aosp12/external/libavc/encoder/armv8/ih264e_evaluate_intra16x16_...

593 lines
16 KiB
ArmAsm

//******************************************************************************
//*
//* Copyright (C) 2015 The Android Open Source Project
//*
//* Licensed under the Apache License, Version 2.0 (the "License");
//* you may not use this file except in compliance with the License.
//* You may obtain a copy of the License at:
//*
//* http://www.apache.org/licenses/LICENSE-2.0
//*
//* Unless required by applicable law or agreed to in writing, software
//* distributed under the License is distributed on an "AS IS" BASIS,
//* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
//* See the License for the specific language governing permissions and
//* limitations under the License.
//*
//*****************************************************************************
//* Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore
//*/
///**
//******************************************************************************
//*
//* @brief :Evaluate best intra 16x16 mode (among VERT, HORZ and DC )
//* and do the prediction.
//*
//* @par Description
//* This function evaluates first three 16x16 modes and compute corresponding sad
//* and return the buffer predicted with best mode.
//*
//* @param[in] pu1_src
//* UWORD8 pointer to the source
//*
//** @param[in] pu1_ngbr_pels_i16
//* UWORD8 pointer to neighbouring pels
//*
//* @param[out] pu1_dst
//* UWORD8 pointer to the destination
//*
//* @param[in] src_strd
//* integer source stride
//*
//* @param[in] dst_strd
//* integer destination stride
//*
//* @param[in] u4_n_avblty
//* availability of neighbouring pixels
//*
//* @param[in] u4_intra_mode
//* Pointer to the variable in which best mode is returned
//*
//* @param[in] pu4_sadmin
//* Pointer to the variable in which minimum sad is returned
//*
//* @param[in] u4_valid_intra_modes
//* Says what all modes are valid
//*
//*
//* @return none
//*
//******************************************************************************
//*/
//
//void ih264e_evaluate_intra16x16_modes(UWORD8 *pu1_src,
// UWORD8 *pu1_ngbr_pels_i16,
// UWORD8 *pu1_dst,
// UWORD32 src_strd,
// UWORD32 dst_strd,
// WORD32 u4_n_avblty,
// UWORD32 *u4_intra_mode,
// WORD32 *pu4_sadmin,
// UWORD32 u4_valid_intra_modes)
//
.text
.p2align 2
.include "ih264_neon_macros.s"
.global ih264e_evaluate_intra16x16_modes_av8
ih264e_evaluate_intra16x16_modes_av8:
//x0 = pu1_src,
//x1 = pu1_ngbr_pels_i16,
//x2 = pu1_dst,
//w3 = src_strd,
//w4 = dst_strd,
//w5 = u4_n_avblty,
//x6 = u4_intra_mode,
//x7 = pu4_sadmin
// STMFD sp!, {x4-x12, x14} //store register values to stack
push_v_regs
sxtw x3, w3
sxtw x4, w4
stp x19, x20, [sp, #-16]!
ldr w16, [sp, #80]
mov x17, x4
mov x14, x6
mov x15, x7
sub v0.16b, v0.16b, v0.16b
sub v1.16b, v1.16b, v1.16b
mov w10, #0
mov w11 , #3
ands w6, w5, #0x01
beq top_available //LEFT NOT AVAILABLE
ld1 {v0.16b}, [x1]
add w10, w10, #8
add w11, w11, #1
top_available:
ands w6, w5, #0x04
beq none_available
add x6, x1, #17
ld1 {v1.16b}, [x6]
add w10, w10, #8
add w11, w11, #1
b summation
none_available:
cmp w5, #0
bne summation
mov w6, #128
dup v30.16b, w6
dup v31.16b, w6
b sad_comp
summation:
uaddl v2.8h, v0.8b, v1.8b
uaddl2 v3.8h, v0.16b, v1.16b
dup v10.8h, w10
neg w11, w11
dup v20.8h, w11
add v0.8h, v2.8h, v3.8h
mov v1.d[0], v0.d[1]
add v0.4h, v0.4h, v1.4h
addp v0.4h, v0.4h , v0.4h
addp v0.4h, v0.4h , v0.4h
add v0.4h, v0.4h, v10.4h
uqshl v0.8h, v0.8h, v20.8h
sqxtun v0.8b, v0.8h
dup v30.16b, v0.b[0]
dup v31.16b, v0.b[0]
sad_comp:
ld1 { v0.2s, v1.2s }, [x0], x3 // source x0w 0
ld1 { v2.2s, v3.2s}, [x0], x3 //row 1
ld1 { v4.2s, v5.2s}, [x0], x3 //row 2
ld1 { v6.2s, v7.2s}, [x0], x3 //row 3
//---------------------
//values for vertical prediction
add x6, x1, #17
ld1 {v10.8b}, [x6], #8
ld1 {v11.8b}, [x6], #8
ld1 {v9.16b}, [x1]
dup v20.8b, v9.b[15] ///HORIZONTAL VALUE ROW=0//
dup v21.8b, v9.b[15] ///HORIZONTAL VALUE ROW=0//
///* computing SADs for all three modes*/
///vertical row 0@
uabdl v16.8h, v0.8b, v10.8b
uabdl v18.8h, v1.8b, v11.8b
///HORZ row 0@
uabdl v26.8h, v0.8b, v20.8b
uabdl v28.8h, v1.8b, v21.8b
///dc row 0@
uabdl v22.8h, v0.8b, v30.8b
uabdl v24.8h, v1.8b, v31.8b
dup v20.8b, v9.b[14] ///HORIZONTAL VALUE ROW=1//
dup v21.8b, v9.b[14]
///vertical row 1@
uabal v16.8h, v2.8b, v10.8b
uabal v18.8h, v3.8b, v11.8b
ld1 { v0.2s, v1.2s }, [x0], x3 //row 4
///HORZ row 1@
uabal v26.8h, v2.8b, v20.8b
uabal v28.8h, v3.8b, v21.8b
///dc row 1@
uabal v22.8h, v2.8b, v30.8b
uabal v24.8h, v3.8b, v31.8b
dup v20.8b, v9.b[13] ///HORIZONTAL VALUE ROW=2//
dup v21.8b, v9.b[13]
///vertical row 2@
uabal v16.8h, v4.8b, v10.8b
uabal v18.8h, v5.8b, v11.8b
ld1 { v2.2s, v3.2s}, [x0], x3 //row 5
///HORZ row 2@
uabal v26.8h, v4.8b, v20.8b
uabal v28.8h, v5.8b, v21.8b
///dc row 2@
uabal v22.8h, v4.8b, v30.8b
uabal v24.8h, v5.8b, v31.8b
dup v20.8b, v9.b[12] ///HORIZONTAL VALUE ROW=3//
dup v21.8b, v9.b[12]
///vertical row 3@
uabal v16.8h, v6.8b, v10.8b
uabal v18.8h, v7.8b, v11.8b
ld1 { v4.2s, v5.2s}, [x0], x3 //row 6
///HORZ row 3@
uabal v26.8h, v6.8b, v20.8b
uabal v28.8h, v7.8b, v21.8b
///dc row 3@
uabal v22.8h, v6.8b, v30.8b
uabal v24.8h, v7.8b, v31.8b
//----------------------------------------------------------------------------------------------
dup v20.8b, v9.b[11] ///HORIZONTAL VALUE ROW=0//
dup v21.8b, v9.b[11]
///vertical row 0@
uabal v16.8h, v0.8b, v10.8b
uabal v18.8h, v1.8b, v11.8b
ld1 { v6.2s, v7.2s}, [x0], x3 //row 7
///HORZ row 0@
uabal v26.8h, v0.8b, v20.8b
uabal v28.8h, v1.8b, v21.8b
///dc row 0@
uabal v22.8h, v0.8b, v30.8b
uabal v24.8h, v1.8b, v31.8b
dup v20.8b, v9.b[10] ///HORIZONTAL VALUE ROW=1//
dup v21.8b, v9.b[10]
///vertical row 1@
uabal v16.8h, v2.8b, v10.8b
uabal v18.8h, v3.8b, v11.8b
ld1 { v0.2s, v1.2s }, [x0], x3 //row 8
///HORZ row 1@
uabal v26.8h, v2.8b, v20.8b
uabal v28.8h, v3.8b, v21.8b
///dc row 1@
uabal v22.8h, v2.8b, v30.8b
uabal v24.8h, v3.8b, v31.8b
dup v20.8b, v9.b[9] ///HORIZONTAL VALUE ROW=2//
dup v21.8b, v9.b[9]
///vertical row 2@
uabal v16.8h, v4.8b, v10.8b
uabal v18.8h, v5.8b, v11.8b
ld1 { v2.2s, v3.2s}, [x0], x3 //row 9
///HORZ row 2@
uabal v26.8h, v4.8b, v20.8b
uabal v28.8h, v5.8b, v21.8b
///dc row 2@
uabal v22.8h, v4.8b, v30.8b
uabal v24.8h, v5.8b, v31.8b
dup v20.8b, v9.b[8] ///HORIZONTAL VALUE ROW=3//
dup v21.8b, v9.b[8]
///vertical row 3@
uabal v16.8h, v6.8b, v10.8b
uabal v18.8h, v7.8b, v11.8b
ld1 { v4.2s, v5.2s}, [x0], x3 //row 10
///HORZ row 3@
uabal v26.8h, v6.8b, v20.8b
uabal v28.8h, v7.8b, v21.8b
///dc row 3@
uabal v22.8h, v6.8b, v30.8b
uabal v24.8h, v7.8b, v31.8b
//-------------------------------------------
dup v20.8b, v9.b[7] ///HORIZONTAL VALUE ROW=0//
dup v21.8b, v9.b[7]
///vertical row 0@
uabal v16.8h, v0.8b, v10.8b
uabal v18.8h, v1.8b, v11.8b
ld1 { v6.2s, v7.2s}, [x0], x3 //row11
///HORZ row 0@
uabal v26.8h, v0.8b, v20.8b
uabal v28.8h, v1.8b, v21.8b
///dc row 0@
uabal v22.8h, v0.8b, v30.8b
uabal v24.8h, v1.8b, v31.8b
dup v20.8b, v9.b[6] ///HORIZONTAL VALUE ROW=1//
dup v21.8b, v9.b[6]
///vertical row 1@
uabal v16.8h, v2.8b, v10.8b
uabal v18.8h, v3.8b, v11.8b
ld1 { v0.2s, v1.2s }, [x0], x3 //row12
///HORZ row 1@
uabal v26.8h, v2.8b, v20.8b
uabal v28.8h, v3.8b, v21.8b
///dc row 1@
uabal v22.8h, v2.8b, v30.8b
uabal v24.8h, v3.8b, v31.8b
dup v20.8b, v9.b[5] ///HORIZONTAL VALUE ROW=2//
dup v21.8b, v9.b[5]
///vertical row 2@
uabal v16.8h, v4.8b, v10.8b
uabal v18.8h, v5.8b, v11.8b
ld1 { v2.2s, v3.2s}, [x0], x3 //row13
///HORZ row 2@
uabal v26.8h, v4.8b, v20.8b
uabal v28.8h, v5.8b, v21.8b
///dc row 2@
uabal v22.8h, v4.8b, v30.8b
uabal v24.8h, v5.8b, v31.8b
dup v20.8b, v9.b[4] ///HORIZONTAL VALUE ROW=3//
dup v21.8b, v9.b[4]
///vertical row 3@
uabal v16.8h, v6.8b, v10.8b
uabal v18.8h, v7.8b, v11.8b
ld1 { v4.2s, v5.2s}, [x0], x3 //row14
///HORZ row 3@
uabal v26.8h, v6.8b, v20.8b
uabal v28.8h, v7.8b, v21.8b
///dc row 3@
uabal v22.8h, v6.8b, v30.8b
uabal v24.8h, v7.8b, v31.8b
//-----------------------------------------------------------------
dup v20.8b, v9.b[3] ///HORIZONTAL VALUE ROW=0//
dup v21.8b, v9.b[3]
///vertical row 0@
uabal v16.8h, v0.8b, v10.8b
uabal v18.8h, v1.8b, v11.8b
ld1 { v6.2s, v7.2s}, [x0], x3 //row15
///HORZ row 0@
uabal v26.8h, v0.8b, v20.8b
uabal v28.8h, v1.8b, v21.8b
///dc row 0@
uabal v22.8h, v0.8b, v30.8b
uabal v24.8h, v1.8b, v31.8b
dup v20.8b, v9.b[2] ///HORIZONTAL VALUE ROW=1//
dup v21.8b, v9.b[2]
///vertical row 1@
uabal v16.8h, v2.8b, v10.8b
uabal v18.8h, v3.8b, v11.8b
///HORZ row 1@
uabal v26.8h, v2.8b, v20.8b
uabal v28.8h, v3.8b, v21.8b
///dc row 1@
uabal v22.8h, v2.8b, v30.8b
uabal v24.8h, v3.8b, v31.8b
dup v20.8b, v9.b[1] ///HORIZONTAL VALUE ROW=2//
dup v21.8b, v9.b[1]
///vertical row 2@
uabal v16.8h, v4.8b, v10.8b
uabal v18.8h, v5.8b, v11.8b
///HORZ row 2@
uabal v26.8h, v4.8b, v20.8b
uabal v28.8h, v5.8b, v21.8b
///dc row 2@
uabal v22.8h, v4.8b, v30.8b
uabal v24.8h, v5.8b, v31.8b
dup v20.8b, v9.b[0] ///HORIZONTAL VALUE ROW=3//
dup v21.8b, v9.b[0]
///vertical row 3@
uabal v16.8h, v6.8b, v10.8b
uabal v18.8h, v7.8b, v11.8b
///HORZ row 3@
uabal v26.8h, v6.8b, v20.8b
uabal v28.8h, v7.8b, v21.8b
///dc row 3@
uabal v22.8h, v6.8b, v30.8b
uabal v24.8h, v7.8b, v31.8b
//------------------------------------------------------------------------------
//vert sum
add v16.8h, v16.8h , v18.8h
mov v18.d[0], v16.d[1]
add v16.4h, v16.4h , v18.4h
uaddlp v16.2s, v16.4h
addp v16.2s, v16.2s, v16.2s
smov x8, v16.s[0] //dc
//horz sum
add v26.8h, v26.8h , v28.8h
mov v28.d[0], v26.d[1]
add v26.4h, v26.4h , v28.4h
uaddlp v26.2s, v26.4h
addp v26.2s, v26.2s, v26.2s
smov x9, v26.s[0]
//dc sum
add v24.8h, v22.8h , v24.8h ///DC
mov v25.d[0], v24.d[1]
add v24.4h, v24.4h , v25.4h ///DC
uaddlp v24.2s, v24.4h ///DC
addp v24.2s, v24.2s, v24.2s ///DC
smov x10, v24.s[0] //dc
//-----------------------
mov x11, #1
lsl x11, x11, #30
mov w0, w16
//--------------------------------------------
ands w7, w0, #01 // vert mode valid????????????
csel x8, x11, x8, eq
ands w6, w0, #02 // horz mode valid????????????
csel x9, x11, x9, eq
ands w6, w0, #04 // dc mode valid????????????
csel x10, x11, x10, eq
//--------------------------------
mov x4, x17
mov x7, x15
mov x6, x14
//---------------------------
//--------------------------
cmp x8, x9
bgt not_vert
cmp x8, x10
bgt do_dc
///----------------------
//DO VERTICAL PREDICTION
str w8 , [x7] //MIN SAD
mov w8, #0
str w8 , [x6] // MODE
add x6, x1, #17
ld1 {v30.16b}, [x6]
b do_dc_vert
//-----------------------------
not_vert: cmp x9, x10
bgt do_dc
///----------------------
//DO HORIZONTAL
str w9 , [x7] //MIN SAD
mov w9, #1
str w9 , [x6] // MODE
ld1 {v0.16b}, [x1]
dup v10.16b, v0.b[15]
dup v11.16b, v0.b[14]
dup v12.16b, v0.b[13]
dup v13.16b, v0.b[12]
st1 {v10.16b}, [x2], x4
dup v14.16b, v0.b[11]
st1 {v11.16b}, [x2], x4
dup v15.16b, v0.b[10]
st1 {v12.16b}, [x2], x4
dup v16.16b, v0.b[9]
st1 {v13.16b}, [x2], x4
dup v17.16b, v0.b[8]
st1 {v14.16b}, [x2], x4
dup v18.16b, v0.b[7]
st1 {v15.16b}, [x2], x4
dup v19.16b, v0.b[6]
st1 {v16.16b}, [x2], x4
dup v20.16b, v0.b[5]
st1 {v17.16b}, [x2], x4
dup v21.16b, v0.b[4]
st1 {v18.16b}, [x2], x4
dup v22.16b, v0.b[3]
st1 {v19.16b}, [x2], x4
dup v23.16b, v0.b[2]
st1 {v20.16b}, [x2], x4
dup v24.16b, v0.b[1]
st1 {v21.16b}, [x2], x4
dup v25.16b, v0.b[0]
st1 {v22.16b}, [x2], x4
st1 {v23.16b}, [x2], x4
st1 {v24.16b}, [x2], x4
st1 {v25.16b}, [x2], x4
b end_func
///-----------------------------
do_dc: ///---------------------------------
//DO DC
str w10 , [x7] //MIN SAD
mov w10, #2
str w10 , [x6] // MODE
do_dc_vert:
st1 {v30.4s}, [x2], x4 //0
st1 {v30.4s}, [x2], x4 //1
st1 {v30.4s}, [x2], x4 //2
st1 {v30.4s}, [x2], x4 //3
st1 {v30.4s}, [x2], x4 //4
st1 {v30.4s}, [x2], x4 //5
st1 {v30.4s}, [x2], x4 //6
st1 {v30.4s}, [x2], x4 //7
st1 {v30.4s}, [x2], x4 //8
st1 {v30.4s}, [x2], x4 //9
st1 {v30.4s}, [x2], x4 //10
st1 {v30.4s}, [x2], x4 //11
st1 {v30.4s}, [x2], x4 //12
st1 {v30.4s}, [x2], x4 //13
st1 {v30.4s}, [x2], x4 //14
st1 {v30.4s}, [x2], x4 //15
///------------------
end_func:
// LDMFD sp!,{x4-x12,PC} //Restoring registers from stack
ldp x19, x20, [sp], #16
pop_v_regs
ret