shithub: dav1d

Download patch

ref: 7cb756ea5131285637259d264925c608a24d3443
parent: 121b3bb34bd20d828bc880dfa6ab6b37b811b9b5
author: Ronald S. Bultje <rsbultje@gmail.com>
date: Mon Dec 17 15:06:19 EST 2018

itx: replace qw_ constant prefix with pw_

Makes it more consistent with the rest of the project.

--- a/src/x86/itx_ssse3.asm
+++ b/src/x86/itx_ssse3.asm
@@ -31,20 +31,20 @@
 
 deint_shuf: db  0,  1,  4,  5,  8,  9, 12, 13,  2,  3,  6,  7, 10, 11, 14, 15
 
-qw_2896x8:      times 8 dw  2896*8
-qw_1567_m3784:  times 4 dw  1567, -3784
-qw_3784_1567:   times 4 dw  3784,  1567
+pw_2896x8:      times 8 dw  2896*8
+pw_1567_m3784:  times 4 dw  1567, -3784
+pw_3784_1567:   times 4 dw  3784,  1567
 
-qw_1321_3803:   times 4 dw  1321,  3803
-qw_2482_m1321:  times 4 dw  2482, -1321
-qw_3344_2482:   times 4 dw  3344,  2482
-qw_3344_m3803:  times 4 dw  3344, -3803
-qw_m6688_m3803: times 4 dw -6688, -3803
-qw_3344x8:      times 8 dw  3344*8
-qw_5793x4:      times 8 dw  5793*4
+pw_1321_3803:   times 4 dw  1321,  3803
+pw_2482_m1321:  times 4 dw  2482, -1321
+pw_3344_2482:   times 4 dw  3344,  2482
+pw_3344_m3803:  times 4 dw  3344, -3803
+pw_m6688_m3803: times 4 dw -6688, -3803
+pw_3344x8:      times 8 dw  3344*8
+pw_5793x4:      times 8 dw  5793*4
 
 pd_2048:        times 4 dd  2048
-qw_2048:        times 8 dw  2048
+pw_2048:        times 8 dw  2048
 
 iadst4_dconly1a: times 2 dw 10568, 19856, 26752, 30424
 iadst4_dconly1b: times 2 dw 30424, 26752, 19856, 10568
@@ -57,7 +57,7 @@
 
 %macro ITX4_END 4-5 2048 ; row[1-4], rnd
 %if %5
-    mova                 m2, [qw_%5]
+    mova                 m2, [pw_%5]
     pmulhrsw             m0, m2
     pmulhrsw             m1, m2
 %endif
@@ -107,11 +107,11 @@
     pmaddwd              m%2, m%4, m%1
     pmaddwd              m%1, m%5
 %elif %6 & 1
-    pmaddwd              m%2, m%1, [qw_%5_%4]
-    pmaddwd              m%1, [qw_%4_m%5]
+    pmaddwd              m%2, m%1, [pw_%5_%4]
+    pmaddwd              m%1, [pw_%4_m%5]
 %else
-    pmaddwd              m%2, m%1, [qw_%4_m%5]
-    pmaddwd              m%1, [qw_%5_%4]
+    pmaddwd              m%2, m%1, [pw_%4_m%5]
+    pmaddwd              m%1, [pw_%5_%4]
 %endif
     paddd                m%2, m%3
     paddd                m%1, m%3
@@ -120,7 +120,7 @@
     packssdw             m%1, m%2
 %endmacro
 
-%macro IDCT4_1D_PACKED 0-1   ;qw_2896x8
+%macro IDCT4_1D_PACKED 0-1   ;pw_2896x8
     punpckhwd            m2, m0, m1           ;unpacked in1 in3
     psubw                m3, m0, m1
     paddw                m0, m1
@@ -132,7 +132,7 @@
 %if %0 == 1
     pmulhrsw             m0, m%1
 %else
-    pmulhrsw             m0, [qw_2896x8]     ;high: t1 ;low: t0
+    pmulhrsw             m0, [pw_2896x8]     ;high: t1 ;low: t0
 %endif
 
     psubsw               m1, m0, m2          ;high: out2 ;low: out3
@@ -146,14 +146,14 @@
     punpckhqdq           m1, m1                    ;
     paddw                m1, m0                    ;low: in0 - in2 + in3
 
-    pmaddwd              m0, m2, [qw_1321_3803]    ;1321 * in0 + 3803 * in2
-    pmaddwd              m2, [qw_2482_m1321]       ;2482 * in0 - 1321 * in2
-    pmaddwd              m4, m3, [qw_3344_2482]    ;3344 * in1 + 2482 * in3
-    pmaddwd              m5, m3, [qw_3344_m3803]   ;3344 * in1 - 3803 * in3
+    pmaddwd              m0, m2, [pw_1321_3803]    ;1321 * in0 + 3803 * in2
+    pmaddwd              m2, [pw_2482_m1321]       ;2482 * in0 - 1321 * in2
+    pmaddwd              m4, m3, [pw_3344_2482]    ;3344 * in1 + 2482 * in3
+    pmaddwd              m5, m3, [pw_3344_m3803]   ;3344 * in1 - 3803 * in3
     paddd                m4, m0                    ;t0 + t3
 
-    pmaddwd              m3, [qw_m6688_m3803]      ;-2 * 3344 * in1 - 3803 * in3
-    pmulhrsw             m1, [qw_3344x8]           ;low: out2
+    pmaddwd              m3, [pw_m6688_m3803]      ;-2 * 3344 * in1 - 3803 * in3
+    pmulhrsw             m1, [pw_3344x8]           ;low: out2
     mova                 m0, [pd_2048]
     paddd                m2, m0
     paddd                m0, m4                    ;t0 + t3 + 2048
@@ -188,10 +188,10 @@
 %macro INV_TXFM_4X4_FN 2-3 -1 ; type1, type2, fast_thresh
     INV_TXFM_FN          %1, %2, %3, 4x4
 %ifidn %1_%2, dct_identity
-    mova                 m0, [qw_2896x8]
+    mova                 m0, [pw_2896x8]
     pmulhrsw             m0, [coeffq]
     paddw                m0, m0
-    pmulhrsw             m0, [qw_5793x4]
+    pmulhrsw             m0, [pw_5793x4]
     punpcklwd            m0, m0
     punpckhdq            m1, m0, m0
     punpckldq            m0, m0
@@ -205,8 +205,8 @@
     punpcklwd            m0, m1
     punpcklqdq           m0, m0
     paddw                m0, m0
-    pmulhrsw             m0, [qw_5793x4]
-    pmulhrsw             m0, [qw_2896x8]
+    pmulhrsw             m0, [pw_5793x4]
+    pmulhrsw             m0, [pw_2896x8]
     mova                 m1, m0
     call m(iadst_4x4_internal).end
     RET
@@ -214,7 +214,7 @@
     pshuflw              m0, [coeffq], q0000
     punpcklqdq           m0, m0
 %ifidn %1, dct
-    mova                 m1, [qw_2896x8]
+    mova                 m1, [pw_2896x8]
     pmulhrsw             m0, m1
 %elifidn %1, adst
     pmulhrsw             m0, [iadst4_dconly1a]
@@ -224,7 +224,7 @@
     mov            [coeffq], eobd                ;0
 %ifidn %2, dct
 %ifnidn %1, dct
-    pmulhrsw             m0, [qw_2896x8]
+    pmulhrsw             m0, [pw_2896x8]
 %else
     pmulhrsw             m0, m1
 %endif
@@ -330,7 +330,7 @@
 cglobal iidentity_4x4_internal, 0, 0, 6, dst, stride, coeff, eob, tx2
     mova                 m0, [coeffq+16*0]
     mova                 m1, [coeffq+16*1]
-    mova                 m2, [qw_5793x4]
+    mova                 m2, [pw_5793x4]
     paddw                m0, m0
     paddw                m1, m1
     pmulhrsw             m0, m2
@@ -343,7 +343,7 @@
     jmp                tx2q
 
 .pass2:
-    mova                 m2, [qw_5793x4]
+    mova                 m2, [pw_5793x4]
     paddw                m0, m0
     paddw                m1, m1
     pmulhrsw             m0, m2