Files
opencv-MIRROR/modules
Arjan van de Ven a75840d19c Merge pull request #10468 from fenrus75:avx512-2
* Add a 512 bit codepath to the AVX512 fastConv function

this patch adds a 512 wide codepath to the fastConv() function for
AVX512 use.
The basic idea is to process the first N * 16 elements of the vector
with avx512, and then run the rest of the vector using the traditional
AVX2 codepath.

* dnn: use unaligned AVX512 load (OpenCV aligns data on 32-byte boundary)

* dnn: change "vecsize" condition for AVX512

* dnn: fix indentation
2018-01-31 16:34:12 +03:00
..
2018-01-31 04:03:35 +00:00
2018-01-19 00:23:02 +01:00
2016-11-29 01:18:10 +03:00
2018-01-19 00:23:02 +01:00
2017-12-22 08:42:21 +09:00
2017-09-04 14:41:29 +03:00
2017-10-15 12:12:00 -04:00
2018-01-19 00:23:02 +01:00
2018-01-19 00:23:02 +01:00
2018-01-19 14:06:05 +03:00
2018-01-26 17:01:02 +01:00
2017-12-08 18:33:48 +03:00
2018-01-12 19:14:00 +00:00
2017-12-21 01:33:25 -08:00