1/* Assemble with ".../intel-gen4asm/src/intel-gen4asm -g 7" */ 2 3 4/* Move pixels into g10-g13. The pixel shaader does not load what you want. It 5 * loads the input data for a plane function to calculate what you want. The 6 * following is boiler plate code to move our normalized texture coordinates 7 * (u,v) into g10-g13. It does this 4 subspans (16 pixels) at a time. 8 * 9 * This should do the same thing, but it doesn't work for some reason. 10 * pln(16) g10 g6<0,1,0>F g2<8,8,1>F { align1 }; 11 * pln(16) g12 g6.16<1>F g2<8,8,1>F { align1 }; 12 */ 13/* U */ 14pln (8) g10<1>F g6.0<0,1,0>F g2.0<8,8,1>F { align1 }; /* pixel 0-7 */ 15pln (8) g11<1>F g6.0<0,1,0>F g4.0<8,8,1>F { align1 }; /* pixel 8-15 */ 16/* V */ 17pln (8) g12<1>F g6.16<0,1,0> g2.0<8,8,1>F { align1 }; /* pixel 0-7 */ 18pln (8) g13<1>F g6.16<0,1,0> g4.0<8,8,1>F { align1 }; /* pixel 8-15 */ 19 20 21/* Next the we want the sampler to fetch the src texture (ie. src buffer). This 22 * is done with a pretty simple send message. The output goes to g112, which is 23 * exactly what we're supposed to use in our final send message. 24 * In intel-gen4asm, we should end up parsed by the following rule: 25 * predicate SEND execsize dst sendleadreg sndopr directsrcoperand instoptions 26 * 27 * Send message descriptor: 28 * 28:25 = message len = 4 // our 4 registers have 16 pixels 29 * 24:20 = response len = 8 // Each pixel is RGBA32, so we need 8 registers 30 * 19:19 = header present = 0 31 * 18:17 = SIMD16 = 2 32 * 16:12 = TYPE = 0 (regular sample) 33 * 11:08 = Sampler index = ignored/0 34 * 7:0 = binding table index = src = 1 35 * 0x8840001 36 * 37 * Send message extra descriptor 38 * 5:5 = End of Thread = 0 39 * 3:0 = Target Function ID = SFID_SAMPLER (2) 40 * 0x2 41 */ 42 43send(16) g112 g10 0x2 0x8840001 { align1 }; 44 45/* g112-g119 now contains the sample source input, and all we must do is write 46 * it out to the destination render target. This is done with the send message 47 * as well. The only extra bits are to terminate the pixel shader. 48 * 49 * Send message descriptor: 50 * 28:25 = message len = 8 // 16 pixels RGBA32 51 * 24:20 = response len = 0 52 * 19:19 = header present = 0 53 * 17:14 = message type = Render Target Write (12) 54 * 12:12 = Last Render Target Select = 1 55 * 10:08 = Message Type = SIMD16 (0) 56 * 07:00 = Binding Table Index = dest = 0 57 * 0x10031000 58 * 59 * Send message extra descriptor 60 * 5:5 = End of Thread = 1 61 * 3:0 = Target Function ID = SFID_DP_RC (5) 62 * 0x25 63 */ 64send(16) null g112 0x25 0x10031000 { align1, EOT }; 65 66/* vim: set ft=c ts=4 sw=2 tw=80 et: */ 67