05 16:17:39 +02:00
2014-02-27 15:01:20 +04:00
2013-12-14 15:07:46 +01:00
2014-08-28 15:15:59 +06:00
2014-08-14 16:31:53 +02:00
2013-08-18 14:16:15 +00:00
2014-03-29 13:03:49 +01:00
2014-05-05 02:19:08 +10:00
2014-09-25 23:28:01 +06:00
2014-08-14 16:09:40 +02:00
2014-09-03 21:28:43 +06:00
2014-10-04 19:00:26 +06:00
2014-04-18 13:51:29 +02:00
2014-06-14 13:49:56 +02:00
2014-09-24 13:34:28 +02:00
2014-09-26 14:23:48 +06:00
2014-08-19 11:40:33 +02:00
2014-08-19 11:40:33 +02:00
2014-08-19 11:40:33 +02:00
2014-02-03 13:56:34 +11:00
2014-09-24 13:34:28 +02:00
2014-10-04 19:00:26 +06:00
2014-10-03 12:55:31 +06:00
2014-08-19 11:40:33 +02:00
2014-08-19 11:40:33 +02:00
2014-09-25 20:07:50 +02:00
2014-08-19 11:40:33 +02:00
2014-08-26 17:02:26 +02:00