| /external/libvpx/vp8/common/arm/neon/ |
| buildintrapredictorsmby_neon.asm | 45 vdup.u8 q0, r12
126 vdup.u8 q0, r5
129 vst1.u8 {q0}, [r1]!
130 vst1.u8 {q0}, [r1]!
131 vst1.u8 {q0}, [r1]!
132 vst1.u8 {q0}, [r1]!
133 vst1.u8 {q0}, [r1]!
134 vst1.u8 {q0}, [r1]!
135 vst1.u8 {q0}, [r1]!
136 vst1.u8 {q0}, [r1]
[all...] |
| loopfiltersimplehorizontaledge_neon.asm | 36 vld1.u8 {q0}, [r12]! ; 0x80
37 vld1.u8 {q7}, [r0], r1 ; q0
42 vabd.u8 q15, q6, q7 ; abs(p0 - q0)
44 vqadd.u8 q15, q15, q15 ; abs(p0 - q0) * 2
46 vqadd.u8 q15, q15, q14 ; abs(p0 - q0) * 2 + abs(p1 - q1) / 2
49 veor q7, q7, q0 ; qs0: q0 offset to convert to a signed value
50 veor q6, q6, q0 ; ps0: p0 offset to convert to a signed value
51 veor q5, q5, q0 ; ps1: p1 offset to convert to a signed value
52 veor q8, q8, q0 ; qs1: q1 offset to convert to a signed valu
[all...] |
| recon16x16mb_neon.asm | 35 vmovl.u8 q0, d24 ;modify Pred data from 8 bits to 16 bits
50 vadd.s16 q0, q0, q8 ;add Diff data and Pred data together
59 vqmovun.s16 d0, q0 ;CLAMP() saturation
65 vst1.u8 {q0}, [r2], r3 ;store result
86 vmovl.u8 q0, d24 ;modify Pred data from 8 bits to 16 bits
97 vadd.s16 q0, q0, q8 ;add Diff data and Pred data together
105 vqmovun.s16 d0, q0 ;CLAMP() saturation
|
| loopfilter_neon.asm | 47 vld1.u8 {q7}, [r2], r1 ; q0
86 vld1.u8 {d14}, [r3], r1 ; q0
98 vld1.u8 {d15}, [r12], r1 ; q0
299 ; q0 flimit
306 ; q7 q0
317 vabd.u8 q14, q8, q7 ; abs(q1 - q0)
320 vabd.u8 q9, q6, q7 ; abs(p0 - q0)
329 vcgt.u8 q14, q14, q2 ; (abs(q1 - q0) > thresh)*-1
332 vadd.u8 q0, q0, q0 ; flimit *
[all...] |
| loopfiltersimpleverticaledge_neon.asm | 44 vld1.u8 {q0}, [r12]! ; 0x80
57 ;vswp q4, q5 ; p1:q3, p0:q5, q0:q4, q1:q6
62 vabd.u8 q15, q5, q4 ; abs(p0 - q0)
64 vqadd.u8 q15, q15, q15 ; abs(p0 - q0) * 2
66 vqadd.u8 q15, q15, q14 ; abs(p0 - q0) * 2 + abs(p1 - q1) / 2
68 veor q4, q4, q0 ; qs0: q0 offset to convert to a signed value
69 veor q5, q5, q0 ; ps0: p0 offset to convert to a signed value
70 veor q3, q3, q0 ; ps1: p1 offset to convert to a signed value
71 veor q6, q6, q0 ; qs1: q1 offset to convert to a signed valu
[all...] |
| mbloopfilter_neon.asm | 47 vld1.u8 {q7}, [r0], r1 ; q0
97 vld1.u8 {d14}, [r0], r1 ; q0
98 vld1.u8 {d15}, [r3], r1 ; q0
369 ; q7 q0
381 vabd.u8 q14, q8, q7 ; abs(q1 - q0)
383 vabd.u8 q0, q10, q9 ; abs(q3 - q2)
387 vmax.u8 q3, q3, q0
390 vabd.u8 q12, q6, q7 ; abs(p0 - q0)
394 vcgt.u8 q14, q14, q2 ; (abs(q1 - q0) > thresh) * -1
399 vld1.u8 {q0}, [r12]
[all...] |
| /bionic/libm/src/ |
| k_rem_pio2.c | 108 * q0 the corresponding exponent of q[0]. Note that the
109 * exponent for q[i] would be q0-24*i.
158 int32_t jz,jx,jv,jp,jk,carry,n,iq[20],i,j,k,m,q0,ih; local
165 /* determine jx,jv,q0, note that 3>q0 */
168 q0 = e0-24*(jv+1);
189 z = scalbn(z,q0); /* actual value of z */
194 if(q0>0) { /* need iq[jz-1] to determine n */
195 i = (iq[jz-1]>>(24-q0)); n += i;
196 iq[jz-1] -= i<<(24-q0);
[all...] |
| k_rem_pio2f.c | 50 int32_t jz,jx,jv,jp,jk,carry,n,iq[20],i,j,k,m,q0,ih; local
57 /* determine jx,jv,q0, note that 3>q0 */
60 q0 = e0-8*(jv+1);
81 z = scalbnf(z,q0); /* actual value of z */
86 if(q0>0) { /* need iq[jz-1] to determine n */
87 i = (iq[jz-1]>>(8-q0)); n += i;
88 iq[jz-1] -= i<<(8-q0);
89 ih = iq[jz-1]>>(7-q0);
91 else if(q0==0) ih = iq[jz-1]>>7
[all...] |
| e_jnf.c | 110 float q0,q1,h,tmp; int32_t k,m; local
112 q0 = w; z = w+h; q1 = w*z - (float)1.0; k=1;
115 tmp = z*q1 - q0;
116 q0 = q1;
|
| e_jn.c | 157 double q0,q1,h,tmp; int32_t k,m; local
159 q0 = w; z = w+h; q1 = w*z - 1.0; k=1;
162 tmp = z*q1 - q0;
163 q0 = q1;
|
| /external/fdlibm/ |
| k_rem_pio2.c | 104 * q0 the corresponding exponent of q[0]. Note that the
105 * exponent for q[i] would be q0-24*i.
170 int jz,jx,jv,jp,jk,carry,n,iq[20],i,j,k,m,q0,ih; local
177 /* determine jx,jv,q0, note that 3>q0 */
180 q0 = e0-24*(jv+1);
201 z = ieee_scalbn(z,q0); /* actual value of z */
206 if(q0>0) { /* need iq[jz-1] to determine n */
207 i = (iq[jz-1]>>(24-q0)); n += i;
208 iq[jz-1] -= i<<(24-q0);
[all...] |
| e_jn.c | 161 double q0,q1,h,tmp; int k,m; local
163 q0 = w; z = w+h; q1 = w*z - 1.0; k=1;
166 tmp = z*q1 - q0;
167 q0 = q1;
|
| /external/libvpx/vp8/encoder/arm/neon/ |
| fastquantizeb_neon.asm | 31 vld1.16 {q0, q1}, [r0] ;load z
34 vabs.s16 q4, q0 ;calculate x = abs(z)
52 vshr.s16 q2, q0, #15 ; sz
66 vld1.16 {q0, q1}, [r0] ;load rvsplus1_scan_order
91 vand q0, q0, q14 ;get all valid number from rvsplus1_scan_order array
94 vmax.u16 q0, q0, q1 ;find maximum value in q0, q1
96 vmovl.u16 q0, d
[all...] |
| vp8_memcpy_neon.asm | 30 vld1.8 {q0, q1}, [r1]! ;load src data
33 vst1.8 {q0, q1}, [r0]! ;copy to dst_ptr
59 vld1.8 {q0}, [r1]! ;load src data
61 vst1.8 {q0}, [r0]!
|
| vp8_subpixelvariance16x16s_neon.asm | 55 vext.8 q1, q0, q1, #1 ;construct src_ptr[1]
60 vrhadd.u8 q0, q0, q1 ;(src_ptr[0]+src_ptr[1])/round/shift right 1
70 vsubl.u8 q0, d4, d26
91 vpadal.s16 q8, q0 ;sum
107 vpaddl.s32 q0, q8 ;accumulate sum
137 vld1.u8 {q0}, [r0], r1 ;load src data
153 vrhadd.u8 q0, q0, q2
163 vsubl.u8 q0, d8, d1
[all...] |
| /external/libvpx/vp8/common/x86/ |
| loopfilter_mmx.asm | 68 movq mm4, [rsi] ; q0
69 movq mm0, mm4 ; q0
70 psubusb mm4, mm3 ; q0-=q1
71 psubusb mm3, mm0 ; q1-=q0
72 por mm4, mm3 ; abs(q0-q1)
118 movq mm3, [rsi] ; q0
119 psubusb mm5, mm3 ; p0-=q0
120 psubusb mm3, mm6 ; q0-=p0
121 por mm5, mm3 ; abs(p0 - q0)
122 paddusb mm5, mm5 ; abs(p0-q0)*
[all...] |
| loopfilter_sse2.asm | 22 movdqa xmm5, [rsi] ; q0
28 movlps xmm5, [rsi + rax] ; q0
54 movdqa xmm0, xmm5 ; q0
57 psubusb xmm5, xmm3 ; q0-=q1
58 psubusb xmm3, xmm0 ; q1-=q0
60 por xmm5, xmm3 ; abs(q0-q1)
128 movdqa xmm3, xmm0 ; q0
136 psubusb xmm5, xmm3 ; p0-=q0
139 psubusb xmm3, xmm6 ; q0-=p0
140 por xmm5, xmm3 ; abs(p0 - q0)
[all...] |
| /external/libvpx/vpx_scale/arm/neon/ |
| vp8_vpxyv12_copyframeyonly_neon.asm | 48 vld1.8 {q0, q1}, [r8]!
59 vst1.8 {q0, q1}, [r9]!
119 vmov q1, q0
130 vst1.8 {q0, q1}, [r5], lr
153 vld1.8 {q0, q1}, [r1]!
167 vst1.8 {q0, q1}, [r5]!
200 vld1.8 {q0}, [r1]!
208 vst1.8 {q0}, [r5], lr
210 vst1.8 {q0}, [r5], lr
212 vst1.8 {q0}, [r5], l
[all...] |
| vp8_vpxyv12_extendframeborders_neon.asm | 61 vmov q1, q0
72 vst1.8 {q0, q1}, [r5], lr
95 vld1.8 {q0, q1}, [r1]!
109 vst1.8 {q0, q1}, [r5]!
177 vst1.8 {q0}, [r5], lr
205 vld1.8 {q0, q1}, [r1]!
215 vst1.8 {q0, q1}, [r5]!
248 vld1.8 {q0}, [r1]!
256 vst1.8 {q0}, [r5], lr
258 vst1.8 {q0}, [r5], l
[all...] |
| vp8_vpxyv12_copyframe_func_neon.asm | 58 vld1.8 {q0, q1}, [r8]!
69 vst1.8 {q0, q1}, [r9]!
116 vld1.8 {q0, q1}, [r8]!
123 vst1.8 {q0, q1}, [r9]!
174 vld1.8 {q0}, [r8]!
179 vst1.8 {q0}, [r9]!
|
| vp8_vpxyv12_copysrcframe_func_neon.asm | 52 vld1.8 {q0, q1}, [r2]!
62 vst1.8 {q0, q1}, [r3]!
109 vld1.8 {q0, q1}, [r2]!
115 vst1.8 {q0, q1}, [r3]!
166 vld1.8 {q0, q1}, [r2]!
172 vst1.8 {q0, q1}, [r3]!
215 vld1.8 {q0, q1}, [r2]!
219 vst1.8 {q0, q1}, [r3]!
|
| /external/libvpx/vp8/common/arm/armv6/ |
| simpleloopfilter_v6.asm | 69 ldr r5, [src] ; q0
84 uqsub8 r10, r4, r5 ; p0 - q0
85 uqsub8 r11, r5, r4 ; q0 - p0
87 orr r10, r10, r11 ; abs(p0 - q0)
88 uqadd8 r10, r10, r10 ; abs(p0 - q0) * 2
90 uqadd8 r10, r10, r8 ; abs(p0 - q0)*2 + abs(p1 - q1)/2
102 eor r5, r5, r2 ; q0 offset to convert to a signed value
105 qsub8 r6, r5, r4 ; q0 - p0
106 qadd8 r3, r3, r6 ; += q0 - p0
108 qadd8 r3, r3, r6 ; += q0 - p
[all...] |
| loopfilter_v6.asm | 101 ldr r9, [src], pstep ; q0
110 uqsub8 r11, r12, r9 ; p0 - q0
111 uqsub8 r12, r9, r12 ; q0 - p0
114 orr r12, r11, r12 ; abs (p0-q0)
116 uqadd8 r12, r12, r12 ; abs (p0-q0) * 2
118 uqsub8 r7, r9, r10 ; q0 - q1
119 uqadd8 r12, r12, r6 ; abs (p0-q0)*2 + abs (p1-q1)/2
120 uqsub8 r6, r10, r9 ; q1 - q0
128 orr r6, r7, r6 ; abs (q1-q0)
168 ldr r9, [src], pstep ; q0
[all...] |
| /external/libvpx/vp8/common/ |
| loopfilter_filters.c | 28 uc p3, uc p2, uc p1, uc p0, uc q0, uc q1, uc q2, uc q3)
34 mask |= (abs(q1 - q0) > limit) * -1;
37 mask |= (abs(p0 - q0) * 2 + abs(p1 - q1) / 2 > flimit * 2 + limit) * -1;
43 static __inline signed char vp8_hevmask(signed char thresh, uc p1, uc p0, uc q0, uc q1)
47 hev |= (abs(q1 - q0) > thresh) * -1;
281 static __inline signed char vp8_simple_filter_mask(signed char limit, signed char flimit, uc p1, uc p0, uc q0, uc q1)
287 signed char mask = (abs(p0 - q0) * 2 + abs(p1 - q1) / 2 <= flimit * 2 + limit) * -1;
296 signed char q0 = (signed char) * oq0 ^ 0x80; local
301 vp8_filter = vp8_signed_char_clamp(vp8_filter + 3 * (q0 - p0));
307 u = vp8_signed_char_clamp(q0 - Filter1)
[all...] |
| /system/core/libpixelflinger/ |
| col32cb16blend_neon.S | 58 vshll.u8 q0, d0, #5 // shift up red and widen
90 vadd.u16 q8, q8, q0 // add src red to dst red0
97 vadd.u16 q13, q13, q0 // add src red to dst red1
|
