[FFmpeg-cvslog] Merge commit 'effc1430b2fe5997d9d55bf28dc507c27125eb27'

[James Almer]

ffmpeg | branch: master | James Almer <jamrial at gmail.com> | Tue Apr  4 15:26:18 2017 -0300| [6747fc436e05c4005d0612a275c996c72b7c2736] | committer: James Almer

Merge commit 'effc1430b2fe5997d9d55bf28dc507c27125eb27'

* commit 'effc1430b2fe5997d9d55bf28dc507c27125eb27':
  Revert "checkasm: vp9dsp: Benchmark the dc-only version of idct_idct separately"

Merged-by: James Almer <jamrial at gmail.com>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=6747fc436e05c4005d0612a275c996c72b7c2736
---

 tests/checkasm/vp9dsp.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/tests/checkasm/vp9dsp.c b/tests/checkasm/vp9dsp.c
index 337ab21..a91577b 100644
--- a/tests/checkasm/vp9dsp.c
+++ b/tests/checkasm/vp9dsp.c
@@ -368,12 +368,6 @@ static void check_itxfm(void)
                         bench_new(dst, sz * SIZEOF_PIXEL, coef, eob);
                     }
                 }
-                if (txtp == 0 && tx != 4) {
-                    if (check_func(dsp.itxfm_add[tx][txtp], "vp9_inv_%s_%dx%d_dc_add_%d",
-                                   txtp_types[txtp], sz, sz, bit_depth)) {
-                        bench_new(dst, sz * SIZEOF_PIXEL, coef, 1);
-                    }
-                }
             }
         }
     }


======================================================================

diff --cc tests/checkasm/vp9dsp.c
index 337ab21,690e0cf..a91577b
--- a/tests/checkasm/vp9dsp.c
+++ b/tests/checkasm/vp9dsp.c
@@@ -323,57 -262,40 +323,51 @@@ static void check_itxfm(void
          [ADST_DCT] = "dct_adst", [ADST_ADST] = "adst_adst"
      };
  
 -    ff_vp9dsp_init(&dsp);
 -
 -    for (tx = TX_4X4; tx <= N_TXFM_SIZES /* 4 = lossless */; tx++) {
 -        int sz = 4 << (tx & 3);
 -        int n_txtps = tx < TX_32X32 ? N_TXFM_TYPES : 1;
 -
 -        for (txtp = 0; txtp < n_txtps; txtp++) {
 -            if (check_func(dsp.itxfm_add[tx][txtp], "vp9_inv_%s_%dx%d_add",
 -                           tx == 4 ? "wht_wht" : txtp_types[txtp], sz, sz)) {
 -                randomize_buffers();
 -                ftx(coef, tx, txtp, sz, BIT_DEPTH);
 -
 -                for (sub = (txtp == 0) ? 1 : 2; sub <= sz; sub <<= 1) {
 -                    int eob;
 +    for (bit_depth = 8; bit_depth <= 12; bit_depth += 2) {
 +        ff_vp9dsp_init(&dsp, bit_depth, 0);
 +
 +        for (tx = TX_4X4; tx <= N_TXFM_SIZES /* 4 = lossless */; tx++) {
 +            int sz = 4 << (tx & 3);
 +            int n_txtps = tx < TX_32X32 ? N_TXFM_TYPES : 1;
 +
 +            for (txtp = 0; txtp < n_txtps; txtp++) {
 +                // skip testing sub-IDCTs for WHT or ADST since they don't
 +                // implement it in any of the SIMD functions. If they do,
 +                // consider changing this to ensure we have complete test
 +                // coverage. Test sub=1 for dc-only, then 2, 4, 8, 12, etc,
 +                // since the arm version can distinguish them at that level.
 +                for (sub = (txtp == 0 && tx < 4) ? 1 : sz; sub <= sz;
 +                     sub < 4 ? (sub <<= 1) : (sub += 4)) {
 +                    if (check_func(dsp.itxfm_add[tx][txtp],
 +                                   "vp9_inv_%s_%dx%d_sub%d_add_%d",
 +                                   tx == 4 ? "wht_wht" : txtp_types[txtp],
 +                                   sz, sz, sub, bit_depth)) {
 +                        int eob;
 +
 +                        randomize_buffers();
 +                        ftx(coef, tx, txtp, sz, bit_depth);
 +
 +                        if (sub < sz) {
 +                            eob = copy_subcoefs(subcoef0, coef, tx, txtp,
 +                                                sz, sub, bit_depth);
 +                        } else {
 +                            eob = sz * sz;
 +                            memcpy(subcoef0, coef, sz * sz * SIZEOF_COEF);
 +                        }
  
 -                    if (sub < sz) {
 -                        eob = copy_subcoefs(subcoef0, coef, tx, txtp,
 -                                            sz, sub, BIT_DEPTH);
 -                    } else {
 -                        eob = sz * sz;
 -                        memcpy(subcoef0, coef, sz * sz * SIZEOF_COEF);
 +                        memcpy(dst0, dst, sz * sz * SIZEOF_PIXEL);
 +                        memcpy(dst1, dst, sz * sz * SIZEOF_PIXEL);
 +                        memcpy(subcoef1, subcoef0, sz * sz * SIZEOF_COEF);
 +                        call_ref(dst0, sz * SIZEOF_PIXEL, subcoef0, eob);
 +                        call_new(dst1, sz * SIZEOF_PIXEL, subcoef1, eob);
 +                        if (memcmp(dst0, dst1, sz * sz * SIZEOF_PIXEL) ||
 +                            !iszero(subcoef0, sz * sz * SIZEOF_COEF) ||
 +                            !iszero(subcoef1, sz * sz * SIZEOF_COEF))
 +                            fail();
 +
 +                        bench_new(dst, sz * SIZEOF_PIXEL, coef, eob);
                      }
 -
 -                    memcpy(dst0, dst, sz * sz * SIZEOF_PIXEL);
 -                    memcpy(dst1, dst, sz * sz * SIZEOF_PIXEL);
 -                    memcpy(subcoef1, subcoef0, sz * sz * SIZEOF_COEF);
 -                    call_ref(dst0, sz * SIZEOF_PIXEL, subcoef0, eob);
 -                    call_new(dst1, sz * SIZEOF_PIXEL, subcoef1, eob);
 -                    if (memcmp(dst0, dst1, sz * sz * SIZEOF_PIXEL) ||
 -                        !iszero(subcoef0, sz * sz * SIZEOF_COEF) ||
 -                        !iszero(subcoef1, sz * sz * SIZEOF_COEF))
 -                        fail();
                  }
-                 if (txtp == 0 && tx != 4) {
-                     if (check_func(dsp.itxfm_add[tx][txtp], "vp9_inv_%s_%dx%d_dc_add_%d",
-                                    txtp_types[txtp], sz, sz, bit_depth)) {
-                         bench_new(dst, sz * SIZEOF_PIXEL, coef, 1);
-                     }
-                 }
 -                bench_new(dst, sz * SIZEOF_PIXEL, coef, sz * sz);
              }
          }
      }