26 #define GMB_MULT_AP1( MP1, MA1, MP2, MA2, MX1 ) \  argument
31 TWO(PMULLW     ( MP2, MA2 ))			/*                  t2 = p2*a2                   */	;\
50 #define GMB_MULT_GS( MP1, MA1, MP2, MA2 ) \  argument
52 TWO(PMULLW     ( MP2, MA2 ))			/*                  t2 = p2*a2                   */	;\
57 TWO(MOVQ       ( MA2, MP2 ))										;\
63 TWO(PADDW      ( MP2, MA2 ))			/*        t2 + (t2 >> 8) ~= (t2/255) << 8        */	;\
78 #define GMB_MULT_GSR( MP1, MA1, MP2, MA2, M80 ) \  argument
82 TWO(PMULLW     ( MP2, MA2 ))			/*                  t2 = p2*a2                   */	;\
88 TWO(MOVQ       ( MA2, MP2 ))										;\
94 TWO(PADDW      ( MP2, MA2 ))			/*        t2 + (t2 >> 8) ~= (t2/255) << 8        */	;\
100 #define GMB_LERP_GS( MP1, MQ1, MA1, MP2, MQ2, MA2) \  argument
105 TWO(PSUBW      ( MQ2, MP2 ))                    /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */…
107 TWO(PMULLW     ( MP2, MA2 ))			/*              t2 = (q2 - p2)*pa2               */	;\
112 TWO(MOVQ       ( MA2, MP2 ))										;\
116 TWO(PADDW      ( MP2, MA2 ))			/*        t2 + (t2 >> 8) ~= (t2/255) << 8        */	;\
131 #define GMB_LERP_GSR( MP1, MQ1, MA1, MP2, MQ2, MA2, M80) \  argument
136 TWO(PSUBW      ( MQ2, MP2 ))                    /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */…
138 TWO(PMULLW     ( MP2, MA2 ))			/*              t2 = (q2 - p2)*pa2               */	;\
141 TWO(PSRLW      ( CONST(15), MP2 ))		/*                 q2 > q2 ? 1 : 0               */	;\
144 TWO(PSLLW      ( CONST(8), MP2 ))		/*             q2 > q2 ? 0x100 : 0               */	;\
147 TWO(PSUBW      ( MP2, MA2 ))			/*                  t2 -=? 0x100                 */	;\
155 TWO(MOVQ       ( MA2, MP2 ))										;\
159 TWO(PADDW      ( MP2, MA2 ))			/*        t2 + (t2 >> 8) ~= (t2/255) << 8        */	;\
176 #define GMB_LERP_GSC( MP1, MQ1, MA1, MP2, MQ2, MA2) \  argument
181 TWO(PSUBW      ( MQ2, MP2 ))                    /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */…
183 TWO(PMULLW     ( MP2, MA2 ))			/*              t2 = (q2 - p2)*pa2               */	;\
188 TWO(MOVQ       ( MA2, MP2 ))										;\
194 TWO(PADDW      ( MA2, MP2 ))			/*        t2 + (t2 >> 8) ~= (t2/255) << 8        */	;\
198 TWO(PADDW      ( MP2, MA2 ))			/*  t2 + (t2 >> 8) + (t2 >>15) ~= (t2/255) << 8  */	;\
220 #define GMB_UNPACK(MP1, MQ1, MP2, MQ2, M00) \  argument
221 TWO(MOVQ       ( MP1, MP2 ))										;\
227 TWO(PUNPCKHBW  ( M00, MP2 ))                    /*    pa2    |    pb2    |    pg2    |    pr2    */
229 #define GMB_ALPHA(MP1, MA1, MP2, MA2) \  argument
231 TWO(MOVQ       ( MP2, MA2 ))										;\