1 /* 2 * Copyright (c) 2019, Intel Corporation 3 * 4 * Permission is hereby granted, free of charge, to any person obtaining a 5 * copy of this software and associated documentation files (the "Software"), 6 * to deal in the Software without restriction, including without limitation 7 * the rights to use, copy, modify, merge, publish, distribute, sublicense, 8 * and/or sell copies of the Software, and to permit persons to whom the 9 * Software is furnished to do so, subject to the following conditions: 10 * 11 * The above copyright notice and this permission notice shall be included 12 * in all copies or substantial portions of the Software. 13 * 14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS 15 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 16 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL 17 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR 18 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, 19 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR 20 * OTHER DEALINGS IN THE SOFTWARE. 21 */ 22 { 23 uchar Buffer_Index1 = (Buffer_Index & 0x0f) << 4; 24 uchar Buffer_Index2 = (Buffer_Index >> 4) << 4; 25 26 /* 27 Buffer layout after shuffle 28 _________________________________________________ 29 |_______Block0__________|_______Block1__________| 30 |_______Block2__________|_______Block3__________| 31 |_______Block4__________|_______Block5__________| 32 |_______Block6__________|_______Block7__________| 33 34 Write back buffer layout correlate to the block number#, each box stands for 1 GRF 35 _______________________________________________ 36 |____R0_________R1_____|____R2_________R3_____| 37 |____G0_________G1_____|____G2_________G3_____| 38 |____B0_________B1_____|____B2_________B3_____| 39 |____A0_________A1_____|____A2_________A3_____| 40 |____R4_________R5_____|____R6_________R7_____| 41 |____G4_________G5_____|____G6_________G7_____| 42 |____B4_________B5_____|____B6_________B7_____| 43 |____A4_________A5_____|____A6_________A7_____| 44 */ 45 46 { 47 matrix<uint, 1, 16> Temp; 48 { 49 // R1/G1/B1/A1 50 matrix<ushort, 1, 16> Alpha = DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + 6, 0); 51 52 { 53 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_0, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_RV_0, 0) * 0xFFFF))) >> 16; 54 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_0, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_0, 0), TempMask[CalculationMask]); 55 } 56 57 { 58 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_0, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_GY_0, 0) * 0xFFFF))) >> 16; 59 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_0, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_0, 0), TempMask[CalculationMask]); 60 } 61 62 { 63 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_0, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_BU_0, 0) * 0xFFFF))) >> 16; 64 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_0, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_0, 0), TempMask[CalculationMask]); 65 } 66 } 67 68 { 69 // R2/G2/B2/A2 70 matrix<ushort, 1, 16> Alpha = DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + 7, 0); 71 72 { 73 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_1, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_RV_1, 0) * 0xFFFF))) >> 16; 74 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_1, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_1, 0), TempMask[CalculationMask + 1]); 75 } 76 77 { 78 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_1, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_GY_1, 0) * 0xFFFF))) >> 16; 79 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_1, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_1, 0), TempMask[CalculationMask + 1]); 80 } 81 82 { 83 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_1, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_BU_1, 0) * 0xFFFF))) >> 16; 84 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_1, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_1, 0), TempMask[CalculationMask + 1]); 85 } 86 } 87 88 { 89 // R3/G3/B3/A3 90 matrix<ushort, 1, 16> Alpha = DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + 14, 0); 91 92 { 93 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_2, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_RV_2, 0) * 0xFFFF))) >> 16; 94 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_2, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_2, 0), TempMask[CalculationMask + 2]); 95 } 96 97 { 98 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_2, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_GY_2, 0) * 0xFFFF))) >> 16; 99 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_2, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_2, 0), TempMask[CalculationMask + 2]); 100 } 101 102 { 103 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_2, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_BU_2, 0) * 0xFFFF))) >> 16; 104 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_2, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_2, 0), TempMask[CalculationMask + 2]); 105 } 106 } 107 108 { 109 // R4/G4/B4/A4 110 matrix<ushort, 1, 16> Alpha = DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + 15, 0); 111 112 { 113 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_3, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_RV_3, 0) * 0xFFFF))) >> 16; 114 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_3, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_RV_3, 0), TempMask[CalculationMask + 3]); 115 } 116 117 { 118 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_3, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_GY_3, 0) * 0xFFFF))) >> 16; 119 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_3, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_GY_3, 0), TempMask[CalculationMask + 3]); 120 } 121 122 { 123 Temp = (((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_3, 0) * (cm_add<ushort>(0xFF00, - Alpha, SAT)))) + ((DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index2 + Channel_Offset_BU_3, 0) * 0xFFFF))) >> 16; 124 DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_3, 0).merge(Temp.format<ushort, 1, 32>().select<1, 1, 16, 2>(0, 1), DataBuffer.format<ushort, 96, 16>().select<1, 1, 16, 1>(Buffer_Index1 + Channel_Offset_BU_3, 0), TempMask[CalculationMask + 3]); 125 } 126 } 127 } 128 129 }