| /external/libvpx/vp8/encoder/arm/neon/ |
| variance_neon.asm | 29 vmov.i8 q8, #0 ;q8 - sum
49 vpadal.s16 q8, q11 ;calculate sum
55 vpadal.s16 q8, q12
58 vpadal.s16 q8, q13
61 vpadal.s16 q8, q14
68 vpaddl.s32 q0, q8 ;accumulate sum
102 vmov.i8 q8, #0 ;q8 - sum
119 vpadal.s16 q8, q11 ;calculate su
[all...] |
| vp8_mse16x16_neon.asm | 31 vmov.i8 q7, #0 ;q7, q8, q9, q10 - sse
32 vmov.i8 q8, #0
50 vmlal.s16 q8, d23, d23
57 vmlal.s16 q8, d27, d27
63 vadd.u32 q7, q7, q8
85 vmov.i8 q8, #0 ;q8 - sum
102 vpadal.s16 q8, q11
108 vpadal.s16 q8, q12
111 vpadal.s16 q8, q1
[all...] |
| vp8_subpixelvariance16x16s_neon.asm | 37 vmov.i8 q8, #0 ;q8 - sum
75 vpadal.s16 q8, q4 ;sum
81 vpadal.s16 q8, q5
84 vpadal.s16 q8, q6
87 vpadal.s16 q8, q7
91 vpadal.s16 q8, q0 ;sum
94 vpadal.s16 q8, q1
97 vpadal.s16 q8, q2
100 vpadal.s16 q8, q
[all...] |
| vp8_subpixelvariance8x8_neon.asm | 51 vmull.u8 q8, d6, d0
61 vmlal.u8 q8, d7, d1
69 vqrshrn.u16 d24, q8, #7
78 vmull.u8 q8, d6, d0
90 vmlal.u8 q8, d7, d1
96 vqrshrn.u16 d28, q8, #7
120 vmull.u8 q8, d29, d0
129 vmlal.u8 q8, d30, d1
138 vqrshrn.u16 d29, q8, #7
159 vmov.i8 q8, #0 ;q8 - su
[all...] |
| fastquantizeb_neon.asm | 56 vld1.s16 {q8, q9}, [r12] ;load quant_ptr [0-15]
63 vqdmulh.s16 q4, q8 ;y = ((Round + abs(z)) * Quant) >> 16
67 vceq.s16 q8, q8 ;set q8 to all 1
86 vtst.16 q14, q4, q8 ;now find eob
87 vtst.16 q15, q5, q8 ;non-zero element is set to all 1 in q4, q5
|
| fastfdct4x4_neon.asm | 56 vqdmulh.s16 q8, q4, d0[0]
60 vshr.s16 q8, q8, #1
62 vadd.s16 q8, q4, q8 ;d16:temp2 = ((c1 * x_c1)>>16) + c1; d17:temp2 = ((d1 * x_c1)>>16) + d1
86 vqdmulh.s16 q8, q4, d0[0]
90 vshr.s16 q8, q8, #1
92 vadd.s16 q8, q4, q8 ;d16:temp2 = ((c1 * x_c1)>>16) + c1; d17:temp2 = ((d1 * x_c1)>>16) + d
[all...] |
| vp8_subpixelvariance16x16_neon.asm | 66 vmull.u8 q8, d3, d0
89 vmlal.u8 q8, d3, d1 ;(src_ptr[0] * Filter[1])
97 vqrshrn.u16 d15, q8, #7
202 vmull.u8 q8, d29, d0
211 vmlal.u8 q8, d31, d1
222 vqrshrn.u16 d9, q8, #7
254 vmull.u8 q8, d3, d0
277 vmlal.u8 q8, d3, d1 ;(src_ptr[0] * Filter[1])
285 vqrshrn.u16 d15, q8, #7
329 vmull.u8 q8, d29, d
[all...] |
| subtract_neon.asm | 65 vsubl.u8 q8, d0, d2
74 vst1.16 {q8}, [r0]! ;store diff
115 vsubl.u8 q8, d0, d1
124 vst1.16 {q8}, [r0]! ;store diff
151 vsubl.u8 q8, d0, d1
160 vst1.16 {q8}, [r0]! ;store diff
|
| fastfdct8x4_neon.asm | 69 vqdmulh.s16 q8, q7, d0[1]
76 vshr.s16 q8, q8, #1
120 vqdmulh.s16 q8, q7, d0[1]
127 vshr.s16 q8, q8, #1
148 vclt.s16 q8, q4, #0
153 vsub.s16 q4, q4, q8
|
| vp8_memcpy_neon.asm | 38 vld1.8 {q8, q9}, [r1]!
41 vst1.8 {q8, q9}, [r0]!
|
| /system/core/libpixelflinger/ |
| col32cb16blend_neon.S | 70 // q8 = dst red0
78 vshr.u16 q8, q10, #11 // shift dst red0 to low 5 bits
83 vmul.u16 q8, q8, q3 // multiply dst red0 by src alpha
90 vadd.u16 q8, q8, q0 // add src red to dst red0
96 vshr.u16 q8, q8, #8 // shift down red0
104 vsli.u16 q10, q8, #11 // shift & insert red0 into blue0
|
| /external/libvpx/vp8/common/arm/neon/ |
| sixtappredict4x4_neon.asm | 75 vmull.u8 q8, d20, d5
85 vmlal.u8 q8, d10, d0
92 vmlsl.u8 q8, d20, d1
99 vmlsl.u8 q8, d10, d4
106 vmlal.u8 q8, d20, d2
117 vqadd.s16 q8, q10
123 vqrshrun.s16 d28, q8, #7
140 vmull.u8 q8, d20, d5
152 vmlal.u8 q8, d10, d0
162 vmlsl.u8 q8, d20, d
[all...] |
| sixtappredict8x4_neon.asm | 69 vmull.u8 q8, d8, d0
79 vmlsl.u8 q8, d29, d1
89 vmlsl.u8 q8, d29, d4
99 vmlal.u8 q8, d29, d2
109 vmlal.u8 q8, d29, d5
124 vqadd.s16 q8, q4
131 vqrshrun.s16 d23, q8, #7
145 vmull.u8 q8, d6, d0 ;(src_ptr[-2] * vp8_filter[0])
157 vmlsl.u8 q8, d27, d1 ;-(src_ptr[-1] * vp8_filter[1])
169 vmlsl.u8 q8, d27, d4 ;-(src_ptr[2] * vp8_filter[4]
[all...] |
| sixtappredict8x8_neon.asm | 73 vmull.u8 q8, d8, d0
83 vmlsl.u8 q8, d29, d1
93 vmlsl.u8 q8, d29, d4
103 vmlal.u8 q8, d29, d2
113 vmlal.u8 q8, d29, d5
130 vqadd.s16 q8, q4
137 vqrshrun.s16 d23, q8, #7
158 vmull.u8 q8, d6, d0 ;(src_ptr[-2] * vp8_filter[0])
170 vmlsl.u8 q8, d27, d1 ;-(src_ptr[-1] * vp8_filter[1])
182 vmlsl.u8 q8, d27, d4 ;-(src_ptr[2] * vp8_filter[4]
[all...] |
| loopfilter_neon.asm | 48 vld1.u8 {q8}, [r2], r1 ; q1
59 vst1.u8 {q8}, [r0], r1 ; store oq1
163 vtrn.32 q4, q8
170 vtrn.16 q8, q10
174 vtrn.8 q7, q8
247 vtrn.32 q4, q8
254 vtrn.16 q8, q10
258 vtrn.8 q7, q8
307 ; q8 q1
317 vabd.u8 q14, q8, q7 ; abs(q1 - q0
[all...] |
| bilinearpredict8x8_neon.asm | 48 vmull.u8 q8, d6, d0
58 vmlal.u8 q8, d7, d1
66 vqrshrn.u16 d24, q8, #7
75 vmull.u8 q8, d6, d0
87 vmlal.u8 q8, d7, d1
93 vqrshrn.u16 d28, q8, #7
118 vmull.u8 q8, d29, d0
127 vmlal.u8 q8, d30, d1
136 vqrshrn.u16 d9, q8, #7
|
| mbloopfilter_neon.asm | 48 vld1.u8 {q8}, [r0], r1 ; q1
64 vst1.u8 {q8}, [r12], r1 ; store oq1
169 vtrn.32 q4, q8
176 vtrn.16 q8, q10
180 vtrn.8 q7, q8
202 vtrn.32 q4, q8
210 vtrn.16 q8, q10
214 vtrn.8 q7, q8
283 vtrn.32 q4, q8
290 vtrn.16 q8, q1
[all...] |
| sixtappredict16x16_neon.asm | 78 vmull.u8 q8, d6, d0 ;(src_ptr[-2] * vp8_filter[0])
89 vmlsl.u8 q8, d28, d1 ;-(src_ptr[-1] * vp8_filter[1])
105 vmlsl.u8 q8, d28, d4 ;-(src_ptr[2] * vp8_filter[4])
121 vmlal.u8 q8, d28, d5 ;(src_ptr[3] * vp8_filter[5])
137 vmlal.u8 q8, d28, d2 ;(src_ptr[0] * vp8_filter[2])
161 vqadd.s16 q8, q4 ;sum of all (src_data*filter_parameters)
175 vqrshrun.s16 d6, q8, #7 ;shift/round/saturate to u8
196 vabs.s32 q8, q6
247 vmull.u8 q8, d22, d3
254 vqadd.s16 q8, q
[all...] |
| loopfiltersimplehorizontaledge_neon.asm | 39 vld1.u8 {q8}, [r0] ; q1
43 vabd.u8 q14, q5, q8 ; abs(p1 - q1)
52 veor q8, q8, q0 ; qs1: q1 offset to convert to a signed value
63 vqsub.s8 q4, q5, q8 ; q4: vp8_filter = vp8_signed_char_clamp(ps1-qs1)
|
| bilinearpredict16x16_neon.asm | 63 vmull.u8 q8, d3, d0
86 vmlal.u8 q8, d3, d1 ;(src_ptr[0] * vp8_filter[1])
94 vqrshrn.u16 d15, q8, #7
196 vmull.u8 q8, d29, d0
205 vmlal.u8 q8, d31, d1
216 vqrshrn.u16 d9, q8, #7
248 vmull.u8 q8, d3, d0
271 vmlal.u8 q8, d3, d1 ;(src_ptr[0] * vp8_filter[1])
279 vqrshrn.u16 d15, q8, #7
320 vmull.u8 q8, d29, d
[all...] |
| recon16x16mb_neon.asm | 31 vld1.16 {q8, q9}, [r1]! ;load data from diff_ptr
50 vadd.s16 q0, q0, q8 ;add Diff data and Pred data together
82 vld1.16 {q8, q9}, [r1]! ;load data from diff_ptr
97 vadd.s16 q0, q0, q8 ;add Diff data and Pred data together
|
| /external/libvpx/vp8/decoder/arm/neon/ |
| idct_dequant_dc_full_2x_neon.asm | 69 ; q8: 4 * cospi
73 vqdmulh.s16 q8, q4, d0[0] ; cospi8sqrt2minus1
87 vshr.s16 q8, q8, #1
92 vqadd.s16 q4, q4, q8
120 ; q8: 1 * sinpi : c1/temp1
124 vqdmulh.s16 q8, q5, d0[2] ; sinpi8sqrt2
141 ; q8: c1 = temp1 - temp2
143 vqsub.s16 q8, q8, q1
[all...] |
| idct_dequant_full_2x_neon.asm | 61 ; q8: 4 * cospi
65 vqdmulh.s16 q8, q4, d0[0] ; cospi8sqrt2minus1
79 vshr.s16 q8, q8, #1
84 vqadd.s16 q4, q4, q8
112 ; q8: 1 * sinpi : c1/temp1
116 vqdmulh.s16 q8, q5, d0[2] ; sinpi8sqrt2
133 ; q8: c1 = temp1 - temp2
135 vqsub.s16 q8, q8, q1
[all...] |
| /external/libvpx/vpx_scale/arm/neon/ |
| vp8_vpxyv12_copyframe_func_neon.asm | 59 vld1.8 {q8, q9}, [r10]!
70 vst1.8 {q8, q9}, [r11]!
117 vld1.8 {q8, q9}, [r10]!
124 vst1.8 {q8, q9}, [r11]!
|
| vp8_vpxyv12_copyframeyonly_neon.asm | 49 vld1.8 {q8, q9}, [r10]!
60 vst1.8 {q8, q9}, [r11]!
123 vmov q9, q8
134 vst1.8 {q8, q9}, [r5], lr
154 vld1.8 {q8, q9}, [r2]!
168 vst1.8 {q8, q9}, [r6]!
264 vst1.8 {q8}, [r5], lr
284 vld1.8 {q8, q9}, [r2]!
298 vst1.8 {q8, q9}, [r6]!
426 vld1.8 {q8, q9}, [r10]
[all...] |
