Row-major vs. column-major mat mult.

vliw4_shader1.txt

#version 430 core

in vec4 pos;
out vec4 pos_out;

buffer mats {
    mat4x4 Matrices[];
};

void main()
{
    pos_out = Matrices[0] * pos;
}

; --------  Disassembly --------------------
00 CALL_FS 
01 ALU: ADDR(32) CNT(8) 
      0  x: MOV         R2.x,  0.0f      
         y: MOV         R0.y,  (0x0000000C, 1.681558157e-44f).x      
         z: MOV         R0.z,  (0x00000008, 1.121038771e-44f).y      
         w: MOV         R2.w,  1.0f      
      1  x: MOV         R0.x,  (0x00000004, 5.605193857e-45f).x      
         w: MOV         R0.w,  0.0f      
02 TEX: ADDR(64) CNT(4) 
      2  VFETCH R3, R0.y, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      3  VFETCH R4, R0.z, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      4  VFETCH R5, R0.x, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      5  VFETCH R0, R0.w, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
03 EXP_DONE: POS0, R2.xxxw
04 ALU: ADDR(40) CNT(16) 
      6  x: MUL         ____,  R1.w,  R3.y      
         y: MUL         ____,  R1.w,  R3.w      
         z: MUL         ____,  R1.w,  R3.z      
         w: MUL         ____,  R1.w,  R3.x      
      7  x: MULADD      ____,  R1.z,  R4.y,  PV6.x      
         y: MULADD      ____,  R1.z,  R4.w,  PV6.y      
         z: MULADD      ____,  R1.z,  R4.z,  PV6.z      
         w: MULADD      ____,  R1.z,  R4.x,  PV6.w      
      8  x: MULADD      ____,  R1.y,  R5.y,  PV7.x      
         y: MULADD      ____,  R1.y,  R5.w,  PV7.y      
         z: MULADD      ____,  R1.y,  R5.z,  PV7.z      
         w: MULADD      ____,  R1.y,  R5.x,  PV7.w      
      9  x: MULADD      R0.x,  R1.x,  R0.x,  PV8.w      
         y: MULADD      R0.y,  R1.x,  R0.y,  PV8.x      
         z: MULADD      R0.z,  R1.x,  R0.z,  PV8.z      
         w: MULADD      R0.w,  R1.x,  R0.w,  PV8.y      
05 EXP_DONE: PARAM0, R0
06 ALU: ADDR(56) CNT(1) 
     10  x: NOP         ____      
07 NOP NO_BARRIER 
08 END 
END_OF_PROGRAM

vliw4_shader2.txt

#version 430 core

in vec4 pos;
out vec4 pos_out;

buffer mats {
    mat4x4 Matrices[];
};

void main()
{
    pos_out = pos * Matrices[0];
}

; --------  Disassembly --------------------
00 CALL_FS 
01 ALU: ADDR(32) CNT(8) 
      0  x: MOV         R0.x,  0.0f      
         y: MOV         R0.y,  0.0f      
         z: MOV         R0.z,  (0x00000004, 5.605193857e-45f).x      
         w: MOV         R0.w,  1.0f      
      1  y: MOV         R2.y,  (0x0000000C, 1.681558157e-44f).x      
         z: MOV         R2.z,  (0x00000008, 1.121038771e-44f).y      
02 TEX: ADDR(64) CNT(4) 
      2  VFETCH R3, R0.y, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      3  VFETCH R4, R0.z, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      4  VFETCH R5, R2.z, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
      5  VFETCH R2, R2.y, fc150  FORMAT(32_32_32_32_FLOAT) 
         FETCH_TYPE(NO_INDEX_OFFSET) 
03 EXP_DONE: POS0, R0.xxxw
04 ALU: ADDR(40) CNT(16) 
      6  x: DOT4_e      R3.x,  R1.x,  R3.x      
         y: DOT4_e      ____,  R1.y,  R3.y      
         z: DOT4_e      ____,  R1.z,  R3.z      
         w: DOT4_e      ____,  R1.w,  R3.w      
      7  x: DOT4_e      ____,  R1.x,  R4.x      
         y: DOT4_e      R3.y,  R1.y,  R4.y      
         z: DOT4_e      ____,  R1.z,  R4.z      
         w: DOT4_e      ____,  R1.w,  R4.w      
      8  x: DOT4_e      ____,  R1.x,  R5.x      
         y: DOT4_e      ____,  R1.y,  R5.y      
         z: DOT4_e      R3.z,  R1.z,  R5.z      
         w: DOT4_e      ____,  R1.w,  R5.w      
      9  x: DOT4_e      ____,  R1.x,  R2.x      
         y: DOT4_e      ____,  R1.y,  R2.y      
         z: DOT4_e      ____,  R1.z,  R2.z      
         w: DOT4_e      R3.w,  R1.w,  R2.w      
05 EXP_DONE: PARAM0, R3
06 ALU: ADDR(56) CNT(1) 
     10  x: NOP         ____      
07 NOP NO_BARRIER 
08 END 
END_OF_PROGRAM