Skip to content

x86: Add prep_8tap_scaled + put/prep_bilin_scaled AVX2 asm

    mct_scaled_8tap_regular_w4_8bpc_c: 872.1
    mct_scaled_8tap_regular_w4_8bpc_avx2: 125.6
    mct_scaled_8tap_regular_w4_dy1_8bpc_c: 886.3
    mct_scaled_8tap_regular_w4_dy1_8bpc_avx2: 84.0
    mct_scaled_8tap_regular_w4_dy2_8bpc_c: 1189.1
    mct_scaled_8tap_regular_w4_dy2_8bpc_avx2: 84.7

    mct_scaled_8tap_regular_w8_8bpc_c: 2261.0
    mct_scaled_8tap_regular_w8_8bpc_avx2: 306.2
    mct_scaled_8tap_regular_w8_dy1_8bpc_c: 2189.9
    mct_scaled_8tap_regular_w8_dy1_8bpc_avx2: 233.8
    mct_scaled_8tap_regular_w8_dy2_8bpc_c: 3060.3
    mct_scaled_8tap_regular_w8_dy2_8bpc_avx2: 282.8

    mct_scaled_8tap_regular_w16_8bpc_c: 4335.3
    mct_scaled_8tap_regular_w16_8bpc_avx2: 680.7
    mct_scaled_8tap_regular_w16_dy1_8bpc_c: 5137.2
    mct_scaled_8tap_regular_w16_dy1_8bpc_avx2: 578.6
    mct_scaled_8tap_regular_w16_dy2_8bpc_c: 7878.4
    mct_scaled_8tap_regular_w16_dy2_8bpc_avx2: 774.6

    mct_scaled_8tap_regular_w32_8bpc_c: 17871.9
    mct_scaled_8tap_regular_w32_8bpc_avx2: 2954.8
    mct_scaled_8tap_regular_w32_dy1_8bpc_c: 18594.7
    mct_scaled_8tap_regular_w32_dy1_8bpc_avx2: 2073.9
    mct_scaled_8tap_regular_w32_dy2_8bpc_c: 28696.0
    mct_scaled_8tap_regular_w32_dy2_8bpc_avx2: 2852.1

    mct_scaled_8tap_regular_w64_8bpc_c: 46967.5
    mct_scaled_8tap_regular_w64_8bpc_avx2: 7527.5
    mct_scaled_8tap_regular_w64_dy1_8bpc_c: 45564.2
    mct_scaled_8tap_regular_w64_dy1_8bpc_avx2: 5262.9
    mct_scaled_8tap_regular_w64_dy2_8bpc_c: 72793.3
    mct_scaled_8tap_regular_w64_dy2_8bpc_avx2: 7535.9

    mct_scaled_8tap_regular_w128_8bpc_c: 111190.8
    mct_scaled_8tap_regular_w128_8bpc_avx2: 19386.8
    mct_scaled_8tap_regular_w128_dy1_8bpc_c: 122625.0
    mct_scaled_8tap_regular_w128_dy1_8bpc_avx2: 15376.1
    mct_scaled_8tap_regular_w128_dy2_8bpc_c: 197120.6
    mct_scaled_8tap_regular_w128_dy2_8bpc_avx2: 21871.0
mc_scaled_bilinear_w2_8bpc_c: 186.8
mc_scaled_bilinear_w2_8bpc_avx2: 159.5
mc_scaled_bilinear_w2_dy1_8bpc_c: 231.3
mc_scaled_bilinear_w2_dy1_8bpc_avx2: 96.3
mc_scaled_bilinear_w2_dy2_8bpc_c: 326.5
mc_scaled_bilinear_w2_dy2_8bpc_avx2: 115.6

mc_scaled_bilinear_w4_8bpc_c: 311.6
mc_scaled_bilinear_w4_8bpc_avx2: 155.5
mc_scaled_bilinear_w4_dy1_8bpc_c: 410.2
mc_scaled_bilinear_w4_dy1_8bpc_avx2: 112.7
mc_scaled_bilinear_w4_dy2_8bpc_c: 575.1
mc_scaled_bilinear_w4_dy2_8bpc_avx2: 115.3

mc_scaled_bilinear_w8_8bpc_c: 357.4
mc_scaled_bilinear_w8_8bpc_avx2: 235.4
mc_scaled_bilinear_w8_dy1_8bpc_c: 522.7
mc_scaled_bilinear_w8_dy1_8bpc_avx2: 215.8
mc_scaled_bilinear_w8_dy2_8bpc_c: 881.9
mc_scaled_bilinear_w8_dy2_8bpc_avx2: 268.5

mc_scaled_bilinear_w16_8bpc_c: 1325.3
mc_scaled_bilinear_w16_8bpc_avx2: 692.0
mc_scaled_bilinear_w16_dy1_8bpc_c: 1368.1
mc_scaled_bilinear_w16_dy1_8bpc_avx2: 552.8
mc_scaled_bilinear_w16_dy2_8bpc_c: 2415.2
mc_scaled_bilinear_w16_dy2_8bpc_avx2: 703.6

mc_scaled_bilinear_w32_8bpc_c: 4986.1
mc_scaled_bilinear_w32_8bpc_avx2: 2275.6
mc_scaled_bilinear_w32_dy1_8bpc_c: 4194.3
mc_scaled_bilinear_w32_dy1_8bpc_avx2: 1640.9
mc_scaled_bilinear_w32_dy2_8bpc_c: 7399.1
mc_scaled_bilinear_w32_dy2_8bpc_avx2: 2231.9

mc_scaled_bilinear_w64_8bpc_c: 10682.7
mc_scaled_bilinear_w64_8bpc_avx2: 6354.5
mc_scaled_bilinear_w64_dy1_8bpc_c: 14611.6
mc_scaled_bilinear_w64_dy1_8bpc_avx2: 5521.6
mc_scaled_bilinear_w64_dy2_8bpc_c: 26594.6
mc_scaled_bilinear_w64_dy2_8bpc_avx2: 7732.4

mc_scaled_bilinear_w128_8bpc_c: 47076.6
mc_scaled_bilinear_w128_8bpc_avx2: 23023.4
mc_scaled_bilinear_w128_dy1_8bpc_c: 39308.0
mc_scaled_bilinear_w128_dy1_8bpc_avx2: 15433.8
mc_scaled_bilinear_w128_dy2_8bpc_c: 70535.7
mc_scaled_bilinear_w128_dy2_8bpc_avx2: 21781.5
mct_scaled_bilinear_w4_8bpc_c: 226.2
mct_scaled_bilinear_w4_8bpc_avx2: 122.8
mct_scaled_bilinear_w4_dy1_8bpc_c: 267.1
mct_scaled_bilinear_w4_dy1_8bpc_avx2: 83.7
mct_scaled_bilinear_w4_dy2_8bpc_c: 378.9
mct_scaled_bilinear_w4_dy2_8bpc_avx2: 84.5

mct_scaled_bilinear_w8_8bpc_c: 561.5
mct_scaled_bilinear_w8_8bpc_avx2: 283.2
mct_scaled_bilinear_w8_dy1_8bpc_c: 605.5
mct_scaled_bilinear_w8_dy1_8bpc_avx2: 233.4
mct_scaled_bilinear_w8_dy2_8bpc_c: 965.1
mct_scaled_bilinear_w8_dy2_8bpc_avx2: 282.3

mct_scaled_bilinear_w16_8bpc_c: 1959.2
mct_scaled_bilinear_w16_8bpc_avx2: 800.4
mct_scaled_bilinear_w16_dy1_8bpc_c: 1559.1
mct_scaled_bilinear_w16_dy1_8bpc_avx2: 580.9
mct_scaled_bilinear_w16_dy2_8bpc_c: 2741.6
mct_scaled_bilinear_w16_dy2_8bpc_avx2: 754.9

mct_scaled_bilinear_w32_8bpc_c: 5066.5
mct_scaled_bilinear_w32_8bpc_avx2: 2685.4
mct_scaled_bilinear_w32_dy1_8bpc_c: 5931.3
mct_scaled_bilinear_w32_dy1_8bpc_avx2: 2088.9
mct_scaled_bilinear_w32_dy2_8bpc_c: 10573.3
mct_scaled_bilinear_w32_dy2_8bpc_avx2: 2975.3

mct_scaled_bilinear_w64_8bpc_c: 12267.7
mct_scaled_bilinear_w64_8bpc_avx2: 6694.1
mct_scaled_bilinear_w64_dy1_8bpc_c: 15369.7
mct_scaled_bilinear_w64_dy1_8bpc_avx2: 5306.8
mct_scaled_bilinear_w64_dy2_8bpc_c: 26437.5
mct_scaled_bilinear_w64_dy2_8bpc_avx2: 7465.3

mct_scaled_bilinear_w128_8bpc_c: 34085.4
mct_scaled_bilinear_w128_8bpc_avx2: 19666.4
mct_scaled_bilinear_w128_dy1_8bpc_c: 41659.7
mct_scaled_bilinear_w128_dy1_8bpc_avx2: 14700.3
mct_scaled_bilinear_w128_dy2_8bpc_c: 75093.0
mct_scaled_bilinear_w128_dy2_8bpc_avx2: 20273.3
Edited by Victorien Le Couviour--Tuffet

Merge request reports