f6224eae74
Tested by jsg@, kettenis@ and myself on a wide range of intel cards.
132 lines
5.5 KiB
Plaintext
132 lines
5.5 KiB
Plaintext
/*
|
|
* Copyright © 2008 Intel Corporation
|
|
*
|
|
* Permission is hereby granted, free of charge, to any person obtaining a
|
|
* copy of this software and associated documentation files (the "Software"),
|
|
* to deal in the Software without restriction, including without limitation
|
|
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
|
* and/or sell copies of the Software, and to permit persons to whom the
|
|
* Software is furnished to do so, subject to the following conditions:
|
|
*
|
|
* The above copyright notice and this permission notice (including the next
|
|
* paragraph) shall be included in all copies or substantial portions of the
|
|
* Software.
|
|
*
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
|
|
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
* SOFTWARE.
|
|
*
|
|
* Author:
|
|
* Zou Nan hai <nanhai.zou@intel.com>
|
|
* Zhang Hua jun <huajun.zhang@intel.com>
|
|
* Xing Dong sheng <dongsheng.xing@intel.com>
|
|
*
|
|
*/
|
|
|
|
/* shader for forward and backward predict mc
|
|
*/
|
|
/* (x', y') = (x, y) + (motion_vector1.x >> 1, motion_vector1.y >> 1)
|
|
/* (x'', y'') = (x, y) + (motion_vector2.x >> 1, motion_vector2.y >> 1)
|
|
/* f(x, y) = (motion_forward(x`, y`) + motion_backward(x'', y'') + 1) / 2 */
|
|
|
|
//Save payload
|
|
mov (8) g76.0<1>UD g1.0<8,8,1>UD {align1};
|
|
//mov (8) g77.0<1>UD g2.0<8,8,1>UD {align1};
|
|
include(`block_clear.g4i')
|
|
|
|
mov (8) g115.0<1>UD g1.0<8,8,1>UD {align1};
|
|
mov (1) g1.8<1>UD 0x0070007UD {align1};
|
|
mov (8) g2.0<1>UD g1.0<8,8,1>UD {align1};
|
|
/* Y */
|
|
//Forward
|
|
asr (2) g1.14<1>W g115.14<2,2,1>W 1W {align1};
|
|
add (2) g2.0<1>UD g115.0<2,2,1>UD g1.14<2,2,1>W {align1};
|
|
define(`input_surface', `4')
|
|
define(`mv1', `g115.14')
|
|
define(`mv2', `g115.16')
|
|
include(`motion_frame_y.g4i')
|
|
mov (16) g52.0<1>UD g28.0<16,16,1>UD {align1 compr};
|
|
mov (16) g54.0<1>UD g30.0<16,16,1>UD {align1 compr};
|
|
mov (16) g56.0<1>UD g32.0<16,16,1>UD {align1 compr};
|
|
mov (16) g58.0<1>UD g34.0<16,16,1>UD {align1 compr};
|
|
mov (16) g60.0<1>UD g36.0<16,16,1>UD {align1 compr};
|
|
mov (16) g62.0<1>UD g38.0<16,16,1>UD {align1 compr};
|
|
mov (16) g64.0<1>UD g40.0<16,16,1>UD {align1 compr};
|
|
mov (16) g66.0<1>UD g42.0<16,16,1>UD {align1 compr};
|
|
//Backward
|
|
asr (2) g1.14<1>W g115.18<2,2,1>W 1W {align1};
|
|
add (2) g2.0<1>UD g115.0<2,2,1>UD g1.14<2,2,1>W {align1};
|
|
define(`input_surface', `7')
|
|
define(`mv1', `g115.18')
|
|
define(`mv2', `g115.20')
|
|
include(`motion_frame_y.g4i')
|
|
|
|
//Average
|
|
avg.sat (16) g28.0<1>UW g28.0<16,16,1>UW g52.0<16,16,1>UW {align1};
|
|
avg.sat (16) g29.0<1>UW g29.0<16,16,1>UW g53.0<16,16,1>UW {align1};
|
|
avg.sat (16) g30.0<1>UW g30.0<16,16,1>UW g54.0<16,16,1>UW {align1};
|
|
avg.sat (16) g31.0<1>UW g31.0<16,16,1>UW g55.0<16,16,1>UW {align1};
|
|
avg.sat (16) g32.0<1>UW g32.0<16,16,1>UW g56.0<16,16,1>UW {align1};
|
|
avg.sat (16) g33.0<1>UW g33.0<16,16,1>UW g57.0<16,16,1>UW {align1};
|
|
avg.sat (16) g34.0<1>UW g34.0<16,16,1>UW g58.0<16,16,1>UW {align1};
|
|
avg.sat (16) g35.0<1>UW g35.0<16,16,1>UW g59.0<16,16,1>UW {align1};
|
|
avg.sat (16) g36.0<1>UW g36.0<16,16,1>UW g60.0<16,16,1>UW {align1};
|
|
avg.sat (16) g37.0<1>UW g37.0<16,16,1>UW g61.0<16,16,1>UW {align1};
|
|
avg.sat (16) g38.0<1>UW g38.0<16,16,1>UW g62.0<16,16,1>UW {align1};
|
|
avg.sat (16) g39.0<1>UW g39.0<16,16,1>UW g63.0<16,16,1>UW {align1};
|
|
avg.sat (16) g40.0<1>UW g40.0<16,16,1>UW g64.0<16,16,1>UW {align1};
|
|
avg.sat (16) g41.0<1>UW g41.0<16,16,1>UW g65.0<16,16,1>UW {align1};
|
|
avg.sat (16) g42.0<1>UW g42.0<16,16,1>UW g66.0<16,16,1>UW {align1};
|
|
avg.sat (16) g43.0<1>UW g43.0<16,16,1>UW g67.0<16,16,1>UW {align1};
|
|
|
|
/* UV */
|
|
//Forward
|
|
shr (2) g1.0<1>UD g115.0<2,2,1>UD 1UD {align1};
|
|
asr (2) g115.14<1>W g115.14<2,2,1>W 1W {align1};
|
|
asr (2) g1.14<1>W g115.14<2,2,1>W 1W {align1};
|
|
add (2) g2.0<1>UD g1.0<2,2,1>UD g1.14<2,2,1>W {align1};
|
|
|
|
define(`input_surface1', `5')
|
|
define(`input_surface2', `6')
|
|
define(`mv1', `g115.14')
|
|
define(`mv2', `g115.16')
|
|
mov (1) g2.8<1>UD 0x007000fUD {align1};
|
|
include(`motion_frame_uv.g4i')
|
|
mov (16) g122.0<1>UB g44.0<16,16,2>UB {align1};
|
|
mov (16) g122.16<1>UB g45.0<16,16,2>UB {align1};
|
|
mov (16) g123.0<1>UB g46.0<16,16,2>UB {align1};
|
|
mov (16) g123.16<1>UB g47.0<16,16,2>UB {align1};
|
|
mov (16) g124.0<1>UB g48.0<16,16,2>UB {align1};
|
|
mov (16) g124.16<1>UB g49.0<16,16,2>UB {align1};
|
|
mov (16) g125.0<1>UB g50.0<16,16,2>UB {align1};
|
|
mov (16) g125.16<1>UB g51.0<16,16,2>UB {align1};
|
|
//Backward
|
|
asr (2) g115.18<1>W g115.18<2,2,1>W 1W {align1};
|
|
asr (2) g1.14<1>W g115.18<2,2,1>W 1W {align1};
|
|
add (2) g2.0<1>UD g1.0<2,2,1>UD g1.14<2,2,1>W {align1};
|
|
define(`input_surface1', `8')
|
|
define(`input_surface2', `9')
|
|
define(`mv1', `g115.18')
|
|
define(`mv2', `g115.20')
|
|
mov (1) g2.8<1>UD 0x007000fUD {align1};
|
|
include(`motion_frame_uv.g4i')
|
|
|
|
//Average
|
|
avg.sat (16) g44.0<1>UW g44.0<16,16,1>UW g122.0<16,16,1>UB {align1};
|
|
avg.sat (16) g45.0<1>UW g45.0<16,16,1>UW g122.16<16,16,1>UB {align1};
|
|
avg.sat (16) g46.0<1>UW g46.0<16,16,1>UW g123.0<16,16,1>UB {align1};
|
|
avg.sat (16) g47.0<1>UW g47.0<16,16,1>UW g123.16<16,16,1>UB {align1};
|
|
avg.sat (16) g48.0<1>UW g48.0<16,16,1>UW g124.0<16,16,1>UB {align1};
|
|
avg.sat (16) g49.0<1>UW g49.0<16,16,1>UW g124.16<16,16,1>UB {align1};
|
|
avg.sat (16) g50.0<1>UW g50.0<16,16,1>UW g125.0<16,16,1>UB {align1};
|
|
avg.sat (16) g51.0<1>UW g51.0<16,16,1>UW g125.16<16,16,1>UB {align1};
|
|
|
|
include(`addidct.g4i')
|
|
|
|
send (16) 0 acc0<1>UW g0<8,8,1>UW
|
|
thread_spawner(0, 0, 0) mlen 1 rlen 0 { align1 EOT};
|