Home
last modified time | relevance | path

Searched refs:transpose (Results 1 – 25 of 28) sorted by relevance

12

/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp9/common/arm/neon/
Dvp9_idct4x4_add_neon.asm30 ; transposed rows) and then transpose the results (so that it goes back
33 ; So, two passes of a transpose followed by a column transform.
49 ; transpose the input data
107 ; transpose the results
Dvp9_iht4x4_add_neon.asm155 ; transpose the input data
172 ; transpose the matrix
188 ; transpose the matrix
203 ; transpose the matrix
Dvp9_iht8x8_add_neon.asm581 ; transpose the input data
597 ; transpose the matrix
615 ; transpose the matrix
633 ; transpose the matrix
Dvp9_loopfilter_neon.asm124 ;transpose to 8x16 matrix
165 ; necessary load, transpose (if necessary) and store. The function does not use
375 ;transpose to 8x16 matrix
427 ; necessary load, transpose (if necessary) and store. The function does not use
Dvp9_idct32x32_add_neon.asm299 ; r8 transpose loop counter
320 ; internal buffer used to transpose 8 lines into before transforming them
323 ; results of the first pass (transpose and transform rows)
326 ; results of the second pass (transpose and transform columns)
355 mov r8, #2 ; initialize transpose loop counter
434 ; transpose pair loop processing
Dvp9_loopfilter_16_neon.asm83 ; necessary load, transpose (if necessary) and store. This function uses
Dvp9_idct8x8_add_neon.asm215 ; transpose the input data
327 ; transpose the input data
Dvp9_convolve8_neon.asm146 ; transpose
Dvp9_convolve8_avg_neon.asm154 ; transpose
Dvp9_idct16x16_add_neon.asm71 ; transpose the input data
318 ; transpose the input data
821 ; transpose the input data
950 ; transpose the input data
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/encoder/arm/neon/
Dshortfdct_neon.asm43 ; transpose d0=ip[0], d1=ip[1], d2=ip[2], d3=ip[3]
72 ; transpose d0=ip[0], d1=ip[4], d2=ip[8], d3=ip[12]
124 ; transpose q0=ip[0], q1=ip[1], q2=ip[2], q3=ip[3]
165 ; transpose q0=ip[0], q1=ip[4], q2=ip[8], q3=ip[12]
Dvp8_shortwalsh4x4_neon.asm31 ;transpose d0, d1, d2, d3. Then, d0=ip[0], d1=ip[1], d2=ip[2], d3=ip[3]
59 ;transpose d0, d1, d2, d3, Then, d0=ip[0], d1=ip[4], d2=ip[8], d3=ip[12]
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/common/x86/
Didctllm_sse2.asm126 ; note the transpose of xmm1 and xmm2, necessary for shuffle
198 ; transpose for the second pass
276 ; transpose to save
458 ; note the transpose of xmm1 and xmm2, necessary for shuffle
535 ; transpose for the second pass
613 ; transpose to save
Dloopfilter_sse2.asm952 ;transpose 16x8 to 8x16, and store the 8-line result on stack.
961 ; transpose and write back - only work on q1, q0, p0, p1
1017 ;transpose 16x8 to 8x16, and store the 8-line result on stack.
1026 ; transpose and write back - only work on q1, q0, p0, p1
1179 ; transpose and write back
1242 ; transpose and write back
1541 ; transpose back to write out
Dloopfilter_mmx.asm530 ; transpose and write back
1292 ; transpose and write back
1655 ; transpose back to write out
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/common/arm/neon/
Dmbloopfilter_neon.asm155 ;transpose to 8x16 matrix
177 ;transpose to 16x8 matrix
250 ;transpose to 8x16 matrix
272 ;transpose to 16x8 matrix
311 ; functions do the necessary load, transpose (if necessary), preserve (if
Dloopfilter_neon.asm148 ;transpose to 8x16 matrix
238 ;transpose to 8x16 matrix
289 ; necessary load, transpose (if necessary) and store.
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/common/ppc/
Didctllm_altivec.asm73 ;# transpose input
128 ;# transpose output
Dloopfilter_filters_altivec.asm26 ;# We often need to perform transposes (and other transpose-like operations)
43 ;# A 16x16 transpose can then be thought of as an operation on
45 ;# memory and the effect of a transpose is to interchange address bit
89 ;# edges together. This requires a single 16x16 transpose, which, in
103 ;# For clarity, and because we can afford it, we do this transpose
137 ;# Whole transpose takes 4*16 = 64 instructions
173 ;# In other words, we transpose each of the four 4x4 submatrices.
216 ;# Normal mb vertical edge filter transpose.
247 ;# Inverse transpose is similar, except here I -> (I+3) mod 7 and the
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/encoder/x86/
Ddct_mmx.asm38 ; transpose for the first stage
114 ; transpose for the second stage
Ddct_sse2.asm181 ; transpose for the first stage
266 ; transpose for the second stage
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/common/arm/armv6/
Dsixtappredict8x4_v6.asm23 ;note: In first pass, store the result in transpose(8linesx9columns) on stack. Temporary stack size…
25 ;and the result is stored in transpose.
Dsimpleloopfilter_v6.asm176 ;transpose r7, r8, r9, r10 to r3, r4, r5, r6
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp9/common/x86/
Dvp9_loopfilter_intrin_sse2.c1533 static INLINE void transpose(unsigned char *src[], int in_p, in transpose() function
1618 transpose(src, 16, dst, p, 2); in vp9_lpf_vertical_4_dual_sse2()
1634 transpose(src, p, dst, 8, 1); in vp9_lpf_vertical_8_sse2()
1643 transpose(src, 8, dst, p, 1); in vp9_lpf_vertical_8_sse2()
1669 transpose(src, 16, dst, p, 2); in vp9_lpf_vertical_8_dual_sse2()
1686 transpose(src, p, dst, 8, 2); in vp9_lpf_vertical_16_sse2()
1697 transpose(src, 8, dst, p, 2); in vp9_lpf_vertical_16_sse2()
/hardware/intel/common/omx-components/videocodec/libvpx_internal/libvpx/vp8/encoder/ppc/
Dfdct_altivec.asm77 ;# forward transform uses transpose.

12