25 #define GMB_MULT_AP1( MP1, MA1, MP2, MA2, MX1 ) \  argument
27     PMULLW     ( MP1, MA1 )			/*                  t1 = p1*a1                   */	;\
49 #define GMB_MULT_GS( MP1, MA1, MP2, MA2 ) \  argument
50     PMULLW     ( MP1, MA1 )			/*                  t1 = p1*a1                   */	;\
53     MOVQ       ( MA1, MP1 )										;\
59     PADDW      ( MP1, MA1 )			/*        t1 + (t1 >> 8) ~= (t1/255) << 8        */	;\
77 #define GMB_MULT_GSR( MP1, MA1, MP2, MA2, M80 ) \  argument
78     PMULLW     ( MP1, MA1 )			/*                  t1 = p1*a1                   */	;\
84     MOVQ       ( MA1, MP1 )										;\
90     PADDW      ( MP1, MA1 )			/*        t1 + (t1 >> 8) ~= (t1/255) << 8        */	;\
99 #define GMB_LERP_GS( MP1, MQ1, MA1, MP2, MQ2, MA2) \  argument
100 … PSUBW      ( MQ1, MP1 )                     /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */	;\
102     PMULLW     ( MP1, MA1 )			/*              t1 = (q1 - p1)*pa1               */	;\
108     MOVQ       ( MA1, MP1 )										;\
114     PADDW      ( MP1, MA1 )			/*        t1 + (t1 >> 8) ~= (t1/255) << 8        */	;\
130 #define GMB_LERP_GSR( MP1, MQ1, MA1, MP2, MQ2, MA2, M80) \  argument
131 … PSUBW      ( MQ1, MP1 )                     /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */	;\
133     PMULLW     ( MP1, MA1 )			/*              t1 = (q1 - p1)*pa1               */	;\
139     PSRLW      ( CONST(15), MP1 )		/*                 q1 > p1 ? 1 : 0               */	;\
142     PSLLW      ( CONST(8), MP1 )		/*             q1 > p1 ? 0x100 : 0               */	;\
145     PSUBW      ( MP1, MA1 )			/*                  t1 -=? 0x100                 */	;\
151     MOVQ       ( MA1, MP1 )										;\
157     PADDW      ( MP1, MA1 )			/*        t1 + (t1 >> 8) ~= (t1/255) << 8        */	;\
175 #define GMB_LERP_GSC( MP1, MQ1, MA1, MP2, MQ2, MA2) \  argument
176 … PSUBW      ( MQ1, MP1 )                     /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */	;\
178     PMULLW     ( MP1, MA1 )			/*              t1 = (q1 - p1)*pa1               */	;\
184     MOVQ       ( MA1, MP1 )										;\
190     PADDW      ( MA1, MP1 )			/*        t1 + (t1 >> 8) ~= (t1/255) << 8        */	;\
196     PADDW      ( MP1, MA1 )			/*  t1 + (t1 >> 8) + (t1 >>15) ~= (t1/255) << 8  */	;\
219 #define GMB_UNPACK(MP1, MQ1, MP2, MQ2, M00) \  argument
220 TWO(MOVQ       ( MP1, MP2 ))										;\
225     PUNPCKLBW  ( M00, MP1 )			/*    pa1    |    pb1    |    pg1    |    pr1    */	;\
228 #define GMB_ALPHA(MP1, MA1, MP2, MA2) \  argument
229     MOVQ       ( MP1, MA1 )										;\