1// **************************************************************************** 2// * 3// * XVID MPEG-4 VIDEO CODEC 4// * - IA64 halfpel refinement - 5// * 6// * Copyright(C) 2002 Johannes Singler, Daniel Winkler 7// * 8// * This program is free software; you can redistribute it and/or modify it 9// * under the terms of the GNU General Public License as published by 10// * the Free Software Foundation; either version 2 of the License, or 11// * (at your option) any later version. 12// * 13// * This program is distributed in the hope that it will be useful, 14// * but WITHOUT ANY WARRANTY; without even the implied warranty of 15// * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 16// * GNU General Public License for more details. 17// * 18// * You should have received a copy of the GNU General Public License 19// * along with this program; if not, write to the Free Software 20// * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 21// * 22// * $Id: calc_delta_2.s,v 1.2 2009-02-19 17:07:29 Isibaar Exp $ 23// * 24// ***************************************************************************/ 25// 26// **************************************************************************** 27// * 28// * calc_delta_2.s, IA-64 halfpel refinement 29// * 30// * This version was implemented during an IA-64 practical training at 31// * the University of Karlsruhe (http://i44w3.info.uni-karlsruhe.de/) 32// * 33// **************************************************************************** 34 35(non0_2) mov sc[0] = 1 36(non0_3) mov sc[1] = 1 37 ;; 38 add mpr[0] = mpr[0], mpr[1] 39(non0_2) shl sc[0] = sc[0], iFcode 40 add mpr[2] = mpr[2], mpr[3] 41(non0_3) shl sc[1] = sc[1], iFcode 42 add mpr[4] = mpr[4], mpr[5] 43 add mpr[6] = mpr[6], mpr[7] 44 ;; 45(non0_2) add sc[0] = -1, sc[0] 46(non0_3) add sc[1] = -1, sc[1] 47 mov ret0 = 2 48 ;; 49(non0_2) add component[0] = component[0], sc[0] 50(non0_3) add component[1] = component[1], sc[1] 51 ;; 52(non0_2) shr component[0] = component[0], iFcode 53(non0_3) shr component[1] = component[1], iFcode 54 add mpr[0] = mpr[0], mpr[2] 55 add mpr[4] = mpr[4], mpr[6] 56 ;; 57(non0_2) cmp.lt cg32_0, p0 = 32, component[0] 58(non0_3) cmp.lt cg32_1, p0 = 32, component[1] 59 ;; 60(cg32_0) mov component[0] = 32 61(cg32_1) mov component[1] = 32 62 ;; 63(non0_2) addl tabaddress[0] = @gprel(mvtab#), gp 64(non0_3) addl tabaddress[1] = @gprel(mvtab#), gp 65 ;; 66(non0_2) shladd tabaddress[0] = component[0], 2, tabaddress[0] 67(non0_3) shladd tabaddress[1] = component[1], 2, tabaddress[1] 68 ;; 69(non0_2) ld4 sc[0] = [tabaddress[0]] 70(non0_3) ld4 sc[1] = [tabaddress[1]] 71 mov component[0] = dx 72 mov component[1] = dy 73 cmp.ne non0_0, p0 = 0, dx 74 cmp.gt neg_0, p0 = 0, dx 75 .pred.rel "mutex", p30, p34 //non0_0, neg_0 76 77 cmp.ne non0_1, p0 = 0, dy 78 cmp.gt neg_1, p0 = 0, dy 79 ;; 80 .pred.rel "mutex", p31, p35 //non0_1, neg_1 81 82(non0_2) add sc[0] = iFcode, sc[0] 83(non0_3) add sc[1] = iFcode, sc[1] 84 ;; 85(non0_2) add ret0 = ret0, sc[0] 86(neg_0) sub component[0] = 0, component[0] //abs 87(neg_1) sub component[1] = 0, component[1] //abs 88 ;; 89(non0_3) add ret0 = ret0, sc[1] 90 add iSAD = mpr[0], mpr[4] 91 ;; 92 93.explicit 94{.mii 95 setf.sig fmv = ret0 96(non0_0) mov sc[0] = 1 97(non0_1) mov sc[1] = 1 98 ;; 99} 100{.mfb 101 xmpy.l fmv = fmv, fQuant 102} 103{.mii 104(non0_0) shl sc[0] = sc[0], iFcode 105(non0_1) shl sc[1] = sc[1], iFcode 106 ;; 107} 108 109.default 110 111(non0_0) add sc[0] = -1, sc[0] 112(non0_1) add sc[1] = -1, sc[1] 113 ;; 114(non0_0) add component[0] = component[0], sc[0] 115(non0_1) add component[1] = component[1], sc[1] 116 ;; 117(non0_0) shr component[0] = component[0], iFcode 118(non0_1) shr component[1] = component[1], iFcode 119 ;; 120(non0_0) cmp.lt cg32_0, p0 = 32, component[0] 121(non0_1) cmp.lt cg32_1, p0 = 32, component[1] 122 ;; 123(cg32_0) mov component[0] = 32 124(cg32_1) mov component[1] = 32 125 ;; 126(non0_0) addl tabaddress[0] = @gprel(mvtab#), gp 127(non0_1) addl tabaddress[1] = @gprel(mvtab#), gp 128 ;; 129(non0_0) shladd tabaddress[0] = component[0], 2, tabaddress[0] 130(non0_1) shladd tabaddress[1] = component[1], 2, tabaddress[1] 131 getf.sig ret0 = fmv 132 ;; 133(non0_0) ld4 sc[0] = [tabaddress[0]] 134(non0_1) ld4 sc[1] = [tabaddress[1]] 135 add mpr[8] = mpr[8], ret0 136 ;; 137