Attachment 121138 Details for Bug 93292 – Shader with peephole optimize_CNDcc_op run

Shader with peephole optimize_CNDcc_op run

shader_with_peephole_optimize_CNDcc_op.log (text/plain), 100.94 KB, created by Jan-Marek Glogowski on 2016-01-19 17:41:14 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Jan-Marek Glogowski

Created: 2016-01-19 17:41:14 UTC

Size: 100.94 KB

patch

obsolete

>696 [1mglClear[0m(mask[0m = [34mGL_DEPTH_BUFFER_BIT[0m | [34mGL_STENCIL_BUFFER_BIT[0m | [34mGL_COLOR_BUFFER_BIT[0m)
>696: warning: no current context
>
>sb: shader 94
>
>===== SHADER #94 =========================================== VS/RV710/R700 =====
>===== 288 dw ===== 12 gprs ===== 1 stack =======================================
>0000  00000000 89800000 CALL_FS @0
>0002  40000009 a4040000 ALU_PUSH_BEFORE 2 @18 KC0[CB0:0-15]
> 0018  811020f8 00800790     1      x: SETNE_DX10         R4.x,  0, KC0[1].z
> 0020  801f00fe 0160229c     2 MP   x: PRED_SETNE_INT     R11.x,  PV.x, 0
>0004  00000004 85000000 JUMP @8
>0006  4000000b a1800000 ALU 97 @22 KC0[CB0:0-15]
> 0022  00002082 00800090     3      x: MUL                R4.x,  KC0[2].x, R1.x
> 0024  00002482 20800090            y: MUL                R4.y,  KC0[2].y, R1.x
> 0026  00002882 40800090            z: MUL                R4.z,  KC0[2].z, R1.x
> 0028  80002c82 60800090            w: MUL                R4.w,  KC0[2].w, R1.x
> 0030  00802083 008200fe     4      x: MULADD             R4.x,  KC0[3].x, R1.y, PV.x
> 0032  00802483 208204fe            y: MULADD             R4.y,  KC0[3].y, R1.y, PV.y
> 0034  00802883 408208fe            z: MULADD             R4.z,  KC0[3].z, R1.y, PV.z
> 0036  80802c83 60820cfe            w: MULADD             R4.w,  KC0[3].w, R1.y, PV.w
> 0038  01002084 008200fe     5      x: MULADD             R4.x,  KC0[4].x, R1.z, PV.x
> 0040  01002484 208204fe            y: MULADD             R4.y,  KC0[4].y, R1.z, PV.y
> 0042  01002884 408208fe            z: MULADD             R4.z,  KC0[4].z, R1.z, PV.z
> 0044  81002c84 60820cfe            w: MULADD             R4.w,  KC0[4].w, R1.z, PV.w
> 0046  0010a0fe 00800010     6      x: ADD                R4.x,  PV.x, KC0[5].x
> 0048  0090a4fe 20800010            y: ADD                R4.y,  PV.y, KC0[5].y
> 0050  8110a8fe 40800010            z: ADD                R4.z,  PV.z, KC0[5].z
> 0052  00000086 00a00c90     7      x: MOV                R5.x,  KC0[6].x
> 0054  80000087 20a00c90            y: MOV                R5.y,  KC0[7].x
> 0056  00000486 00c00c90     8      x: MOV                R6.x,  KC0[6].y
> 0058  80000088 40a00c90            z: MOV                R5.z,  KC0[8].x
> 0060  00000487 20c00c90     9      y: MOV                R6.y,  KC0[7].y
> 0062  80000488 40c00c90            z: MOV                R6.z,  KC0[8].y
> 0064  00000886 00e00c90    10      x: MOV                R7.x,  KC0[6].z
> 0066  80000887 20e00c90            y: MOV                R7.y,  KC0[7].z
> 0068  00004005 00a00090    11      x: MUL                R5.x,  R5.x, R2.x
> 0070  00004405 20a00090            y: MUL                R5.y,  R5.y, R2.x
> 0072  00000888 40e00c90            z: MOV                R7.z,  KC0[8].z
> 0074  80004805 40a00090            t: MUL                R5.z,  R5.z, R2.x
> 0076  00804006 00a200fe    12      x: MULADD             R5.x,  R6.x, R2.y, PV.x
> 0078  00804406 20a204fe            y: MULADD             R5.y,  R6.y, R2.y, PV.y
> 0080  80804806 40a200ff            z: MULADD             R5.z,  R6.z, R2.y, PS
> 0082  01004007 00a200fe    13      x: MULADD             R5.x,  R7.x, R2.z, PV.x
> 0084  01004407 20a204fe            y: MULADD             R5.y,  R7.y, R2.z, PV.y
> 0086  81004807 40a208fe            z: MULADD             R5.z,  R7.z, R2.z, PV.z
> 0088  001fc0fe 00c02810    14      x: DOT4               R6.x,  PV.x, PV.x
> 0090  009fc4fe 20c02800            y: DOT4               __.y,  PV.y, PV.y
> 0092  011fc8fe 40c02800            z: DOT4               __.z,  PV.z, PV.z
> 0094  801f00f8 60c02800            w: DOT4               __.w,  0, 0
> 0096  800000fe 01603391    15      t: RECIPSQRT_CLAMPED  R11.x,  |PV.x|
> 0098  000000ff 00c00c90    16      x: MOV                R6.x,  PS
> 0100  000000ff 20c00c80            y: MOV                __.y,  PS
> 0102  000000ff 40c00c80            z: MOV                __.z,  PS
> 0104  800000ff 60c00c80            w: MOV                __.w,  PS
> 0106  001fc005 00a00090    17      x: MUL                R5.x,  R5.x, PV.x
> 0108  001fc405 20a00090            y: MUL                R5.y,  R5.y, PV.x
> 0110  801fc805 40a00090            z: MUL                R5.z,  R5.z, PV.x
> 0112  01900004 00c00090    18      x: MUL                R6.x,  R4.x, KC0[0].w
> 0114  01900404 20c00090            y: MUL                R6.y,  R4.y, KC0[0].w
> 0116  81900804 40c00090            z: MUL                R6.z,  R4.z, KC0[0].w
> 0118  021fc080 00c00010    19      x: ADD                R6.x,  KC0[0].x, -PV.x
> 0120  029fc480 20c00010            y: ADD                R6.y,  KC0[0].y, -PV.y
> 0122  831fc880 40c00010            z: ADD                R6.z,  KC0[0].z, -PV.z
> 0124  001fc0fe 00e02810    20      x: DOT4               R7.x,  PV.x, PV.x
> 0126  009fc4fe 20e02800            y: DOT4               __.y,  PV.y, PV.y
> 0128  011fc8fe 40e02800            z: DOT4               __.z,  PV.z, PV.z
> 0130  801f00f8 60e02800            w: DOT4               __.w,  0, 0
> 0132  800000fe 01603391    21      t: RECIPSQRT_CLAMPED  R11.x,  |PV.x|
> 0134  000000ff 00e00c90    22      x: MOV                R7.x,  PS
> 0136  000000ff 20e00c80            y: MOV                __.y,  PS
> 0138  000000ff 40e00c80            z: MOV                __.z,  PS
> 0140  800000ff 60e00c80            w: MOV                __.w,  PS
> 0142  001fc006 00c00090    23      x: MUL                R6.x,  R6.x, PV.x
> 0144  001fc406 20c00090            y: MUL                R6.y,  R6.y, PV.x
> 0146  801fc806 40c00090            z: MUL                R6.z,  R6.z, PV.x
> 0148  001fc005 00c02810    24      x: DOT4               R6.x,  R5.x, PV.x
> 0150  009fc405 20c02800            y: DOT4               __.y,  R5.y, PV.y
> 0152  011fc805 40c02800            z: DOT4               __.z,  R5.z, PV.z
> 0154  801f00f8 60c02800            w: DOT4               __.w,  0, 0
> 0156  801fc0fe 00c00090    25      x: MUL                R6.x,  PV.x, PV.x
> 0158  821fc4f9 00c00010    26      x: ADD                R6.x,  1.0, -PV.x
> 0160  800000fe 01603510    27      t: SQRT_IEEE          R11.x,  PV.x
> 0162  000000ff 00c00c90    28      x: MOV                R6.x,  PS
> 0164  000000ff 20c00c80            y: MOV                __.y,  PS
> 0166  000000ff 40c00c80            z: MOV                __.z,  PS
> 0168  800000ff 60c00c80            w: MOV                __.w,  PS
> 0170  801fc881 00c00090    29      x: MUL                R6.x,  KC0[1].z, PV.x
> 0172  001fc005 00a00090    30      x: MUL                R5.x,  R5.x, PV.x
> 0174  001fc405 20a00090            y: MUL                R5.y,  R5.y, PV.x
> 0176  801fc805 40a00090            z: MUL                R5.z,  R5.z, PV.x
> 0178  021fc004 00800010    31      x: ADD                R4.x,  R4.x, -PV.x
> 0180  029fc404 20800010            y: ADD                R4.y,  R4.y, -PV.y
> 0182  831fc804 40800010            z: ADD                R4.z,  R4.z, -PV.z
> 0184  001fc08a 00a00090    32      x: MUL                R5.x,  KC0[10].x, PV.x
> 0186  001fc48a 20a00090            y: MUL                R5.y,  KC0[10].y, PV.x
> 0188  001fc88a 40a00090            z: MUL                R5.z,  KC0[10].z, PV.x
> 0190  801fcc8a 60a00090            w: MUL                R5.w,  KC0[10].w, PV.x
> 0192  0080808b 00a200fe    33      x: MULADD             R5.x,  KC0[11].x, R4.y, PV.x
> 0194  0080848b 20a204fe            y: MULADD             R5.y,  KC0[11].y, R4.y, PV.y
> 0196  0080888b 40a208fe            z: MULADD             R5.z,  KC0[11].z, R4.y, PV.z
> 0198  80808c8b 60a20cfe            w: MULADD             R5.w,  KC0[11].w, R4.y, PV.w
> 0200  0100808c 008200fe    34      x: MULADD             R4.x,  KC0[12].x, R4.z, PV.x
> 0202  0100848c 208204fe            y: MULADD             R4.y,  KC0[12].y, R4.z, PV.y
> 0204  0100888c 408208fe            z: MULADD             R4.z,  KC0[12].z, R4.z, PV.z
> 0206  81008c8c 60820cfe            w: MULADD             R4.w,  KC0[12].w, R4.z, PV.w
> 0208  0011a0fe 00800010    35      x: ADD                R4.x,  PV.x, KC0[13].x
> 0210  0091a4fe 20800010            y: ADD                R4.y,  PV.y, KC0[13].y
> 0212  0111a8fe 40800010            z: ADD                R4.z,  PV.z, KC0[13].z
> 0214  8191acfe 60800010            w: ADD                R4.w,  PV.w, KC0[13].w
>0008  00000006 86800001 ELSE @12 POP:1
>0010  8000006c a83c0000 ALU_POP_AFTER 16 @216 KC0[CB0:0-31]
> 0216  0000208e 00a00090    36      x: MUL                R5.x,  KC0[14].x, R1.x
> 0218  0000248e 20a00090            y: MUL                R5.y,  KC0[14].y, R1.x
> 0220  0000288e 40a00090            z: MUL                R5.z,  KC0[14].z, R1.x
> 0222  80002c8e 60a00090            w: MUL                R5.w,  KC0[14].w, R1.x
> 0224  0080208f 00a200fe    37      x: MULADD             R5.x,  KC0[15].x, R1.y, PV.x
> 0226  0080248f 20a204fe            y: MULADD             R5.y,  KC0[15].y, R1.y, PV.y
> 0228  0080288f 40a208fe            z: MULADD             R5.z,  KC0[15].z, R1.y, PV.z
> 0230  80802c8f 60a20cfe            w: MULADD             R5.w,  KC0[15].w, R1.y, PV.w
> 0232  01002090 00a200fe    38      x: MULADD             R5.x,  KC0[16].x, R1.z, PV.x
> 0234  01002490 20a204fe            y: MULADD             R5.y,  KC0[16].y, R1.z, PV.y
> 0236  01002890 40a208fe            z: MULADD             R5.z,  KC0[16].z, R1.z, PV.z
> 0238  81002c90 60a20cfe            w: MULADD             R5.w,  KC0[16].w, R1.z, PV.w
> 0240  001220fe 00800010    39      x: ADD                R4.x,  PV.x, KC0[17].x
> 0242  009224fe 20800010            y: ADD                R4.y,  PV.y, KC0[17].y
> 0244  011228fe 40800010            z: ADD                R4.z,  PV.z, KC0[17].z
> 0246  81922cfe 60800010            w: ADD                R4.w,  PV.w, KC0[17].w
>0012  4000007c a04c0000 ALU 20 @248 KC0[CB0:0-15]
> 0248  00000004 00a00c90    40      x: MOV                R5.x,  R4.x
> 0250  00000404 20a00c90            y: MOV                R5.y,  R4.y
> 0252  00000c04 60a00c90            w: MOV                R5.w,  R4.w
> 0254  80000c04 01603210            t: RECIP_CLAMPED      R11.x,  R4.w
> 0256  000000ff 00c00c90    41      x: MOV                R6.x,  PS
> 0258  000000ff 20c00c80            y: MOV                __.y,  PS
> 0260  000000ff 40c00c80            z: MOV                __.z,  PS
> 0262  800000ff 60c00c80            w: MOV                __.w,  PS
> 0264  801fc081 00c00090    42      x: MUL                R6.x,  KC0[1].x, PV.x
> 0266  800000fe 80c00c90    43      x: MOV_sat            R6.x,  PV.x
> 0268  801fc804 00c00010    44      x: ADD                R6.x,  R4.z, PV.x
> 0270  038080fe 00800190    45      x: MAX                R4.x,  PV.x, -R4.w
> 0272  829020f9 01600010            t: ADD                R11.x,  1.0, -KC0[1].y
> 0274  8000c0ff 01600090    46      x: MUL                R11.x,  PS, R6.x
> 0276  80008481 008200fe    47      x: MULADD             R4.x,  KC0[1].y, R4.x, PV.x
> 0278  800000fe 40a00c90    48      z: MOV                R5.z,  PV.x
> 0280  00000005 00600c90    49      x: MOV                R3.x,  R5.x
> 0282  00000405 20600c90            y: MOV                R3.y,  R5.y
> 0284  000008fe 40600c90            z: MOV                R3.z,  PV.z
> 0286  80000c05 60600c90            w: MOV                R3.w,  R5.w
>0014  c001a03c 94000688 EXPORT_DONE        POS   60    R3.xyzw
>0016  c0004000 94200fff EXPORT_DONE        PARAM 0     R0.____  EOP
>===== SHADER_END ===============================================================
>
>
>
>###### after parse
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.xF@R1.x, R1.yF@R1.y, R1.zF@R1.z, R1.wF@R1.w, R2.xF@R2.x, R2.yF@R2.y, R2.zF@R2.z, R2.wF@R2.w,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    ALU_PUSH_BEFORE     
>    <  
>        [  
>            SETNE_DX10     R4.x,    0|00000000, C1.z
>        ]  
>        [  
>            PRED_SETNE_INT     R11.x, PR, EM,    R4.x, 0|00000000
>        ]  
>    >  
>    region #0   
>        depart region #0 after {    
>            if EM       
>            {
>                depart region #0 after {    
>                    JUMP      @8
>                    ALU     
>                    <  
>                        [  
>                            MUL     R4.x,    C2.x, R1.xF@R1.x
>                            MUL     R4.y,    C2.y, R1.xF@R1.x
>                            MUL     R4.z,    C2.z, R1.xF@R1.x
>                            MUL     R4.w,    C2.w, R1.xF@R1.x
>                        ]  
>                        [  
>                            MULADD     R4.x,    C3.x, R1.yF@R1.y, R4.x
>                            MULADD     R4.y,    C3.y, R1.yF@R1.y, R4.y
>                            MULADD     R4.z,    C3.z, R1.yF@R1.y, R4.z
>                            MULADD     R4.w,    C3.w, R1.yF@R1.y, R4.w
>                        ]  
>                        [  
>                            MULADD     R4.x,    C4.x, R1.zF@R1.z, R4.x
>                            MULADD     R4.y,    C4.y, R1.zF@R1.z, R4.y
>                            MULADD     R4.z,    C4.z, R1.zF@R1.z, R4.z
>                            MULADD     R4.w,    C4.w, R1.zF@R1.z, R4.w
>                        ]  
>                        [  
>                            ADD     R4.x,    R4.x, C5.x
>                            ADD     R4.y,    R4.y, C5.y
>                            ADD     R4.z,    R4.z, C5.z
>                        ]  
>                        [  
>                            MOV     R5.x,    C6.x
>                            MOV     R5.y,    C7.x
>                        ]  
>                        [  
>                            MOV     R6.x,    C6.y
>                            MOV     R5.z,    C8.x
>                        ]  
>                        [  
>                            MOV     R6.y,    C7.y
>                            MOV     R6.z,    C8.y
>                        ]  
>                        [  
>                            MOV     R7.x,    C6.z
>                            MOV     R7.y,    C7.z
>                        ]  
>                        [  
>                            MUL     R5.x,    R5.x, R2.xF@R2.x
>                            MUL     R5.y,    R5.y, R2.xF@R2.x
>                            MOV     R7.z,    C8.z
>                            MUL     R5.z,    R5.z, R2.xF@R2.x
>                        ]  
>                        [  
>                            MULADD     R5.x,    R6.x, R2.yF@R2.y, R5.x
>                            MULADD     R5.y,    R6.y, R2.yF@R2.y, R5.y
>                            MULADD     R5.z,    R6.z, R2.yF@R2.y, R5.z
>                        ]  
>                        [  
>                            MULADD     R5.x,    R7.x, R2.zF@R2.z, R5.x
>                            MULADD     R5.y,    R7.y, R2.zF@R2.z, R5.y
>                            MULADD     R5.z,    R7.z, R2.zF@R2.z, R5.z
>                        ]  
>                        [  
>                            DOT4       
>                                4S  DOT4     R6.x,    R5.x, R5.x
>                                4S  DOT4     __,    R5.y, R5.y
>                                4S  DOT4     __,    R5.z, R5.z
>                                4S  DOT4     __,    0|00000000, 0|00000000
>                        ]  
>                        [  
>                            RECIPSQRT_CLAMPED     R11.x,    |R6.x|
>                        ]  
>                        [  
>                            MOV     R6.x,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                        ]  
>                        [  
>                            MUL     R5.x,    R5.x, R6.x
>                            MUL     R5.y,    R5.y, R6.x
>                            MUL     R5.z,    R5.z, R6.x
>                        ]  
>                        [  
>                            MUL     R6.x,    R4.x, C0.w
>                            MUL     R6.y,    R4.y, C0.w
>                            MUL     R6.z,    R4.z, C0.w
>                        ]  
>                        [  
>                            ADD     R6.x,    C0.x, -R6.x
>                            ADD     R6.y,    C0.y, -R6.y
>                            ADD     R6.z,    C0.z, -R6.z
>                        ]  
>                        [  
>                            DOT4       
>                                4S  DOT4     R7.x,    R6.x, R6.x
>                                4S  DOT4     __,    R6.y, R6.y
>                                4S  DOT4     __,    R6.z, R6.z
>                                4S  DOT4     __,    0|00000000, 0|00000000
>                        ]  
>                        [  
>                            RECIPSQRT_CLAMPED     R11.x,    |R7.x|
>                        ]  
>                        [  
>                            MOV     R7.x,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                        ]  
>                        [  
>                            MUL     R6.x,    R6.x, R7.x
>                            MUL     R6.y,    R6.y, R7.x
>                            MUL     R6.z,    R6.z, R7.x
>                        ]  
>                        [  
>                            DOT4       
>                                4S  DOT4     R6.x,    R5.x, R6.x
>                                4S  DOT4     __,    R5.y, R6.y
>                                4S  DOT4     __,    R5.z, R6.z
>                                4S  DOT4     __,    0|00000000, 0|00000000
>                        ]  
>                        [  
>                            MUL     R6.x,    R6.x, R6.x
>                        ]  
>                        [  
>                            ADD     R6.x,    1|3f800000, -R6.x
>                        ]  
>                        [  
>                            SQRT_IEEE     R11.x,    R6.x
>                        ]  
>                        [  
>                            MOV     R6.x,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                            MOV     __,    R11.x
>                        ]  
>                        [  
>                            MUL     R6.x,    C1.z, R6.x
>                        ]  
>                        [  
>                            MUL     R5.x,    R5.x, R6.x
>                            MUL     R5.y,    R5.y, R6.x
>                            MUL     R5.z,    R5.z, R6.x
>                        ]  
>                        [  
>                            ADD     R4.x,    R4.x, -R5.x
>                            ADD     R4.y,    R4.y, -R5.y
>                            ADD     R4.z,    R4.z, -R5.z
>                        ]  
>                        [  
>                            MUL     R5.x,    C10.x, R4.x
>                            MUL     R5.y,    C10.y, R4.x
>                            MUL     R5.z,    C10.z, R4.x
>                            MUL     R5.w,    C10.w, R4.x
>                        ]  
>                        [  
>                            MULADD     R5.x,    C11.x, R4.y, R5.x
>                            MULADD     R5.y,    C11.y, R4.y, R5.y
>                            MULADD     R5.z,    C11.z, R4.y, R5.z
>                            MULADD     R5.w,    C11.w, R4.y, R5.w
>                        ]  
>                        [  
>                            MULADD     R4.x,    C12.x, R4.z, R5.x
>                            MULADD     R4.y,    C12.y, R4.z, R5.y
>                            MULADD     R4.z,    C12.z, R4.z, R5.z
>                            MULADD     R4.w,    C12.w, R4.z, R5.w
>                        ]  
>                        [  
>                            ADD     R4.x,    R4.x, C13.x
>                            ADD     R4.y,    R4.y, C13.y
>                            ADD     R4.z,    R4.z, C13.z
>                            ADD     R4.w,    R4.w, C13.w
>                        ]  
>                    >  
>                } end_depart   
>            } endif   
>            ELSE      @12
>            ALU_POP_AFTER     
>            <  
>                [  
>                    MUL     R5.x,    C14.x, R1.xF@R1.x
>                    MUL     R5.y,    C14.y, R1.xF@R1.x
>                    MUL     R5.z,    C14.z, R1.xF@R1.x
>                    MUL     R5.w,    C14.w, R1.xF@R1.x
>                ]  
>                [  
>                    MULADD     R5.x,    C15.x, R1.yF@R1.y, R5.x
>                    MULADD     R5.y,    C15.y, R1.yF@R1.y, R5.y
>                    MULADD     R5.z,    C15.z, R1.yF@R1.y, R5.z
>                    MULADD     R5.w,    C15.w, R1.yF@R1.y, R5.w
>                ]  
>                [  
>                    MULADD     R5.x,    C16.x, R1.zF@R1.z, R5.x
>                    MULADD     R5.y,    C16.y, R1.zF@R1.z, R5.y
>                    MULADD     R5.z,    C16.z, R1.zF@R1.z, R5.z
>                    MULADD     R5.w,    C16.w, R1.zF@R1.z, R5.w
>                ]  
>                [  
>                    ADD     R4.x,    R5.x, C17.x
>                    ADD     R4.y,    R5.y, C17.y
>                    ADD     R4.z,    R5.z, C17.z
>                    ADD     R4.w,    R5.w, C17.w
>                ]  
>            >  
>        } end_depart   
>    
>    ALU     
>    <  
>        [  
>            MOV     R5.x,    R4.x
>            MOV     R5.y,    R4.y
>            MOV     R5.w,    R4.w
>            RECIP_CLAMPED     R11.x,    R4.w
>        ]  
>        [  
>            MOV     R6.x,    R11.x
>            MOV     __,    R11.x
>            MOV     __,    R11.x
>            MOV     __,    R11.x
>        ]  
>        [  
>            MUL     R6.x,    C1.x, R6.x
>        ]  
>        [  
>            MOV_sat     R6.x,    R6.x
>        ]  
>        [  
>            ADD     R6.x,    R4.z, R6.x
>        ]  
>        [  
>            MAX     R4.x,    R6.x, -R4.w
>            ADD     R11.x,    1|3f800000, -C1.y
>        ]  
>        [  
>            MUL     R11.x,    R11.x, R6.x
>        ]  
>        [  
>            MULADD     R4.x,    C1.y, R4.x, R11.x
>        ]  
>        [  
>            MOV     R5.z,    R4.x
>        ]  
>        [  
>            MOV     R3.x,    R5.x
>            MOV     R3.y,    R5.y
>            MOV     R3.z,    R5.z
>            MOV     R3.w,    R5.w
>        ]  
>    >  
>    EXPORT  POS   60     R3.x, R3.y, R3.z, R3.w
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after ssa_rename
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, R1.w.1, R2.x.1, R2.y.1, R2.z.1, R2.w.1,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    ALU_PUSH_BEFORE     
>    <  
>        [  
>            SETNE_DX10     R4.x.1,    0|00000000, C1.z
>        ]  
>        [  
>            PRED_SETNE_INT     R11.x.1, PR.1, EM.1,    R4.x.1, 0|00000000
>        ]  
>    >  
>    region #0   
>        depart region #0 after {    
>            if EM.1       
>            {
>                depart region #0 after {    
>                    JUMP      @8
>                    ALU     
>                    <  
>                        [  
>                            MUL     R4.x.2,    C2.x, R1.x.1
>                            MUL     R4.y.1,    C2.y, R1.x.1
>                            MUL     R4.z.1,    C2.z, R1.x.1
>                            MUL     R4.w.1,    C2.w, R1.x.1
>                        ]  
>                        [  
>                            MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>                            MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>                            MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>                            MULADD     R4.w.2,    C3.w, R1.y.1, R4.w.1
>                        ]  
>                        [  
>                            MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>                            MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>                            MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>                            MULADD     R4.w.3,    C4.w, R1.z.1, R4.w.2
>                        ]  
>                        [  
>                            ADD     R4.x.5,    R4.x.4, C5.x
>                            ADD     R4.y.4,    R4.y.3, C5.y
>                            ADD     R4.z.4,    R4.z.3, C5.z
>                        ]  
>                        [  
>                            MOV     R5.x.1,    C6.x
>                            MOV     R5.y.1,    C7.x
>                        ]  
>                        [  
>                            MOV     R6.x.1,    C6.y
>                            MOV     R5.z.1,    C8.x
>                        ]  
>                        [  
>                            MOV     R6.y.1,    C7.y
>                            MOV     R6.z.1,    C8.y
>                        ]  
>                        [  
>                            MOV     R7.x.1,    C6.z
>                            MOV     R7.y.1,    C7.z
>                        ]  
>                        [  
>                            MUL     R5.x.2,    R5.x.1, R2.x.1
>                            MUL     R5.y.2,    R5.y.1, R2.x.1
>                            MOV     R7.z.1,    C8.z
>                            MUL     R5.z.2,    R5.z.1, R2.x.1
>                        ]  
>                        [  
>                            MULADD     R5.x.3,    R6.x.1, R2.y.1, R5.x.2
>                            MULADD     R5.y.3,    R6.y.1, R2.y.1, R5.y.2
>                            MULADD     R5.z.3,    R6.z.1, R2.y.1, R5.z.2
>                        ]  
>                        [  
>                            MULADD     R5.x.4,    R7.x.1, R2.z.1, R5.x.3
>                            MULADD     R5.y.4,    R7.y.1, R2.z.1, R5.y.3
>                            MULADD     R5.z.4,    R7.z.1, R2.z.1, R5.z.3
>                        ]  
>                        [  
>                            DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>                        ]  
>                        [  
>                            RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>                        ]  
>                        [  
>                            MOV     R6.x.3,    R11.x.2
>                            MOV     __,    R11.x.2
>                            MOV     __,    R11.x.2
>                            MOV     __,    R11.x.2
>                        ]  
>                        [  
>                            MUL     R5.x.5,    R5.x.4, R6.x.3
>                            MUL     R5.y.5,    R5.y.4, R6.x.3
>                            MUL     R5.z.5,    R5.z.4, R6.x.3
>                        ]  
>                        [  
>                            MUL     R6.x.4,    R4.x.5, C0.w
>                            MUL     R6.y.2,    R4.y.4, C0.w
>                            MUL     R6.z.2,    R4.z.4, C0.w
>                        ]  
>                        [  
>                            ADD     R6.x.5,    C0.x, -R6.x.4
>                            ADD     R6.y.3,    C0.y, -R6.y.2
>                            ADD     R6.z.3,    C0.z, -R6.z.2
>                        ]  
>                        [  
>                            DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>                        ]  
>                        [  
>                            RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>                        ]  
>                        [  
>                            MOV     R7.x.3,    R11.x.3
>                            MOV     __,    R11.x.3
>                            MOV     __,    R11.x.3
>                            MOV     __,    R11.x.3
>                        ]  
>                        [  
>                            MUL     R6.x.6,    R6.x.5, R7.x.3
>                            MUL     R6.y.4,    R6.y.3, R7.x.3
>                            MUL     R6.z.4,    R6.z.3, R7.x.3
>                        ]  
>                        [  
>                            DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>                        ]  
>                        [  
>                            MUL     R6.x.8,    R6.x.7, R6.x.7
>                        ]  
>                        [  
>                            ADD     R6.x.9,    1|3f800000, -R6.x.8
>                        ]  
>                        [  
>                            SQRT_IEEE     R11.x.4,    R6.x.9
>                        ]  
>                        [  
>                            MOV     R6.x.10,    R11.x.4
>                            MOV     __,    R11.x.4
>                            MOV     __,    R11.x.4
>                            MOV     __,    R11.x.4
>                        ]  
>                        [  
>                            MUL     R6.x.11,    C1.z, R6.x.10
>                        ]  
>                        [  
>                            MUL     R5.x.6,    R5.x.5, R6.x.11
>                            MUL     R5.y.6,    R5.y.5, R6.x.11
>                            MUL     R5.z.6,    R5.z.5, R6.x.11
>                        ]  
>                        [  
>                            ADD     R4.x.6,    R4.x.5, -R5.x.6
>                            ADD     R4.y.5,    R4.y.4, -R5.y.6
>                            ADD     R4.z.5,    R4.z.4, -R5.z.6
>                        ]  
>                        [  
>                            MUL     R5.x.7,    C10.x, R4.x.6
>                            MUL     R5.y.7,    C10.y, R4.x.6
>                            MUL     R5.z.7,    C10.z, R4.x.6
>                            MUL     R5.w.1,    C10.w, R4.x.6
>                        ]  
>                        [  
>                            MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>                            MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>                            MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>                            MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>                        ]  
>                        [  
>                            MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>                            MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>                            MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>                            MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>                        ]  
>                        [  
>                            ADD     R4.x.8,    R4.x.7, C13.x
>                            ADD     R4.y.7,    R4.y.6, C13.y
>                            ADD     R4.z.7,    R4.z.6, C13.z
>                            ADD     R4.w.5,    R4.w.4, C13.w
>                        ]  
>                    >  
>                } end_depart   
>            } endif   
>            ELSE      @12
>            ALU_POP_AFTER     
>            <  
>                [  
>                    MUL     R5.x.9,    C14.x, R1.x.1
>                    MUL     R5.y.9,    C14.y, R1.x.1
>                    MUL     R5.z.9,    C14.z, R1.x.1
>                    MUL     R5.w.3,    C14.w, R1.x.1
>                ]  
>                [  
>                    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>                    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>                    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>                    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>                ]  
>                [  
>                    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>                    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>                    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>                    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>                ]  
>                [  
>                    ADD     R4.x.9,    R5.x.11, C17.x
>                    ADD     R4.y.8,    R5.y.11, C17.y
>                    ADD     R4.z.8,    R5.z.11, C17.z
>                    ADD     R4.w.6,    R5.w.5, C17.w
>                ]  
>            >  
>        } end_depart   
>    {  
>        * phi     R4.x.10,       R4.x.8, R4.x.9
>        * phi     R11.x.5,       R11.x.4, R11.x.1
>        * phi     R4.y.9,       R4.y.7, R4.y.8
>        * phi     R4.z.9,       R4.z.7, R4.z.8
>        * phi     R4.w.7,       R4.w.5, R4.w.6
>        * phi     R5.x.12,       R5.x.8, R5.x.11
>        * phi     R5.y.12,       R5.y.8, R5.y.11
>        * phi     R6.x.12,       R6.x.11, R6.x
>        * phi     R5.z.12,       R5.z.8, R5.z.11
>        * phi     R6.y.5,       R6.y.4, R6.y
>        * phi     R6.z.5,       R6.z.4, R6.z
>        * phi     R7.x.4,       R7.x.3, R7.x
>        * phi     R7.y.2,       R7.y.1, R7.y
>        * phi     R7.z.2,       R7.z.1, R7.z
>        * phi     R5.w.6,       R5.w.2, R5.w.5
>    }  
>    
>    ALU     
>    <  
>        [  
>            MOV     R5.x.13,    R4.x.10
>            MOV     R5.y.13,    R4.y.9
>            MOV     R5.w.7,    R4.w.7
>            RECIP_CLAMPED     R11.x.6,    R4.w.7
>        ]  
>        [  
>            MOV     R6.x.13,    R11.x.6
>            MOV     __,    R11.x.6
>            MOV     __,    R11.x.6
>            MOV     __,    R11.x.6
>        ]  
>        [  
>            MUL     R6.x.14,    C1.x, R6.x.13
>        ]  
>        [  
>            MOV_sat     R6.x.15,    R6.x.14
>        ]  
>        [  
>            ADD     R6.x.16,    R4.z.9, R6.x.15
>        ]  
>        [  
>            MAX     R4.x.11,    R6.x.16, -R4.w.7
>            ADD     R11.x.7,    1|3f800000, -C1.y
>        ]  
>        [  
>            MUL     R11.x.8,    R11.x.7, R6.x.16
>        ]  
>        [  
>            MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>        ]  
>        [  
>            MOV     R5.z.13,    R4.x.12
>        ]  
>        [  
>            MOV     R3.x.1,    R5.x.13
>            MOV     R3.y.1,    R5.y.13
>            MOV     R3.z.1,    R5.z.13
>            MOV     R3.w.1,    R5.w.7
>        ]  
>    >  
>    EXPORT  POS   60     R3.x.1, R3.y.1, R3.z.1, R3.w.1
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after if_conversion
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, __, R2.x.1, R2.y.1, R2.z.1, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    SETNE_DX10     R4.x.1,    0|00000000, C1.z
>    PRED_SETNE_INT     __, __, EM.1,    R4.x.1, 0|00000000
>    SETE_INT     t1,    R4.x.1, 0|00000000
>    MUL     R4.x.2,    C2.x, R1.x.1
>    MUL     R4.y.1,    C2.y, R1.x.1
>    MUL     R4.z.1,    C2.z, R1.x.1
>    MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>    MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>    MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>    MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>    MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>    MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>    ADD     R4.x.5,    R4.x.4, C5.x
>    ADD     R4.y.4,    R4.y.3, C5.y
>    ADD     R4.z.4,    R4.z.3, C5.z
>    MOV     R5.x.1,    C6.x
>    MOV     R5.y.1,    C7.x
>    MOV     R6.x.1,    C6.y
>    MOV     R5.z.1,    C8.x
>    MOV     R6.y.1,    C7.y
>    MOV     R6.z.1,    C8.y
>    MOV     R7.x.1,    C6.z
>    MOV     R7.y.1,    C7.z
>    MUL     R5.x.2,    R5.x.1, R2.x.1
>    MUL     R5.y.2,    R5.y.1, R2.x.1
>    MOV     R7.z.1,    C8.z
>    MUL     R5.z.2,    R5.z.1, R2.x.1
>    MULADD     R5.x.3,    R6.x.1, R2.y.1, R5.x.2
>    MULADD     R5.y.3,    R6.y.1, R2.y.1, R5.y.2
>    MULADD     R5.z.3,    R6.z.1, R2.y.1, R5.z.2
>    MULADD     R5.x.4,    R7.x.1, R2.z.1, R5.x.3
>    MULADD     R5.y.4,    R7.y.1, R2.z.1, R5.y.3
>    MULADD     R5.z.4,    R7.z.1, R2.z.1, R5.z.3
>    DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>    MOV     R6.x.3,    R11.x.2
>    MUL     R5.x.5,    R5.x.4, R6.x.3
>    MUL     R5.y.5,    R5.y.4, R6.x.3
>    MUL     R5.z.5,    R5.z.4, R6.x.3
>    MUL     R6.x.4,    R4.x.5, C0.w
>    MUL     R6.y.2,    R4.y.4, C0.w
>    MUL     R6.z.2,    R4.z.4, C0.w
>    ADD     R6.x.5,    C0.x, -R6.x.4
>    ADD     R6.y.3,    C0.y, -R6.y.2
>    ADD     R6.z.3,    C0.z, -R6.z.2
>    DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>    MOV     R7.x.3,    R11.x.3
>    MUL     R6.x.6,    R6.x.5, R7.x.3
>    MUL     R6.y.4,    R6.y.3, R7.x.3
>    MUL     R6.z.4,    R6.z.3, R7.x.3
>    DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>    MUL     R6.x.8,    R6.x.7, R6.x.7
>    ADD     R6.x.9,    1|3f800000, -R6.x.8
>    SQRT_IEEE     R11.x.4,    R6.x.9
>    MOV     R6.x.10,    R11.x.4
>    MUL     R6.x.11,    C1.z, R6.x.10
>    MUL     R5.x.6,    R5.x.5, R6.x.11
>    MUL     R5.y.6,    R5.y.5, R6.x.11
>    MUL     R5.z.6,    R5.z.5, R6.x.11
>    ADD     R4.x.6,    R4.x.5, -R5.x.6
>    ADD     R4.y.5,    R4.y.4, -R5.y.6
>    ADD     R4.z.5,    R4.z.4, -R5.z.6
>    MUL     R5.x.7,    C10.x, R4.x.6
>    MUL     R5.y.7,    C10.y, R4.x.6
>    MUL     R5.z.7,    C10.z, R4.x.6
>    MUL     R5.w.1,    C10.w, R4.x.6
>    MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>    MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>    MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>    MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>    MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>    MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>    MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>    MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>    ADD     R4.x.8,    R4.x.7, C13.x
>    ADD     R4.y.7,    R4.y.6, C13.y
>    ADD     R4.z.7,    R4.z.6, C13.z
>    ADD     R4.w.5,    R4.w.4, C13.w
>    MUL     R5.x.9,    C14.x, R1.x.1
>    MUL     R5.y.9,    C14.y, R1.x.1
>    MUL     R5.z.9,    C14.z, R1.x.1
>    MUL     R5.w.3,    C14.w, R1.x.1
>    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>    ADD     R4.x.9,    R5.x.11, C17.x
>    ADD     R4.y.8,    R5.y.11, C17.y
>    ADD     R4.z.8,    R5.z.11, C17.z
>    ADD     R4.w.6,    R5.w.5, C17.w
>    CNDE_INT     R4.w.7,    t1, R4.w.5, R4.w.6
>    CNDE_INT     R4.z.9,    t1, R4.z.7, R4.z.8
>    CNDE_INT     R4.y.9,    t1, R4.y.7, R4.y.8
>    CNDE_INT     R4.x.10,    t1, R4.x.8, R4.x.9
>    MOV     R5.x.13,    R4.x.10
>    MOV     R5.y.13,    R4.y.9
>    MOV     R5.w.7,    R4.w.7
>    RECIP_CLAMPED     R11.x.6,    R4.w.7
>    MOV     R6.x.13,    R11.x.6
>    MUL     R6.x.14,    C1.x, R6.x.13
>    MOV_sat     R6.x.15,    R6.x.14
>    ADD     R6.x.16,    R4.z.9, R6.x.15
>    MAX     R4.x.11,    R6.x.16, -R4.w.7
>    ADD     R11.x.7,    1|3f800000, -C1.y
>    MUL     R11.x.8,    R11.x.7, R6.x.16
>    MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>    MOV     R5.z.13,    R4.x.12
>    MOV     R3.x.1,    R5.x.13
>    MOV     R3.y.1,    R5.y.13
>    MOV     R3.z.1,    R5.z.13
>    MOV     R3.w.1,    R5.w.7
>    EXPORT  POS   60     R3.x.1, R3.y.1, R3.z.1, R3.w.1
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after peephole
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, __, R2.x.1, R2.y.1, R2.z.1, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    SETNE_DX10     R4.x.1,    C1.z, 0|00000000
>    PRED_SETNE     __, __, EM.1,    C1.z, 0|00000000
>    SETE_DX10     t1,    C1.z, 0|00000000
>    MUL     R4.x.2,    C2.x, R1.x.1
>    MUL     R4.y.1,    C2.y, R1.x.1
>    MUL     R4.z.1,    C2.z, R1.x.1
>    MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>    MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>    MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>    MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>    MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>    MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>    ADD     R4.x.5,    R4.x.4, C5.x
>    ADD     R4.y.4,    R4.y.3, C5.y
>    ADD     R4.z.4,    R4.z.3, C5.z
>    MOV     R5.x.1,    C6.x
>    MOV     R5.y.1,    C7.x
>    MOV     R6.x.1,    C6.y
>    MOV     R5.z.1,    C8.x
>    MOV     R6.y.1,    C7.y
>    MOV     R6.z.1,    C8.y
>    MOV     R7.x.1,    C6.z
>    MOV     R7.y.1,    C7.z
>    MUL     R5.x.2,    R5.x.1, R2.x.1
>    MUL     R5.y.2,    R5.y.1, R2.x.1
>    MOV     R7.z.1,    C8.z
>    MUL     R5.z.2,    R5.z.1, R2.x.1
>    MULADD     R5.x.3,    R6.x.1, R2.y.1, R5.x.2
>    MULADD     R5.y.3,    R6.y.1, R2.y.1, R5.y.2
>    MULADD     R5.z.3,    R6.z.1, R2.y.1, R5.z.2
>    MULADD     R5.x.4,    R7.x.1, R2.z.1, R5.x.3
>    MULADD     R5.y.4,    R7.y.1, R2.z.1, R5.y.3
>    MULADD     R5.z.4,    R7.z.1, R2.z.1, R5.z.3
>    DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>    MOV     R6.x.3,    R11.x.2
>    MUL     R5.x.5,    R5.x.4, R6.x.3
>    MUL     R5.y.5,    R5.y.4, R6.x.3
>    MUL     R5.z.5,    R5.z.4, R6.x.3
>    MUL     R6.x.4,    R4.x.5, C0.w
>    MUL     R6.y.2,    R4.y.4, C0.w
>    MUL     R6.z.2,    R4.z.4, C0.w
>    ADD     R6.x.5,    C0.x, -R6.x.4
>    ADD     R6.y.3,    C0.y, -R6.y.2
>    ADD     R6.z.3,    C0.z, -R6.z.2
>    DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>    MOV     R7.x.3,    R11.x.3
>    MUL     R6.x.6,    R6.x.5, R7.x.3
>    MUL     R6.y.4,    R6.y.3, R7.x.3
>    MUL     R6.z.4,    R6.z.3, R7.x.3
>    DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>    MUL     R6.x.8,    R6.x.7, R6.x.7
>    ADD     R6.x.9,    1|3f800000, -R6.x.8
>    SQRT_IEEE     R11.x.4,    R6.x.9
>    MOV     R6.x.10,    R11.x.4
>    MUL     R6.x.11,    C1.z, R6.x.10
>    MUL     R5.x.6,    R5.x.5, R6.x.11
>    MUL     R5.y.6,    R5.y.5, R6.x.11
>    MUL     R5.z.6,    R5.z.5, R6.x.11
>    ADD     R4.x.6,    R4.x.5, -R5.x.6
>    ADD     R4.y.5,    R4.y.4, -R5.y.6
>    ADD     R4.z.5,    R4.z.4, -R5.z.6
>    MUL     R5.x.7,    C10.x, R4.x.6
>    MUL     R5.y.7,    C10.y, R4.x.6
>    MUL     R5.z.7,    C10.z, R4.x.6
>    MUL     R5.w.1,    C10.w, R4.x.6
>    MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>    MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>    MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>    MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>    MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>    MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>    MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>    MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>    ADD     R4.x.8,    R4.x.7, C13.x
>    ADD     R4.y.7,    R4.y.6, C13.y
>    ADD     R4.z.7,    R4.z.6, C13.z
>    ADD     R4.w.5,    R4.w.4, C13.w
>    MUL     R5.x.9,    C14.x, R1.x.1
>    MUL     R5.y.9,    C14.y, R1.x.1
>    MUL     R5.z.9,    C14.z, R1.x.1
>    MUL     R5.w.3,    C14.w, R1.x.1
>    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>    ADD     R4.x.9,    R5.x.11, C17.x
>    ADD     R4.y.8,    R5.y.11, C17.y
>    ADD     R4.z.8,    R5.z.11, C17.z
>    ADD     R4.w.6,    R5.w.5, C17.w
>    CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>    CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>    CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>    CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>    MOV     R5.x.13,    R4.x.10
>    MOV     R5.y.13,    R4.y.9
>    MOV     R5.w.7,    R4.w.7
>    RECIP_CLAMPED     R11.x.6,    R4.w.7
>    MOV     R6.x.13,    R11.x.6
>    MUL     R6.x.14,    C1.x, R6.x.13
>    MOV_sat     R6.x.15,    R6.x.14
>    ADD     R6.x.16,    R4.z.9, R6.x.15
>    MAX     R4.x.11,    R6.x.16, -R4.w.7
>    ADD     R11.x.7,    1|3f800000, -C1.y
>    MUL     R11.x.8,    R11.x.7, R6.x.16
>    MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>    MOV     R5.z.13,    R4.x.12
>    MOV     R3.x.1,    R5.x.13
>    MOV     R3.y.1,    R5.y.13
>    MOV     R3.z.1,    R5.z.13
>    MOV     R3.w.1,    R5.w.7
>    EXPORT  POS   60     R3.x.1, R3.y.1, R3.z.1, R3.w.1
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after gvn
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, __, R2.x.1, R2.y.1, R2.z.1, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    SETNE_DX10     R4.x.1,    C1.z, 0|00000000
>    PRED_SETNE     __, __, EM.1,    C1.z, 0|00000000
>    SETE_DX10     t1,    C1.z, 0|00000000
>    MUL     R4.x.2,    C2.x, R1.x.1
>    MUL     R4.y.1,    C2.y, R1.x.1
>    MUL     R4.z.1,    C2.z, R1.x.1
>    MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>    MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>    MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>    MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>    MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>    MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>    ADD     R4.x.5,    R4.x.4, C5.x
>    ADD     R4.y.4,    R4.y.3, C5.y
>    ADD     R4.z.4,    R4.z.3, C5.z
>    MOV     R5.x.1,    C6.x
>    MOV     R5.y.1,    C7.x
>    MOV     R6.x.1,    C6.y
>    MOV     R5.z.1,    C8.x
>    MOV     R6.y.1,    C7.y
>    MOV     R6.z.1,    C8.y
>    MOV     R7.x.1,    C6.z
>    MOV     R7.y.1,    C7.z
>    MUL     R5.x.2,    C6.x, R2.x.1
>    MUL     R5.y.2,    C7.x, R2.x.1
>    MOV     R7.z.1,    C8.z
>    MUL     R5.z.2,    C8.x, R2.x.1
>    MULADD     R5.x.3,    C6.y, R2.y.1, R5.x.2
>    MULADD     R5.y.3,    C7.y, R2.y.1, R5.y.2
>    MULADD     R5.z.3,    C8.y, R2.y.1, R5.z.2
>    MULADD     R5.x.4,    C6.z, R2.z.1, R5.x.3
>    MULADD     R5.y.4,    C7.z, R2.z.1, R5.y.3
>    MULADD     R5.z.4,    C8.z, R2.z.1, R5.z.3
>    DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>    MOV     R6.x.3,    R11.x.2
>    MUL     R5.x.5,    R5.x.4, R11.x.2
>    MUL     R5.y.5,    R5.y.4, R11.x.2
>    MUL     R5.z.5,    R5.z.4, R11.x.2
>    MUL     R6.x.4,    R4.x.5, C0.w
>    MUL     R6.y.2,    R4.y.4, C0.w
>    MUL     R6.z.2,    R4.z.4, C0.w
>    MULADD     R6.x.5,    R4.x.5, -C0.w, C0.x
>    MULADD     R6.y.3,    R4.y.4, -C0.w, C0.y
>    MULADD     R6.z.3,    R4.z.4, -C0.w, C0.z
>    DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>    MOV     R7.x.3,    R11.x.3
>    MUL     R6.x.6,    R6.x.5, R11.x.3
>    MUL     R6.y.4,    R6.y.3, R11.x.3
>    MUL     R6.z.4,    R6.z.3, R11.x.3
>    DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>    MUL     R6.x.8,    R6.x.7, R6.x.7
>    MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>    SQRT_IEEE     R11.x.4,    R6.x.9
>    MOV     R6.x.10,    R11.x.4
>    MUL     R6.x.11,    C1.z, R11.x.4
>    MUL     R5.x.6,    R5.x.5, R6.x.11
>    MUL     R5.y.6,    R5.y.5, R6.x.11
>    MUL     R5.z.6,    R5.z.5, R6.x.11
>    ADD     R4.x.6,    R4.x.5, -R5.x.6
>    ADD     R4.y.5,    R4.y.4, -R5.y.6
>    ADD     R4.z.5,    R4.z.4, -R5.z.6
>    MUL     R5.x.7,    C10.x, R4.x.6
>    MUL     R5.y.7,    C10.y, R4.x.6
>    MUL     R5.z.7,    C10.z, R4.x.6
>    MUL     R5.w.1,    C10.w, R4.x.6
>    MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>    MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>    MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>    MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>    MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>    MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>    MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>    MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>    ADD     R4.x.8,    R4.x.7, C13.x
>    ADD     R4.y.7,    R4.y.6, C13.y
>    ADD     R4.z.7,    R4.z.6, C13.z
>    ADD     R4.w.5,    R4.w.4, C13.w
>    MUL     R5.x.9,    C14.x, R1.x.1
>    MUL     R5.y.9,    C14.y, R1.x.1
>    MUL     R5.z.9,    C14.z, R1.x.1
>    MUL     R5.w.3,    C14.w, R1.x.1
>    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>    ADD     R4.x.9,    R5.x.11, C17.x
>    ADD     R4.y.8,    R5.y.11, C17.y
>    ADD     R4.z.8,    R5.z.11, C17.z
>    ADD     R4.w.6,    R5.w.5, C17.w
>    CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>    CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>    CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>    CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>    MOV     R5.x.13,    R4.x.10
>    MOV     R5.y.13,    R4.y.9
>    MOV     R5.w.7,    R4.w.7
>    RECIP_CLAMPED     R11.x.6,    R4.w.7
>    MOV     R6.x.13,    R11.x.6
>    MUL     R6.x.14,    C1.x, R11.x.6
>    MOV_sat     R6.x.15,    R6.x.14
>    ADD     R6.x.16,    R4.z.9, R6.x.15
>    MAX     R4.x.11,    R6.x.16, -R4.w.7
>    ADD     R11.x.7,    -C1.y, 1|3f800000
>    MUL     R11.x.8,    R11.x.7, R6.x.16
>    MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>    MOV     R5.z.13,    R4.x.12
>    MOV     R3.x.1,    R4.x.10
>    MOV     R3.y.1,    R4.y.9
>    MOV     R3.z.1,    R4.x.12
>    MOV     R3.w.1,    R4.w.7
>    EXPORT  POS   60     R4.x.10, R4.y.9, R4.x.12, R4.w.7
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after def_use
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, __, R2.x.1, R2.y.1, R2.z.1, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    SETNE_DX10     R4.x.1,    C1.z, 0|00000000
>    PRED_SETNE     __, __, EM.1,    C1.z, 0|00000000
>    SETE_DX10     t1,    C1.z, 0|00000000
>    MUL     R4.x.2,    C2.x, R1.x.1
>    MUL     R4.y.1,    C2.y, R1.x.1
>    MUL     R4.z.1,    C2.z, R1.x.1
>    MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>    MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>    MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>    MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>    MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>    MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>    ADD     R4.x.5,    R4.x.4, C5.x
>    ADD     R4.y.4,    R4.y.3, C5.y
>    ADD     R4.z.4,    R4.z.3, C5.z
>    MOV     R5.x.1,    C6.x
>    MOV     R5.y.1,    C7.x
>    MOV     R6.x.1,    C6.y
>    MOV     R5.z.1,    C8.x
>    MOV     R6.y.1,    C7.y
>    MOV     R6.z.1,    C8.y
>    MOV     R7.x.1,    C6.z
>    MOV     R7.y.1,    C7.z
>    MUL     R5.x.2,    C6.x, R2.x.1
>    MUL     R5.y.2,    C7.x, R2.x.1
>    MOV     R7.z.1,    C8.z
>    MUL     R5.z.2,    C8.x, R2.x.1
>    MULADD     R5.x.3,    C6.y, R2.y.1, R5.x.2
>    MULADD     R5.y.3,    C7.y, R2.y.1, R5.y.2
>    MULADD     R5.z.3,    C8.y, R2.y.1, R5.z.2
>    MULADD     R5.x.4,    C6.z, R2.z.1, R5.x.3
>    MULADD     R5.y.4,    C7.z, R2.z.1, R5.y.3
>    MULADD     R5.z.4,    C8.z, R2.z.1, R5.z.3
>    DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>    MOV     R6.x.3,    R11.x.2
>    MUL     R5.x.5,    R5.x.4, R11.x.2
>    MUL     R5.y.5,    R5.y.4, R11.x.2
>    MUL     R5.z.5,    R5.z.4, R11.x.2
>    MUL     R6.x.4,    R4.x.5, C0.w
>    MUL     R6.y.2,    R4.y.4, C0.w
>    MUL     R6.z.2,    R4.z.4, C0.w
>    MULADD     R6.x.5,    R4.x.5, -C0.w, C0.x
>    MULADD     R6.y.3,    R4.y.4, -C0.w, C0.y
>    MULADD     R6.z.3,    R4.z.4, -C0.w, C0.z
>    DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>    MOV     R7.x.3,    R11.x.3
>    MUL     R6.x.6,    R6.x.5, R11.x.3
>    MUL     R6.y.4,    R6.y.3, R11.x.3
>    MUL     R6.z.4,    R6.z.3, R11.x.3
>    DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>    MUL     R6.x.8,    R6.x.7, R6.x.7
>    MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>    SQRT_IEEE     R11.x.4,    R6.x.9
>    MOV     R6.x.10,    R11.x.4
>    MUL     R6.x.11,    C1.z, R11.x.4
>    MUL     R5.x.6,    R5.x.5, R6.x.11
>    MUL     R5.y.6,    R5.y.5, R6.x.11
>    MUL     R5.z.6,    R5.z.5, R6.x.11
>    ADD     R4.x.6,    R4.x.5, -R5.x.6
>    ADD     R4.y.5,    R4.y.4, -R5.y.6
>    ADD     R4.z.5,    R4.z.4, -R5.z.6
>    MUL     R5.x.7,    C10.x, R4.x.6
>    MUL     R5.y.7,    C10.y, R4.x.6
>    MUL     R5.z.7,    C10.z, R4.x.6
>    MUL     R5.w.1,    C10.w, R4.x.6
>    MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>    MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>    MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>    MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>    MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>    MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>    MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>    MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>    ADD     R4.x.8,    R4.x.7, C13.x
>    ADD     R4.y.7,    R4.y.6, C13.y
>    ADD     R4.z.7,    R4.z.6, C13.z
>    ADD     R4.w.5,    R4.w.4, C13.w
>    MUL     R5.x.9,    C14.x, R1.x.1
>    MUL     R5.y.9,    C14.y, R1.x.1
>    MUL     R5.z.9,    C14.z, R1.x.1
>    MUL     R5.w.3,    C14.w, R1.x.1
>    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>    ADD     R4.x.9,    R5.x.11, C17.x
>    ADD     R4.y.8,    R5.y.11, C17.y
>    ADD     R4.z.8,    R5.z.11, C17.z
>    ADD     R4.w.6,    R5.w.5, C17.w
>    CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>    CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>    CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>    CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>    MOV     R5.x.13,    R4.x.10
>    MOV     R5.y.13,    R4.y.9
>    MOV     R5.w.7,    R4.w.7
>    RECIP_CLAMPED     R11.x.6,    R4.w.7
>    MOV     R6.x.13,    R11.x.6
>    MUL     R6.x.14,    C1.x, R11.x.6
>    MOV_sat     R6.x.15,    R6.x.14
>    ADD     R6.x.16,    R4.z.9, R6.x.15
>    MAX     R4.x.11,    R6.x.16, -R4.w.7
>    ADD     R11.x.7,    -C1.y, 1|3f800000
>    MUL     R11.x.8,    R11.x.7, R6.x.16
>    MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>    MOV     R5.z.13,    R4.x.12
>    MOV     R3.x.1,    R4.x.10
>    MOV     R3.y.1,    R4.y.9
>    MOV     R3.z.1,    R4.x.12
>    MOV     R3.w.1,    R4.w.7
>    EXPORT  POS   60     R4.x.10, R4.y.9, R4.x.12, R4.w.7
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after dce_cleanup
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    CALL_FS     R1.x.1, R1.y.1, R1.z.1, __, R2.x.1, R2.y.1, R2.z.1, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>    MUL     R4.x.2,    C2.x, R1.x.1
>    MUL     R4.y.1,    C2.y, R1.x.1
>    MUL     R4.z.1,    C2.z, R1.x.1
>    MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>    MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>    MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>    MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>    MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>    MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>    ADD     R4.x.5,    R4.x.4, C5.x
>    ADD     R4.y.4,    R4.y.3, C5.y
>    ADD     R4.z.4,    R4.z.3, C5.z
>    MUL     R5.x.2,    C6.x, R2.x.1
>    MUL     R5.y.2,    C7.x, R2.x.1
>    MUL     R5.z.2,    C8.x, R2.x.1
>    MULADD     R5.x.3,    C6.y, R2.y.1, R5.x.2
>    MULADD     R5.y.3,    C7.y, R2.y.1, R5.y.2
>    MULADD     R5.z.3,    C8.y, R2.y.1, R5.z.2
>    MULADD     R5.x.4,    C6.z, R2.z.1, R5.x.3
>    MULADD     R5.y.4,    C7.z, R2.z.1, R5.y.3
>    MULADD     R5.z.4,    C8.z, R2.z.1, R5.z.3
>    DOT4     R6.x.2, __, __, __,       R5.x.4, R5.x.4, R5.y.4, R5.y.4, R5.z.4, R5.z.4, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>    MUL     R5.x.5,    R5.x.4, R11.x.2
>    MUL     R5.y.5,    R5.y.4, R11.x.2
>    MUL     R5.z.5,    R5.z.4, R11.x.2
>    MULADD     R6.x.5,    R4.x.5, -C0.w, C0.x
>    MULADD     R6.y.3,    R4.y.4, -C0.w, C0.y
>    MULADD     R6.z.3,    R4.z.4, -C0.w, C0.z
>    DOT4     R7.x.2, __, __, __,       R6.x.5, R6.x.5, R6.y.3, R6.y.3, R6.z.3, R6.z.3, 0|00000000, 0|00000000  
>    RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>    MUL     R6.x.6,    R6.x.5, R11.x.3
>    MUL     R6.y.4,    R6.y.3, R11.x.3
>    MUL     R6.z.4,    R6.z.3, R11.x.3
>    DOT4     R6.x.7, __, __, __,       R5.x.5, R6.x.6, R5.y.5, R6.y.4, R5.z.5, R6.z.4, 0|00000000, 0|00000000  
>    MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>    SQRT_IEEE     R11.x.4,    R6.x.9
>    MUL     R6.x.11,    C1.z, R11.x.4
>    MUL     R5.x.6,    R5.x.5, R6.x.11
>    MUL     R5.y.6,    R5.y.5, R6.x.11
>    MUL     R5.z.6,    R5.z.5, R6.x.11
>    ADD     R4.x.6,    R4.x.5, -R5.x.6
>    ADD     R4.y.5,    R4.y.4, -R5.y.6
>    ADD     R4.z.5,    R4.z.4, -R5.z.6
>    MUL     R5.x.7,    C10.x, R4.x.6
>    MUL     R5.y.7,    C10.y, R4.x.6
>    MUL     R5.z.7,    C10.z, R4.x.6
>    MUL     R5.w.1,    C10.w, R4.x.6
>    MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>    MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>    MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>    MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>    MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>    MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>    MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>    MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>    ADD     R4.x.8,    R4.x.7, C13.x
>    ADD     R4.y.7,    R4.y.6, C13.y
>    ADD     R4.z.7,    R4.z.6, C13.z
>    ADD     R4.w.5,    R4.w.4, C13.w
>    MUL     R5.x.9,    C14.x, R1.x.1
>    MUL     R5.y.9,    C14.y, R1.x.1
>    MUL     R5.z.9,    C14.z, R1.x.1
>    MUL     R5.w.3,    C14.w, R1.x.1
>    MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>    MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>    MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>    MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>    MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>    MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>    MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>    MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>    ADD     R4.x.9,    R5.x.11, C17.x
>    ADD     R4.y.8,    R5.y.11, C17.y
>    ADD     R4.z.8,    R5.z.11, C17.z
>    ADD     R4.w.6,    R5.w.5, C17.w
>    CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>    CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>    CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>    CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>    RECIP_CLAMPED     R11.x.6,    R4.w.7
>    MUL     R6.x.14,    C1.x, R11.x.6
>    MOV_sat     R6.x.15,    R6.x.14
>    ADD     R6.x.16,    R4.z.9, R6.x.15
>    MAX     R4.x.11,    R6.x.16, -R4.w.7
>    ADD     R11.x.7,    -C1.y, 1|3f800000
>    MUL     R11.x.8,    R11.x.7, R6.x.16
>    MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>    EXPORT  POS   60     R4.x.10, R4.y.9, R4.x.12, R4.w.7
>    EXPORT  PARAM 0     __, __, __, __
>}  
>
>
>###### after gcm
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    { BB_0    loop_level = 0  
>        CALL_FS     t18F@R1.x, t19F@R1.y, t20F@R1.z, __, t21F@R2.x, t22F@R2.y, t23F@R2.z, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>        ALU     
>        <  
>            (copy) MOV     R1.x.1,    t18F@R1.x
>            MUL     R4.z.1,    C2.z, R1.x.1
>            MUL     R4.y.1,    C2.y, R1.x.1
>            MUL     R4.x.2,    C2.x, R1.x.1
>            (copy) MOV     R1.y.1,    t19F@R1.y
>            MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>            MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>            MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>            (copy) MOV     R1.z.1,    t20F@R1.z
>            MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>            MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>            MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>            (copy) MOV     R2.x.1,    t21F@R2.x
>            MUL     R5.z.2,    C8.x, R2.x.1
>            MUL     R5.y.2,    C7.x, R2.x.1
>            MUL     R5.x.2,    C6.x, R2.x.1
>            ADD     R4.z.4,    R4.z.3, C5.z
>            ADD     R4.y.4,    R4.y.3, C5.y
>            ADD     R4.x.5,    R4.x.4, C5.x
>            (copy) MOV     R2.y.1,    t22F@R2.y
>            MULADD     R5.z.3,    C8.y, R2.y.1, R5.z.2
>            MULADD     R5.y.3,    C7.y, R2.y.1, R5.y.2
>            MULADD     R5.x.3,    C6.y, R2.y.1, R5.x.2
>            MULADD     R6.z.3,    R4.z.4, -C0.w, C0.z
>            MULADD     R6.y.3,    R4.y.4, -C0.w, C0.y
>            MULADD     R6.x.5,    R4.x.5, -C0.w, C0.x
>            (copy) MOV     R2.z.1,    t23F@R2.z
>            MULADD     R5.z.4,    C8.z, R2.z.1, R5.z.3
>            MULADD     R5.y.4,    C7.z, R2.z.1, R5.y.3
>            MULADD     R5.x.4,    C6.z, R2.z.1, R5.x.3
>            (copy) MOV     t14,    R6.z.3
>            (copy) MOV     t13,    R6.y.3
>            (copy) MOV     t12,    R6.x.5
>            DOT4     R7.x.2, __, __, __,       t12, t12, t13, t13, t14, t14, 0|00000000, 0|00000000  
>            (copy) MOV     t17,    R5.z.4
>            (copy) MOV     t16,    R5.y.4
>            (copy) MOV     t15,    R5.x.4
>            DOT4     R6.x.2, __, __, __,       t15, t15, t16, t16, t17, t17, 0|00000000, 0|00000000  
>            RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>            RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>            MUL     R6.z.4,    R6.z.3, R11.x.3
>            MUL     R5.z.5,    R5.z.4, R11.x.2
>            MUL     R6.y.4,    R6.y.3, R11.x.3
>            MUL     R5.y.5,    R5.y.4, R11.x.2
>            MUL     R6.x.6,    R6.x.5, R11.x.3
>            MUL     R5.x.5,    R5.x.4, R11.x.2
>            (copy) MOV     t11,    R6.z.4
>            (copy) MOV     t10,    R5.z.5
>            (copy) MOV     t9,    R6.y.4
>            (copy) MOV     t8,    R5.y.5
>            (copy) MOV     t7,    R6.x.6
>            (copy) MOV     t6,    R5.x.5
>            DOT4     R6.x.7, __, __, __,       t6, t7, t8, t9, t10, t11, 0|00000000, 0|00000000  
>            MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>            SQRT_IEEE     R11.x.4,    R6.x.9
>            MUL     R6.x.11,    C1.z, R11.x.4
>            MUL     R5.x.6,    R5.x.5, R6.x.11
>            MUL     R5.y.6,    R5.y.5, R6.x.11
>            ADD     R4.x.6,    R4.x.5, -R5.x.6
>            MUL     R5.w.3,    C14.w, R1.x.1
>            MUL     R5.z.6,    R5.z.5, R6.x.11
>            ADD     R4.y.5,    R4.y.4, -R5.y.6
>            MUL     R5.w.1,    C10.w, R4.x.6
>            MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>            ADD     R4.z.5,    R4.z.4, -R5.z.6
>            MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>            MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>            MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>            MUL     R5.z.9,    C14.z, R1.x.1
>            MUL     R5.z.7,    C10.z, R4.x.6
>            ADD     R4.w.6,    R5.w.5, C17.w
>            ADD     R4.w.5,    R4.w.4, C13.w
>            MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>            MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>            CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>            MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>            MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>            RECIP_CLAMPED     R11.x.6,    R4.w.7
>            ADD     R4.z.8,    R5.z.11, C17.z
>            ADD     R4.z.7,    R4.z.6, C13.z
>            MUL     R6.x.14,    C1.x, R11.x.6
>            MUL     R5.y.9,    C14.y, R1.x.1
>            MUL     R5.y.7,    C10.y, R4.x.6
>            MUL     R5.x.9,    C14.x, R1.x.1
>            MUL     R5.x.7,    C10.x, R4.x.6
>            CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>            MOV_sat     R6.x.15,    R6.x.14
>            MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>            MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>            MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>            MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>            ADD     R6.x.16,    R4.z.9, R6.x.15
>            ADD     R11.x.7,    -C1.y, 1|3f800000
>            MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>            MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>            MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>            MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>            MAX     R4.x.11,    R6.x.16, -R4.w.7
>            MUL     R11.x.8,    R11.x.7, R6.x.16
>            ADD     R4.y.8,    R5.y.11, C17.y
>            ADD     R4.y.7,    R4.y.6, C13.y
>            ADD     R4.x.9,    R5.x.11, C17.x
>            ADD     R4.x.8,    R4.x.7, C13.x
>            MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>            CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>            CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>            (copy) MOV     t5,    R4.w.7
>            (copy) MOV     t4,    R4.x.12
>            (copy) MOV     t3,    R4.y.9
>            (copy) MOV     t2,    R4.x.10
>        >  
>        EXPORT  POS   60     t2, t3, t4, t5
>        EXPORT  PARAM 0     __, __, __, __
>    } end BB_0  
>}  
>
>
>###### after dce_cleanup
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    { BB_0    loop_level = 0  live_before: [R0.xF@R0.x R0.yF@R0.y R0.zF@R0.z R0.wF@R0.w ]
>        CALL_FS     t18F@R1.x, t19F@R1.y, t20F@R1.z, __, t21F@R2.x, t22F@R2.y, t23F@R2.z, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>        ALU     
>        <  live_before: [t18F@R1.x t19F@R1.y t20F@R1.z t21F@R2.x t22F@R2.y t23F@R2.z ]
>            (copy) MOV     R1.x.1,    t18F@R1.x
>            MUL     R4.z.1,    C2.z, R1.x.1
>            MUL     R4.y.1,    C2.y, R1.x.1
>            MUL     R4.x.2,    C2.x, R1.x.1
>            (copy) MOV     R1.y.1,    t19F@R1.y
>            MULADD     R4.z.2,    C3.z, R1.y.1, R4.z.1
>            MULADD     R4.y.2,    C3.y, R1.y.1, R4.y.1
>            MULADD     R4.x.3,    C3.x, R1.y.1, R4.x.2
>            (copy) MOV     R1.z.1,    t20F@R1.z
>            MULADD     R4.z.3,    C4.z, R1.z.1, R4.z.2
>            MULADD     R4.y.3,    C4.y, R1.z.1, R4.y.2
>            MULADD     R4.x.4,    C4.x, R1.z.1, R4.x.3
>            (copy) MOV     R2.x.1,    t21F@R2.x
>            MUL     R5.z.2,    C8.x, R2.x.1
>            MUL     R5.y.2,    C7.x, R2.x.1
>            MUL     R5.x.2,    C6.x, R2.x.1
>            ADD     R4.z.4,    R4.z.3, C5.z
>            ADD     R4.y.4,    R4.y.3, C5.y
>            ADD     R4.x.5,    R4.x.4, C5.x
>            (copy) MOV     R2.y.1,    t22F@R2.y
>            MULADD     R5.z.3,    C8.y, R2.y.1, R5.z.2
>            MULADD     R5.y.3,    C7.y, R2.y.1, R5.y.2
>            MULADD     R5.x.3,    C6.y, R2.y.1, R5.x.2
>            MULADD     R6.z.3,    R4.z.4, -C0.w, C0.z
>            MULADD     R6.y.3,    R4.y.4, -C0.w, C0.y
>            MULADD     R6.x.5,    R4.x.5, -C0.w, C0.x
>            (copy) MOV     R2.z.1,    t23F@R2.z
>            MULADD     R5.z.4,    C8.z, R2.z.1, R5.z.3
>            MULADD     R5.y.4,    C7.z, R2.z.1, R5.y.3
>            MULADD     R5.x.4,    C6.z, R2.z.1, R5.x.3
>            (copy) MOV     t14,    R6.z.3
>            (copy) MOV     t13,    R6.y.3
>            (copy) MOV     t12,    R6.x.5
>            DOT4     R7.x.2, __, __, __,       t12, t12, t13, t13, t14, t14, 0|00000000, 0|00000000  
>            (copy) MOV     t17,    R5.z.4
>            (copy) MOV     t16,    R5.y.4
>            (copy) MOV     t15,    R5.x.4
>            DOT4     R6.x.2, __, __, __,       t15, t15, t16, t16, t17, t17, 0|00000000, 0|00000000  
>            RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>            RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>            MUL     R6.z.4,    R6.z.3, R11.x.3
>            MUL     R5.z.5,    R5.z.4, R11.x.2
>            MUL     R6.y.4,    R6.y.3, R11.x.3
>            MUL     R5.y.5,    R5.y.4, R11.x.2
>            MUL     R6.x.6,    R6.x.5, R11.x.3
>            MUL     R5.x.5,    R5.x.4, R11.x.2
>            (copy) MOV     t11,    R6.z.4
>            (copy) MOV     t10,    R5.z.5
>            (copy) MOV     t9,    R6.y.4
>            (copy) MOV     t8,    R5.y.5
>            (copy) MOV     t7,    R6.x.6
>            (copy) MOV     t6,    R5.x.5
>            DOT4     R6.x.7, __, __, __,       t6, t7, t8, t9, t10, t11, 0|00000000, 0|00000000  
>            MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>            SQRT_IEEE     R11.x.4,    R6.x.9
>            MUL     R6.x.11,    C1.z, R11.x.4
>            MUL     R5.x.6,    R5.x.5, R6.x.11
>            MUL     R5.y.6,    R5.y.5, R6.x.11
>            ADD     R4.x.6,    R4.x.5, -R5.x.6
>            MUL     R5.w.3,    C14.w, R1.x.1
>            MUL     R5.z.6,    R5.z.5, R6.x.11
>            ADD     R4.y.5,    R4.y.4, -R5.y.6
>            MUL     R5.w.1,    C10.w, R4.x.6
>            MULADD     R5.w.4,    C15.w, R1.y.1, R5.w.3
>            ADD     R4.z.5,    R4.z.4, -R5.z.6
>            MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>            MULADD     R5.w.5,    C16.w, R1.z.1, R5.w.4
>            MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>            MUL     R5.z.9,    C14.z, R1.x.1
>            MUL     R5.z.7,    C10.z, R4.x.6
>            ADD     R4.w.6,    R5.w.5, C17.w
>            ADD     R4.w.5,    R4.w.4, C13.w
>            MULADD     R5.z.10,    C15.z, R1.y.1, R5.z.9
>            MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>            CNDE     R4.w.7,    C1.z, R4.w.6, R4.w.5
>            MULADD     R5.z.11,    C16.z, R1.z.1, R5.z.10
>            MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>            RECIP_CLAMPED     R11.x.6,    R4.w.7
>            ADD     R4.z.8,    R5.z.11, C17.z
>            ADD     R4.z.7,    R4.z.6, C13.z
>            MUL     R6.x.14,    C1.x, R11.x.6
>            MUL     R5.y.9,    C14.y, R1.x.1
>            MUL     R5.y.7,    C10.y, R4.x.6
>            MUL     R5.x.9,    C14.x, R1.x.1
>            MUL     R5.x.7,    C10.x, R4.x.6
>            CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>            MOV_sat     R6.x.15,    R6.x.14
>            MULADD     R5.y.10,    C15.y, R1.y.1, R5.y.9
>            MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>            MULADD     R5.x.10,    C15.x, R1.y.1, R5.x.9
>            MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>            ADD     R6.x.16,    R4.z.9, R6.x.15
>            ADD     R11.x.7,    -C1.y, 1|3f800000
>            MULADD     R5.y.11,    C16.y, R1.z.1, R5.y.10
>            MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>            MULADD     R5.x.11,    C16.x, R1.z.1, R5.x.10
>            MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>            MAX     R4.x.11,    R6.x.16, -R4.w.7
>            MUL     R11.x.8,    R11.x.7, R6.x.16
>            ADD     R4.y.8,    R5.y.11, C17.y
>            ADD     R4.y.7,    R4.y.6, C13.y
>            ADD     R4.x.9,    R5.x.11, C17.x
>            ADD     R4.x.8,    R4.x.7, C13.x
>            MULADD     R4.x.12,    C1.y, R4.x.11, R11.x.8
>            CNDE     R4.y.9,    C1.z, R4.y.8, R4.y.7
>            CNDE     R4.x.10,    C1.z, R4.x.9, R4.x.8
>            (copy) MOV     t5,    R4.w.7
>            (copy) MOV     t4,    R4.x.12
>            (copy) MOV     t3,    R4.y.9
>            (copy) MOV     t2,    R4.x.10
>        >  live_after: [t2 t3 t4 t5 ]
>        EXPORT  POS   60     t2, t3, t4, t5
>        EXPORT  PARAM 0     __, __, __, __
>    } end BB_0  
>}  
>
>
>###### after ra_coalesce
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    { BB_0    loop_level = 0  live_before: [R0.xF@R0.x R0.yF@R0.y R0.zF@R0.z R0.wF@R0.w ]
>        CALL_FS     t18F@R1.x, t19F@R1.y, t20F@R1.z, __, t21F@R2.x, t22F@R2.y, t23F@R2.z, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>        ALU     
>        <  live_before: [t18F@R1.x t19F@R1.y t20F@R1.z t21F@R2.x t22F@R2.y t23F@R2.z ]
>            (copy) MOV     R1.x.1F@R3.x,    t18F@R1.x
>            MUL     R4.z.1,    C2.z, R1.x.1F@R3.x
>            MUL     R4.y.1,    C2.y, R1.x.1F@R3.x
>            MUL     R4.x.2,    C2.x, R1.x.1F@R3.x
>            (copy) MOV     R1.y.1F@R3.y,    t19F@R1.y
>            MULADD     R4.z.2,    C3.z, R1.y.1F@R3.y, R4.z.1
>            MULADD     R4.y.2,    C3.y, R1.y.1F@R3.y, R4.y.1
>            MULADD     R4.x.3,    C3.x, R1.y.1F@R3.y, R4.x.2
>            (copy) MOV     R1.z.1F@R3.z,    t20F@R1.z
>            MULADD     R4.z.3,    C4.z, R1.z.1F@R3.z, R4.z.2
>            MULADD     R4.y.3,    C4.y, R1.z.1F@R3.z, R4.y.2
>            MULADD     R4.x.4,    C4.x, R1.z.1F@R3.z, R4.x.3
>            (copy) MOV     R2.x.1F@R2.x,    t21F@R2.x
>            MUL     R5.z.2,    C8.x, R2.x.1F@R2.x
>            MUL     R5.y.2,    C7.x, R2.x.1F@R2.x
>            MUL     R5.x.2,    C6.x, R2.x.1F@R2.x
>            ADD     R4.z.4,    R4.z.3, C5.z
>            ADD     R4.y.4,    R4.y.3, C5.y
>            ADD     R4.x.5,    R4.x.4, C5.x
>            (copy) MOV     R2.y.1F@R2.y,    t22F@R2.y
>            MULADD     R5.z.3,    C8.y, R2.y.1F@R2.y, R5.z.2
>            MULADD     R5.y.3,    C7.y, R2.y.1F@R2.y, R5.y.2
>            MULADD     R5.x.3,    C6.y, R2.y.1F@R2.y, R5.x.2
>            MULADD     R6.z.3@R0.z,    R4.z.4, -C0.w, C0.z
>            MULADD     R6.y.3@R0.y,    R4.y.4, -C0.w, C0.y
>            MULADD     R6.x.5@R0.x,    R4.x.5, -C0.w, C0.x
>            (copy) MOV     R2.z.1F@R2.z,    t23F@R2.z
>            MULADD     R5.z.4@R0.w,    C8.z, R2.z.1F@R2.z, R5.z.3
>            MULADD     R5.y.4@R1.w,    C7.z, R2.z.1F@R2.z, R5.y.3
>            MULADD     R5.x.4@R1.z,    C6.z, R2.z.1F@R2.z, R5.x.3
>            (copy) MOV     t14@R0.z,    R6.z.3@R0.z
>            (copy) MOV     t13@R0.y,    R6.y.3@R0.y
>            (copy) MOV     t12@R0.x,    R6.x.5@R0.x
>            DOT4     R7.x.2, __, __, __,       t12@R0.x, t12@R0.x, t13@R0.y, t13@R0.y, t14@R0.z, t14@R0.z, 0|00000000, 0|00000000  
>            (copy) MOV     t17@R0.w,    R5.z.4@R0.w
>            (copy) MOV     t16@R1.w,    R5.y.4@R1.w
>            (copy) MOV     t15@R1.z,    R5.x.4@R1.z
>            DOT4     R6.x.2, __, __, __,       t15@R1.z, t15@R1.z, t16@R1.w, t16@R1.w, t17@R0.w, t17@R0.w, 0|00000000, 0|00000000  
>            RECIPSQRT_CLAMPED     R11.x.3,    |R7.x.2|
>            RECIPSQRT_CLAMPED     R11.x.2,    |R6.x.2|
>            MUL     R6.z.4@R1.y,    R6.z.3@R0.z, R11.x.3
>            MUL     R5.z.5@R1.x,    R5.z.4@R0.w, R11.x.2
>            MUL     R6.y.4@R0.w,    R6.y.3@R0.y, R11.x.3
>            MUL     R5.y.5@R0.z,    R5.y.4@R1.w, R11.x.2
>            MUL     R6.x.6@R0.y,    R6.x.5@R0.x, R11.x.3
>            MUL     R5.x.5@R0.x,    R5.x.4@R1.z, R11.x.2
>            (copy) MOV     t11@R1.y,    R6.z.4@R1.y
>            (copy) MOV     t10@R1.x,    R5.z.5@R1.x
>            (copy) MOV     t9@R0.w,    R6.y.4@R0.w
>            (copy) MOV     t8@R0.z,    R5.y.5@R0.z
>            (copy) MOV     t7@R0.y,    R6.x.6@R0.y
>            (copy) MOV     t6@R0.x,    R5.x.5@R0.x
>            DOT4     R6.x.7, __, __, __,       t6@R0.x, t7@R0.y, t8@R0.z, t9@R0.w, t10@R1.x, t11@R1.y, 0|00000000, 0|00000000  
>            MULADD     R6.x.9,    -R6.x.7, R6.x.7, 1|3f800000
>            SQRT_IEEE     R11.x.4,    R6.x.9
>            MUL     R6.x.11,    C1.z, R11.x.4
>            MUL     R5.x.6,    R5.x.5@R0.x, R6.x.11
>            MUL     R5.y.6,    R5.y.5@R0.z, R6.x.11
>            ADD     R4.x.6,    R4.x.5, -R5.x.6
>            MUL     R5.w.3,    C14.w, R1.x.1F@R3.x
>            MUL     R5.z.6,    R5.z.5@R1.x, R6.x.11
>            ADD     R4.y.5,    R4.y.4, -R5.y.6
>            MUL     R5.w.1,    C10.w, R4.x.6
>            MULADD     R5.w.4,    C15.w, R1.y.1F@R3.y, R5.w.3
>            ADD     R4.z.5,    R4.z.4, -R5.z.6
>            MULADD     R5.w.2,    C11.w, R4.y.5, R5.w.1
>            MULADD     R5.w.5,    C16.w, R1.z.1F@R3.z, R5.w.4
>            MULADD     R4.w.4,    C12.w, R4.z.5, R5.w.2
>            MUL     R5.z.9,    C14.z, R1.x.1F@R3.x
>            MUL     R5.z.7,    C10.z, R4.x.6
>            ADD     R4.w.6,    R5.w.5, C17.w
>            ADD     R4.w.5,    R4.w.4, C13.w
>            MULADD     R5.z.10,    C15.z, R1.y.1F@R3.y, R5.z.9
>            MULADD     R5.z.8,    C11.z, R4.y.5, R5.z.7
>            CNDE     R4.w.7FP@R0.w,    C1.z, R4.w.6, R4.w.5
>            MULADD     R5.z.11,    C16.z, R1.z.1F@R3.z, R5.z.10
>            MULADD     R4.z.6,    C12.z, R4.z.5, R5.z.8
>            RECIP_CLAMPED     R11.x.6,    R4.w.7FP@R0.w
>            ADD     R4.z.8,    R5.z.11, C17.z
>            ADD     R4.z.7,    R4.z.6, C13.z
>            MUL     R6.x.14,    C1.x, R11.x.6
>            MUL     R5.y.9,    C14.y, R1.x.1F@R3.x
>            MUL     R5.y.7,    C10.y, R4.x.6
>            MUL     R5.x.9,    C14.x, R1.x.1F@R3.x
>            MUL     R5.x.7,    C10.x, R4.x.6
>            CNDE     R4.z.9,    C1.z, R4.z.8, R4.z.7
>            MOV_sat     R6.x.15,    R6.x.14
>            MULADD     R5.y.10,    C15.y, R1.y.1F@R3.y, R5.y.9
>            MULADD     R5.y.8,    C11.y, R4.y.5, R5.y.7
>            MULADD     R5.x.10,    C15.x, R1.y.1F@R3.y, R5.x.9
>            MULADD     R5.x.8,    C11.x, R4.y.5, R5.x.7
>            ADD     R6.x.16,    R4.z.9, R6.x.15
>            ADD     R11.x.7,    -C1.y, 1|3f800000
>            MULADD     R5.y.11,    C16.y, R1.z.1F@R3.z, R5.y.10
>            MULADD     R4.y.6,    C12.y, R4.z.5, R5.y.8
>            MULADD     R5.x.11,    C16.x, R1.z.1F@R3.z, R5.x.10
>            MULADD     R4.x.7,    C12.x, R4.z.5, R5.x.8
>            MAX     R4.x.11,    R6.x.16, -R4.w.7FP@R0.w
>            MUL     R11.x.8,    R11.x.7, R6.x.16
>            ADD     R4.y.8,    R5.y.11, C17.y
>            ADD     R4.y.7,    R4.y.6, C13.y
>            ADD     R4.x.9,    R5.x.11, C17.x
>            ADD     R4.x.8,    R4.x.7, C13.x
>            MULADD     R4.x.12FP@R0.z,    C1.y, R4.x.11, R11.x.8
>            CNDE     R4.y.9FP@R0.y,    C1.z, R4.y.8, R4.y.7
>            CNDE     R4.x.10FP@R0.x,    C1.z, R4.x.9, R4.x.8
>            (copy) MOV     t5FP@R0.w,    R4.w.7FP@R0.w
>            (copy) MOV     t4FP@R0.z,    R4.x.12FP@R0.z
>            (copy) MOV     t3FP@R0.y,    R4.y.9FP@R0.y
>            (copy) MOV     t2FP@R0.x,    R4.x.10FP@R0.x
>        >  live_after: [t2FP@R0.x t3FP@R0.y t4FP@R0.z t5FP@R0.w ]
>        EXPORT  POS   60     t2FP@R0.x, t3FP@R0.y, t4FP@R0.z, t5FP@R0.w
>        EXPORT  PARAM 0     __, __, __, __
>    } end BB_0  
>}  
>
>
>###### after ra_init
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    { BB_0    loop_level = 0  live_before: [R0.xF@R0.x R0.yF@R0.y R0.zF@R0.z R0.wF@R0.w ]
>        CALL_FS     t18F@R1.x, t19F@R1.y, t20F@R1.z, __, t21F@R2.x, t22F@R2.y, t23F@R2.z, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>        ALU     
>        <  live_before: [t18F@R1.x t19F@R1.y t20F@R1.z t21F@R2.x t22F@R2.y t23F@R2.z ]
>            (copy) MOV     R1.x.1F@R3.x,    t18F@R1.x
>            MUL     R4.z.1@R0.x,    C2.z, R1.x.1F@R3.x
>            MUL     R4.y.1@R0.y,    C2.y, R1.x.1F@R3.x
>            MUL     R4.x.2@R0.z,    C2.x, R1.x.1F@R3.x
>            (copy) MOV     R1.y.1F@R3.y,    t19F@R1.y
>            MULADD     R4.z.2@R0.x,    C3.z, R1.y.1F@R3.y, R4.z.1@R0.x
>            MULADD     R4.y.2@R0.y,    C3.y, R1.y.1F@R3.y, R4.y.1@R0.y
>            MULADD     R4.x.3@R0.z,    C3.x, R1.y.1F@R3.y, R4.x.2@R0.z
>            (copy) MOV     R1.z.1F@R3.z,    t20F@R1.z
>            MULADD     R4.z.3@R0.x,    C4.z, R1.z.1F@R3.z, R4.z.2@R0.x
>            MULADD     R4.y.3@R0.y,    C4.y, R1.z.1F@R3.z, R4.y.2@R0.y
>            MULADD     R4.x.4@R0.z,    C4.x, R1.z.1F@R3.z, R4.x.3@R0.z
>            (copy) MOV     R2.x.1F@R2.x,    t21F@R2.x
>            MUL     R5.z.2@R0.w,    C8.x, R2.x.1F@R2.x
>            MUL     R5.y.2@R1.x,    C7.x, R2.x.1F@R2.x
>            MUL     R5.x.2@R1.y,    C6.x, R2.x.1F@R2.x
>            ADD     R4.z.4@R2.x,    R4.z.3@R0.x, C5.z
>            ADD     R4.y.4@R2.w,    R4.y.3@R0.y, C5.y
>            ADD     R4.x.5@R3.w,    R4.x.4@R0.z, C5.x
>            (copy) MOV     R2.y.1F@R2.y,    t22F@R2.y
>            MULADD     R5.z.3@R0.w,    C8.y, R2.y.1F@R2.y, R5.z.2@R0.w
>            MULADD     R5.y.3@R1.x,    C7.y, R2.y.1F@R2.y, R5.y.2@R1.x
>            MULADD     R5.x.3@R1.y,    C6.y, R2.y.1F@R2.y, R5.x.2@R1.y
>            MULADD     R6.z.3@R0.z,    R4.z.4@R2.x, -C0.w, C0.z
>            MULADD     R6.y.3@R0.y,    R4.y.4@R2.w, -C0.w, C0.y
>            MULADD     R6.x.5@R0.x,    R4.x.5@R3.w, -C0.w, C0.x
>            (copy) MOV     R2.z.1F@R2.z,    t23F@R2.z
>            MULADD     R5.z.4@R0.w,    C8.z, R2.z.1F@R2.z, R5.z.3@R0.w
>            MULADD     R5.y.4@R1.w,    C7.z, R2.z.1F@R2.z, R5.y.3@R1.x
>            MULADD     R5.x.4@R1.z,    C6.z, R2.z.1F@R2.z, R5.x.3@R1.y
>            (copy) MOV     t14@R0.z,    R6.z.3@R0.z
>            (copy) MOV     t13@R0.y,    R6.y.3@R0.y
>            (copy) MOV     t12@R0.x,    R6.x.5@R0.x
>            DOT4     R7.x.2@R1.x, __, __, __,       t12@R0.x, t12@R0.x, t13@R0.y, t13@R0.y, t14@R0.z, t14@R0.z, 0|00000000, 0|00000000  
>            (copy) MOV     t17@R0.w,    R5.z.4@R0.w
>            (copy) MOV     t16@R1.w,    R5.y.4@R1.w
>            (copy) MOV     t15@R1.z,    R5.x.4@R1.z
>            DOT4     R6.x.2@R1.y, __, __, __,       t15@R1.z, t15@R1.z, t16@R1.w, t16@R1.w, t17@R0.w, t17@R0.w, 0|00000000, 0|00000000  
>            RECIPSQRT_CLAMPED     R11.x.3@R2.y,    |R7.x.2@R1.x|
>            RECIPSQRT_CLAMPED     R11.x.2@R2.z,    |R6.x.2@R1.y|
>            MUL     R6.z.4@R1.y,    R6.z.3@R0.z, R11.x.3@R2.y
>            MUL     R5.z.5@R1.x,    R5.z.4@R0.w, R11.x.2@R2.z
>            MUL     R6.y.4@R0.w,    R6.y.3@R0.y, R11.x.3@R2.y
>            MUL     R5.y.5@R0.z,    R5.y.4@R1.w, R11.x.2@R2.z
>            MUL     R6.x.6@R0.y,    R6.x.5@R0.x, R11.x.3@R2.y
>            MUL     R5.x.5@R0.x,    R5.x.4@R1.z, R11.x.2@R2.z
>            (copy) MOV     t11@R1.y,    R6.z.4@R1.y
>            (copy) MOV     t10@R1.x,    R5.z.5@R1.x
>            (copy) MOV     t9@R0.w,    R6.y.4@R0.w
>            (copy) MOV     t8@R0.z,    R5.y.5@R0.z
>            (copy) MOV     t7@R0.y,    R6.x.6@R0.y
>            (copy) MOV     t6@R0.x,    R5.x.5@R0.x
>            DOT4     R6.x.7@R0.y, __, __, __,       t6@R0.x, t7@R0.y, t8@R0.z, t9@R0.w, t10@R1.x, t11@R1.y, 0|00000000, 0|00000000  
>            MULADD     R6.x.9@R0.y,    -R6.x.7@R0.y, R6.x.7@R0.y, 1|3f800000
>            SQRT_IEEE     R11.x.4@R0.y,    R6.x.9@R0.y
>            MUL     R6.x.11@R0.y,    C1.z, R11.x.4@R0.y
>            MUL     R5.x.6@R0.x,    R5.x.5@R0.x, R6.x.11@R0.y
>            MUL     R5.y.6@R0.z,    R5.y.5@R0.z, R6.x.11@R0.y
>            ADD     R4.x.6@R0.x,    R4.x.5@R3.w, -R5.x.6@R0.x
>            MUL     R5.w.3@R0.w,    C14.w, R1.x.1F@R3.x
>            MUL     R5.z.6@R0.y,    R5.z.5@R1.x, R6.x.11@R0.y
>            ADD     R4.y.5@R0.z,    R4.y.4@R2.w, -R5.y.6@R0.z
>            MUL     R5.w.1@R1.x,    C10.w, R4.x.6@R0.x
>            MULADD     R5.w.4@R0.w,    C15.w, R1.y.1F@R3.y, R5.w.3@R0.w
>            ADD     R4.z.5@R0.y,    R4.z.4@R2.x, -R5.z.6@R0.y
>            MULADD     R5.w.2@R1.x,    C11.w, R4.y.5@R0.z, R5.w.1@R1.x
>            MULADD     R5.w.5@R0.w,    C16.w, R1.z.1F@R3.z, R5.w.4@R0.w
>            MULADD     R4.w.4@R1.x,    C12.w, R4.z.5@R0.y, R5.w.2@R1.x
>            MUL     R5.z.9@R1.y,    C14.z, R1.x.1F@R3.x
>            MUL     R5.z.7@R1.z,    C10.z, R4.x.6@R0.x
>            ADD     R4.w.6@R0.w,    R5.w.5@R0.w, C17.w
>            ADD     R4.w.5@R1.x,    R4.w.4@R1.x, C13.w
>            MULADD     R5.z.10@R1.y,    C15.z, R1.y.1F@R3.y, R5.z.9@R1.y
>            MULADD     R5.z.8@R1.z,    C11.z, R4.y.5@R0.z, R5.z.7@R1.z
>            CNDE     R4.w.7FP@R0.w,    C1.z, R4.w.6@R0.w, R4.w.5@R1.x
>            MULADD     R5.z.11@R1.x,    C16.z, R1.z.1F@R3.z, R5.z.10@R1.y
>            MULADD     R4.z.6@R1.y,    C12.z, R4.z.5@R0.y, R5.z.8@R1.z
>            RECIP_CLAMPED     R11.x.6@R1.z,    R4.w.7FP@R0.w
>            ADD     R4.z.8@R1.x,    R5.z.11@R1.x, C17.z
>            ADD     R4.z.7@R1.y,    R4.z.6@R1.y, C13.z
>            MUL     R6.x.14@R1.z,    C1.x, R11.x.6@R1.z
>            MUL     R5.y.9@R1.w,    C14.y, R1.x.1F@R3.x
>            MUL     R5.y.7@R2.x,    C10.y, R4.x.6@R0.x
>            MUL     R5.x.9@R2.y,    C14.x, R1.x.1F@R3.x
>            MUL     R5.x.7@R0.x,    C10.x, R4.x.6@R0.x
>            CNDE     R4.z.9@R1.x,    C1.z, R4.z.8@R1.x, R4.z.7@R1.y
>            MOV_sat     R6.x.15@R1.y,    R6.x.14@R1.z
>            MULADD     R5.y.10@R1.z,    C15.y, R1.y.1F@R3.y, R5.y.9@R1.w
>            MULADD     R5.y.8@R1.w,    C11.y, R4.y.5@R0.z, R5.y.7@R2.x
>            MULADD     R5.x.10@R2.x,    C15.x, R1.y.1F@R3.y, R5.x.9@R2.y
>            MULADD     R5.x.8@R0.x,    C11.x, R4.y.5@R0.z, R5.x.7@R0.x
>            ADD     R6.x.16@R0.z,    R4.z.9@R1.x, R6.x.15@R1.y
>            ADD     R11.x.7@R1.x,    -C1.y, 1|3f800000
>            MULADD     R5.y.11@R1.y,    C16.y, R1.z.1F@R3.z, R5.y.10@R1.z
>            MULADD     R4.y.6@R1.z,    C12.y, R4.z.5@R0.y, R5.y.8@R1.w
>            MULADD     R5.x.11@R1.w,    C16.x, R1.z.1F@R3.z, R5.x.10@R2.x
>            MULADD     R4.x.7@R0.x,    C12.x, R4.z.5@R0.y, R5.x.8@R0.x
>            MAX     R4.x.11@R0.y,    R6.x.16@R0.z, -R4.w.7FP@R0.w
>            MUL     R11.x.8@R0.z,    R11.x.7@R1.x, R6.x.16@R0.z
>            ADD     R4.y.8@R1.x,    R5.y.11@R1.y, C17.y
>            ADD     R4.y.7@R1.y,    R4.y.6@R1.z, C13.y
>            ADD     R4.x.9@R1.z,    R5.x.11@R1.w, C17.x
>            ADD     R4.x.8@R0.x,    R4.x.7@R0.x, C13.x
>            MULADD     R4.x.12FP@R0.z,    C1.y, R4.x.11@R0.y, R11.x.8@R0.z
>            CNDE     R4.y.9FP@R0.y,    C1.z, R4.y.8@R1.x, R4.y.7@R1.y
>            CNDE     R4.x.10FP@R0.x,    C1.z, R4.x.9@R1.z, R4.x.8@R0.x
>            (copy) MOV     t5FP@R0.w,    R4.w.7FP@R0.w
>            (copy) MOV     t4FP@R0.z,    R4.x.12FP@R0.z
>            (copy) MOV     t3FP@R0.y,    R4.y.9FP@R0.y
>            (copy) MOV     t2FP@R0.x,    R4.x.10FP@R0.x
>        >  live_after: [t2FP@R0.x t3FP@R0.y t4FP@R0.z t5FP@R0.w ]
>        EXPORT  POS   60     t2FP@R0.x, t3FP@R0.y, t4FP@R0.z, t5FP@R0.w
>        EXPORT  PARAM 0     __, __, __, __
>    } end BB_0  
>}  
>
>
>###### after post_scheduler
>{   preloaded inputs [R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w]  
>    { BB_0    loop_level = 0  live_before: [R0.xF@R0.x R0.yF@R0.y R0.zF@R0.z R0.wF@R0.w ]
>        CALL_FS     t18||FP@R1.x, t19||FP@R1.y, t20||FP@R1.z, __, t21||FP@R2.x, t22||FP@R2.y, t23||FP@R2.z, __,       R0.xF@R0.x, R0.yF@R0.y, R0.zF@R0.z, R0.wF@R0.w
>        ALU     
>        <  
>            [  
>                (copy) MOV     R1.x.1F@R0.x,    t18||FP@R1.x
>            ]  
>            [  
>                MUL     R4.z.1F@R124.x,    C2.z, R1.x.1F@R0.x
>                MUL     R4.y.1F@R124.y,    C2.y, R1.x.1F@R0.x
>                MUL     R4.x.2F@R124.z,    C2.x, R1.x.1F@R0.x
>                (copy) MOV     R1.y.1F@R0.y,    t19||FP@R1.y
>            ]  
>            [  
>                MULADD     R4.z.2F@R124.x,    C3.z, R1.y.1F@R0.y, R4.z.1F@R124.x
>                MULADD     R4.y.2F@R124.y,    C3.y, R1.y.1F@R0.y, R4.y.1F@R124.y
>                MULADD     R4.x.3F@R124.z,    C3.x, R1.y.1F@R0.y, R4.x.2F@R124.z
>                (copy) MOV     R1.z.1F@R127.z,    t20||FP@R1.z
>            ]  
>            [  
>                MULADD     R4.y.3F@R125.y,    C4.y, R1.z.1F@R127.z, R4.y.2F@R124.y
>                MULADD     R4.x.4F@R124.z,    C4.x, R1.z.1F@R127.z, R4.x.3F@R124.z
>                MUL     R5.z.2F@R124.w,    C8.x, R2.x.1||FP@R2.x
>            ]  
>            [  
>                MULADD     R4.z.3F@R125.x,    C4.z, R1.z.1F@R127.z, R4.z.2F@R124.x
>                MUL     R5.x.2F@R124.y,    C6.x, R2.x.1||FP@R2.x
>            ]  
>            [  
>                MUL     R5.y.2F@R124.x,    C7.x, R2.x.1||FP@R2.x
>                ADD     R4.x.5F@R126.w,    R4.x.4F@R124.z, C5.x
>                ADD     R4.y.4F@R127.w,    R4.y.3F@R125.y, C5.y
>            ]  
>            [  
>                ADD     R4.z.4F@R127.x,    R4.z.3F@R125.x, C5.z
>                MULADD     R5.z.3F@R124.w,    C8.y, R2.y.1||FP@R2.y, R5.z.2F@R124.w
>            ]  
>            [  
>                MULADD     R5.y.3F@R124.x,    C7.y, R2.y.1||FP@R2.y, R5.y.2F@R124.x
>                MULADD     R5.x.3F@R124.y,    C6.y, R2.y.1||FP@R2.y, R5.x.2F@R124.y
>            ]  
>            [  
>                MULADD     R6.x.5F@R125.x,    R4.x.5F@R126.w, -C0.w, C0.x
>                MULADD     R6.y.3F@R127.y,    R4.y.4F@R127.w, -C0.w, C0.y
>            ]  
>            [  
>                MULADD     R6.z.3F@R124.z,    R4.z.4F@R127.x, -C0.w, C0.z
>                MULADD     R5.z.4F@R125.w,    C8.z, R2.z.1||FP@R2.z, R5.z.3F@R124.w
>            ]  
>            [  
>                MULADD     R5.x.4F@R125.z,    C6.z, R2.z.1||FP@R2.z, R5.x.3F@R124.y
>                MULADD     R5.y.4F@R124.w,    C7.z, R2.z.1||FP@R2.z, R5.y.3F@R124.x
>            ]  
>            [  
>                4S  DOT4     R7.x.2F@R124.x,    t12F@R125.x, t12F@R125.x
>                4S  DOT4     __,    t13F@R127.y, t13F@R127.y
>                4S  DOT4     __,    t14F@R124.z, t14F@R124.z
>                4S  DOT4     __,    0|00000000, 0|00000000
>            ]  
>            [  
>                4S  DOT4     __,    t15F@R125.z, t15F@R125.z
>                4S  DOT4     R6.x.2F@R124.y,    t16F@R124.w, t16F@R124.w
>                4S  DOT4     __,    t17F@R125.w, t17F@R125.w
>                4S  DOT4     __,    0|00000000, 0|00000000
>                RECIPSQRT_CLAMPED     R11.x.3F@R126.y,    |R7.x.2F@R124.x|
>            ]  
>            [  
>                MUL     R6.z.4F@R125.y,    R6.z.3F@R124.z, R11.x.3F@R126.y
>                RECIPSQRT_CLAMPED     R11.x.2F@R124.z,    |R6.x.2F@R124.y|
>            ]  
>            [  
>                MUL     R5.x.5F@R124.x,    R5.x.4F@R125.z, R11.x.2F@R124.z
>                MUL     R6.x.6F@R124.y,    R6.x.5F@R125.x, R11.x.3F@R126.y
>                MUL     R5.y.5F@R124.z,    R5.y.4F@R124.w, R11.x.2F@R124.z
>                MUL     R6.y.4F@R124.w,    R6.y.3F@R127.y, R11.x.3F@R126.y
>                MUL     R5.z.5F@R126.x,    R5.z.4F@R125.w, R11.x.2F@R124.z
>            ]  
>            [  
>                4S  DOT4     __,    t6F@R124.x, t7F@R124.y
>                4S  DOT4     R6.x.7F@R124.y,    t8F@R124.z, t9F@R124.w
>                4S  DOT4     __,    t10F@R126.x, t11F@R125.y
>                4S  DOT4     __,    0|00000000, 0|00000000
>            ]  
>            [  
>                MULADD     R6.x.9F@R124.y,    -R6.x.7F@R124.y, R6.x.7F@R124.y, 1|3f800000
>            ]  
>            [  
>                SQRT_IEEE     R11.x.4F@R124.y,    R6.x.9F@R124.y
>            ]  
>            [  
>                MUL     R6.x.11F@R124.y,    C1.z, R11.x.4F@R124.y
>            ]  
>            [  
>                MUL     R5.x.6F@R124.x,    R5.x.5F@R124.x, R6.x.11F@R124.y
>            ]  
>            [  
>                ADD     R4.x.6F@R125.x,    R4.x.5F@R126.w, -R5.x.6F@R124.x
>                MUL     R5.y.6F@R124.z,    R5.y.5F@R124.z, R6.x.11F@R124.y
>                MUL     R5.w.3F@R124.w,    C14.w, R1.x.1F@R0.x
>            ]  
>            [  
>                MUL     R5.w.1F@R124.x,    C10.w, R4.x.6F@R125.x
>                MUL     R5.z.6F@R124.y,    R5.z.5F@R126.x, R6.x.11F@R124.y
>                MUL     R5.z.7F@R124.z,    C10.z, R4.x.6F@R125.x
>                MULADD     R5.w.4F@R124.w,    C15.w, R1.y.1F@R0.y, R5.w.3F@R124.w
>                ADD     R4.y.5F@R126.z,    R4.y.4F@R127.w, -R5.y.6F@R124.z
>            ]  
>            [  
>                MULADD     R5.w.2F@R124.x,    C11.w, R4.y.5F@R126.z, R5.w.1F@R124.x
>                ADD     R4.z.5F@R126.y,    R4.z.4F@R127.x, -R5.z.6F@R124.y
>                MULADD     R5.w.5F@R124.w,    C16.w, R1.z.1F@R127.z, R5.w.4F@R124.w
>            ]  
>            [  
>                MULADD     R4.w.4F@R124.x,    C12.w, R4.z.5F@R126.y, R5.w.2F@R124.x
>                MULADD     R5.z.8F@R124.z,    C11.z, R4.y.5F@R126.z, R5.z.7F@R124.z
>            ]  
>            [  
>                ADD     R4.w.5F@R124.x,    R4.w.4F@R124.x, C13.w
>                ADD     R4.w.6F@R124.w,    R5.w.5F@R124.w, C17.w
>            ]  
>            [  
>                MULADD     R4.z.6F@R125.y,    C12.z, R4.z.5F@R126.y, R5.z.8F@R124.z
>                MUL     R5.z.9F@R124.y,    C14.z, R1.x.1F@R0.x
>            ]  
>            [  
>                MULADD     R5.z.10F@R124.y,    C15.z, R1.y.1F@R0.y, R5.z.9F@R124.y
>                CNDE     R4.w.7||FP@R0.w,    C1.z, R4.w.6F@R124.w, R4.w.5F@R124.x
>            ]  
>            [  
>                MULADD     R5.z.11F@R124.x,    C16.z, R1.z.1F@R127.z, R5.z.10F@R124.y
>                ADD     R4.z.7F@R125.y,    R4.z.6F@R125.y, C13.z
>            ]  
>            [  
>                ADD     R4.z.8F@R126.x,    R5.z.11F@R124.x, C17.z
>                MUL     R5.x.9F@R124.y,    C14.x, R1.x.1F@R0.x
>                MUL     R5.y.9F@R124.w,    C14.y, R1.x.1F@R0.x
>                RECIP_CLAMPED     R11.x.6F@R124.z,    R4.w.7||FP@R0.w
>            ]  
>            [  
>                MUL     R5.x.7F@R124.x,    C10.x, R4.x.6F@R125.x
>                MUL     R6.x.14F@R125.z,    C1.x, R11.x.6F@R124.z
>                MUL     R5.y.7F@R125.x,    C10.y, R4.x.6F@R125.x
>            ]  
>            [  
>                MULADD     R5.x.10F@R126.x,    C15.x, R1.y.1F@R0.y, R5.x.9F@R124.y
>                MULADD     R5.y.10F@R124.z,    C15.y, R1.y.1F@R0.y, R5.y.9F@R124.w
>                CNDE     R4.z.9F@R127.x,    C1.z, R4.z.8F@R126.x, R4.z.7F@R125.y
>            ]  
>            [  
>                MULADD     R5.x.8F@R124.x,    C11.x, R4.y.5F@R126.z, R5.x.7F@R124.x
>                MOV_sat     R6.x.15F@R124.y,    R6.x.14F@R125.z
>                MULADD     R5.y.8F@R124.w,    C11.y, R4.y.5F@R126.z, R5.y.7F@R125.x
>                ADD     R11.x.7F@R125.x,    -C1.y, 1|3f800000
>            ]  
>            [  
>                MULADD     R5.y.11F@R124.y,    C16.y, R1.z.1F@R127.z, R5.y.10F@R124.z
>                MULADD     R4.y.6F@R124.z,    C12.y, R4.z.5F@R126.y, R5.y.8F@R124.w
>                ADD     R6.x.16F@R125.z,    R4.z.9F@R127.x, R6.x.15F@R124.y
>            ]  
>            [  
>                MULADD     R4.x.7F@R124.x,    C12.x, R4.z.5F@R126.y, R5.x.8F@R124.x
>                MAX     R4.x.11F@R125.y,    R6.x.16F@R125.z, -R4.w.7||FP@R0.w
>                MUL     R11.x.8F@R125.z,    R11.x.7F@R125.x, R6.x.16F@R125.z
>                MULADD     R5.x.11F@R124.w,    C16.x, R1.z.1F@R127.z, R5.x.10F@R126.x
>            ]  
>            [  
>                ADD     R4.x.8F@R124.x,    R4.x.7F@R124.x, C13.x
>                ADD     R4.y.7F@R124.y,    R4.y.6F@R124.z, C13.y
>                ADD     R4.x.9F@R124.z,    R5.x.11F@R124.w, C17.x
>                ADD     R4.y.8F@R125.x,    R5.y.11F@R124.y, C17.y
>            ]  
>            [  
>                CNDE     R4.x.10||FP@R0.x,    C1.z, R4.x.9F@R124.z, R4.x.8F@R124.x
>                CNDE     R4.y.9||FP@R0.y,    C1.z, R4.y.8F@R125.x, R4.y.7F@R124.y
>                MULADD     R4.x.12||FP@R0.z,    C1.y, R4.x.11F@R125.y, R11.x.8F@R125.z
>            ]  
>        >  
>        EXPORT  POS   60     t2||FP@R0.x, t3||FP@R0.y, t4||FP@R0.z, t5||FP@R0.w
>        EXPORT  PARAM 0     __, __, __, __
>    } end BB_0  
>}  
>
>===== SHADER #94 OPT ======================================= VS/RV710/R700 =====
>===== 214 dw ===== 3 gprs ===== 1 stack ========================================
>0000  00000000 89800000 CALL_FS @0
>0002  80000006 a18c0000 ALU 100 @12 KC0[CB0:0-31]
> 0012  80000001 00000c90     1      x: MOV                R0.x,  R1.x
> 0014  00000882 0f800090     2      x: MUL                T0.x,  KC0[2].z, R0.x
> 0016  00000482 2f800090            y: MUL                T0.y,  KC0[2].y, R0.x
> 0018  00000082 4f800090            z: MUL                T0.z,  KC0[2].x, R0.x
> 0020  80000401 20000c90            t: MOV                R0.y,  R1.y
> 0022  00800883 0f82007c     3      x: MULADD             T0.x,  KC0[3].z, R0.y, T0.x
> 0024  00800483 2f82047c            y: MULADD             T0.y,  KC0[3].y, R0.y, T0.y
> 0026  00800083 4f82087c            z: MULADD             T0.z,  KC0[3].x, R0.y, T0.z
> 0028  80000801 4fe40c90            t: MOV                T3.z,  R1.z                   SCL_122
> 0030  010fe484 2fa2047c     4      y: MULADD             T1.y,  KC0[4].y, T3.z, T0.y
> 0032  010fe084 4f82087c            z: MULADD             T0.z,  KC0[4].x, T3.z, T0.z
> 0034  80004088 6f800090            w: MUL                T0.w,  KC0[8].x, R2.x
> 0036  010fe884 0fa2007c     5      x: MULADD             T1.x,  KC0[4].z, T3.z, T0.x
> 0038  80004086 2f800090            y: MUL                T0.y,  KC0[6].x, R2.x
> 0040  00004087 0f800090     6      x: MUL                T0.x,  KC0[7].x, R2.x
> 0042  0010a87c 6fc00010            w: ADD                T2.w,  T0.z, KC0[5].x
> 0044  8090a47d 6fe00010            t: ADD                T3.w,  T1.y, KC0[5].y
> 0046  0110a07d 0fe00010     7      x: ADD                T3.x,  T1.x, KC0[5].z
> 0048  80804488 6f820c7c            w: MULADD             T0.w,  KC0[8].y, R2.y, T0.w
> 0050  00804487 0f82007c     8      x: MULADD             T0.x,  KC0[7].y, R2.y, T0.x
> 0052  80804486 2f82047c            y: MULADD             T0.y,  KC0[6].y, R2.y, T0.y
> 0054  03900c7e 0fa20080     9      x: MULADD             T1.x,  T2.w, -KC0[0].w, KC0[0].x
> 0056  83900c7f 2fea0480            y: MULADD             T3.y,  T3.w, -KC0[0].w, KC0[0].y  VEC_120
> 0058  0390007f 4f820880    10      z: MULADD             T0.z,  T3.x, -KC0[0].w, KC0[0].z
> 0060  81004888 6fa20c7c            w: MULADD             T1.w,  KC0[8].z, R2.z, T0.w
> 0062  01004886 4fa2047c    11      z: MULADD             T1.z,  KC0[6].z, R2.z, T0.y
> 0064  81004887 6f82007c            w: MULADD             T0.w,  KC0[7].z, R2.z, T0.x
> 0066  000fa07d 0f802810    12      x: DOT4               T0.x,  T1.x, T1.x
> 0068  008fe47f 20002800            y: DOT4               __.y,  T3.y, T3.y
> 0070  010f887c 40002800            z: DOT4               __.z,  T0.z, T0.z
> 0072  801f00f8 60002800            w: DOT4               __.w,  0, 0
> 0074  010fa87d 00002800    13      x: DOT4               __.x,  T1.z, T1.z
> 0076  018f8c7c 2f802810            y: DOT4               T0.y,  T0.w, T0.w
> 0078  018fac7d 40082800            z: DOT4               __.z,  T1.w, T1.w             VEC_120
> 0080  001f00f8 60002800            w: DOT4               __.w,  0, 0
> 0082  8000007c 2fc03391            t: RECIPSQRT_CLAMPED  T2.y,  |T0.x|
> 0084  008fc87c 2fa00090    14      y: MUL                T1.y,  T0.z, T2.y
> 0086  8000047c 4f803391            t: RECIPSQRT_CLAMPED  T0.z,  |T0.y|
> 0088  010f887d 0f800090    15      x: MUL                T0.x,  T1.z, T0.z
> 0090  008fc07d 2f800090            y: MUL                T0.y,  T1.x, T2.y
> 0092  010f8c7c 4f800090            z: MUL                T0.z,  T0.w, T0.z
> 0094  008fc47f 6f800090            w: MUL                T0.w,  T3.y, T2.y
> 0096  810f8c7d 0fc00090            t: MUL                T2.x,  T1.w, T0.z
> 0098  008f807c 00002800    16      x: DOT4               __.x,  T0.x, T0.y
> 0100  018f887c 2f802810            y: DOT4               T0.y,  T0.z, T0.w
> 0102  008fa07e 40082800            z: DOT4               __.z,  T2.x, T1.y             VEC_120
> 0104  801f00f8 60002800            w: DOT4               __.w,  0, 0
> 0106  808f947c 2f8200f9    17      y: MULADD             T0.y,  -T0.y, T0.y, 1.0
> 0108  8000047c 2f803510    18      t: SQRT_IEEE          T0.y,  T0.y
> 0110  808f8881 2f800090    19      y: MUL                T0.y,  KC0[1].z, T0.y
> 0112  808f807c 0f800090    20      x: MUL                T0.x,  T0.x, T0.y
> 0114  020f8c7e 0fa00010    21      x: ADD                T1.x,  T2.w, -T0.x
> 0116  008f887c 4f800090            z: MUL                T0.z,  T0.z, T0.y
> 0118  80000c8e 6f840090            w: MUL                T0.w,  KC0[14].w, R0.x        VEC_021
> 0120  000fac8a 0f800090    22      x: MUL                T0.x,  KC0[10].w, T1.x
> 0122  008f807e 2f800090            y: MUL                T0.y,  T2.x, T0.y
> 0124  000fa88a 4f800090            z: MUL                T0.z,  KC0[10].z, T1.x
> 0126  00800c8f 6f860c7c            w: MULADD             T0.w,  KC0[15].w, R0.y, T0.w  VEC_021
> 0128  830f8c7f 4fc00010            t: ADD                T2.z,  T3.w, -T0.z
> 0130  010fcc8b 0f86007c    23      x: MULADD             T0.x,  KC0[11].w, T2.z, T0.x  VEC_021
> 0132  028f807f 2fc00010            y: ADD                T2.y,  T3.x, -T0.y
> 0134  810fec90 6f820c7c            w: MULADD             T0.w,  KC0[16].w, T3.z, T0.w
> 0136  008fcc8c 0f82007c    24      x: MULADD             T0.x,  KC0[12].w, T2.y, T0.x
> 0138  810fc88b 4f82087c            z: MULADD             T0.z,  KC0[11].z, T2.z, T0.z
> 0140  0191a07c 0f800010    25      x: ADD                T0.x,  T0.x, KC0[13].w
> 0142  81922c7c 6f800010            w: ADD                T0.w,  T0.w, KC0[17].w
> 0144  008fc88c 2fa2087c    26      y: MULADD             T1.y,  KC0[12].z, T2.y, T0.z
> 0146  8000088e 2f800090            t: MUL                T0.y,  KC0[14].z, R0.x
> 0148  0080088f 2f82047c    27      y: MULADD             T0.y,  KC0[15].z, R0.y, T0.y
> 0150  818f8881 6003007c            w: CNDE               R0.w,  KC0[1].z, T0.w, T0.x
> 0152  010fe890 0f82047c    28      x: MULADD             T0.x,  KC0[16].z, T3.z, T0.y
> 0154  8111a47d 2fa00010            y: ADD                T1.y,  T1.y, KC0[13].z
> 0156  0112207c 0fc00010    29      x: ADD                T2.x,  T0.x, KC0[17].z
> 0158  0000008e 2f800090            y: MUL                T0.y,  KC0[14].x, R0.x
> 0160  0000048e 6f800090            w: MUL                T0.w,  KC0[14].y, R0.x
> 0162  80000c00 4f803210            t: RECIP_CLAMPED      T0.z,  R0.w
> 0164  000fa08a 0f800090    30      x: MUL                T0.x,  KC0[10].x, T1.x
> 0166  010f8081 4fa00090            z: MUL                T1.z,  KC0[1].x, T0.z
> 0168  800fa48a 0fa00090            t: MUL                T1.x,  KC0[10].y, T1.x
> 0170  0080008f 0fca047c    31      x: MULADD             T2.x,  KC0[15].x, R0.y, T0.y  VEC_120
> 0172  0080048f 4f860c7c            z: MULADD             T0.z,  KC0[15].y, R0.y, T0.w  VEC_021
> 0174  800fc881 0fef047d            t: CNDE               T3.x,  KC0[1].z, T2.x, T1.y   SCL_221
> 0176  010fc08b 0f82007c    32      x: MULADD             T0.x,  KC0[11].x, T2.z, T0.x
> 0178  0000087d af800c90            y: MOV_sat            T0.y,  T1.z
> 0180  010fc48b 6f86007d            w: MULADD             T0.w,  KC0[11].y, T2.z, T1.x  VEC_021
> 0182  801f3481 0fa00010            t: ADD                T1.x,  -KC0[1].y, 1.0
> 0184  010fe490 2f82087c    33      y: MULADD             T0.y,  KC0[16].y, T3.z, T0.z
> 0186  008fc48c 4f820c7c            z: MULADD             T0.z,  KC0[12].y, T2.y, T0.w
> 0188  808f807f 4fa40010            t: ADD                T1.z,  T3.x, T0.y             SCL_122
> 0190  008fc08c 0f82007c    34      x: MULADD             T0.x,  KC0[12].x, T2.y, T0.x
> 0192  0380087d 2fa00190            y: MAX                T1.y,  T1.z, -R0.w
> 0194  010fa07d 4fa00090            z: MUL                T1.z,  T1.x, T1.z
> 0196  810fe090 6f86007e            w: MULADD             T0.w,  KC0[16].x, T3.z, T2.x  VEC_021
> 0198  0011a07c 0f800010    35      x: ADD                T0.x,  T0.x, KC0[13].x
> 0200  0091a87c 2f800010            y: ADD                T0.y,  T0.z, KC0[13].y
> 0202  00122c7c 4f800010            z: ADD                T0.z,  T0.w, KC0[17].x
> 0204  8092247c 0fa00010            t: ADD                T1.x,  T0.y, KC0[17].y
> 0206  010f8881 0003007c    36      x: CNDE               R0.x,  KC0[1].z, T0.z, T0.x
> 0208  000fa881 2003047c            y: CNDE               R0.y,  KC0[1].z, T1.x, T0.y
> 0210  808fa481 4002087d            z: MULADD             R0.z,  KC0[1].y, T1.y, T1.z
>0004  c000203c 94000688 EXPORT_DONE        POS   60    R0.xyzw
>0006  c0004000 94000fff EXPORT_DONE        PARAM 0     R0.____
>0008  0000006a a0000000 ALU 1 @212
> 0212  80000000 00000d00    37      x: ALU_NOP            __.x
>0010  00000000 80200000 NOP @0  EOP
>===== SHADER_END ===============================================================
>
>sb: dry run shader 94: optimized bytecode is not used
>sb: processing shader 94 done ( 104.756 ms ).
>src stats: dw:288, gpr:12, stk:1, alu groups:49, alu clauses: 4, alu:135, fetch:0, fetch clauses:0, cf:5
>opt stats: dw:288, gpr:3, stk:1, alu groups:37, alu clauses: 2, alu:101, fetch:0, fetch clauses:0, cf:4
>diff: dw:0%, gpr:-75%, stk:0%, alu groups:-24%, alu clauses: -50%, alu:-25%, fetch:0%, fetch clauses:0%, cf:-20%
>Rendered 629 frames in 15.9118 secs, average of 39.5305 fps

Actions: View

Attachments on bug 93292: 120410 | 121137 | 121138 | 121139

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.