max/dav1d
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

riscv64/itx: Add 16-point 8bpc RVV flipadst transform

Browse Source 
inv_txfm_add_16x16_adst_flipadst_0_8bpc_c:        15272.2 ( 1.00x)
inv_txfm_add_16x16_adst_flipadst_0_8bpc_rvv:       1824.4 ( 8.37x)
inv_txfm_add_16x16_adst_flipadst_1_8bpc_c:        15261.2 ( 1.00x)
inv_txfm_add_16x16_adst_flipadst_1_8bpc_rvv:       1824.5 ( 8.36x)
inv_txfm_add_16x16_adst_flipadst_2_8bpc_c:        15260.0 ( 1.00x)
inv_txfm_add_16x16_adst_flipadst_2_8bpc_rvv:       1824.5 ( 8.36x)
inv_txfm_add_16x16_dct_flipadst_0_8bpc_c:         14497.2 ( 1.00x)
inv_txfm_add_16x16_dct_flipadst_0_8bpc_rvv:        1637.3 ( 8.85x)
inv_txfm_add_16x16_dct_flipadst_1_8bpc_c:         14490.5 ( 1.00x)
inv_txfm_add_16x16_dct_flipadst_1_8bpc_rvv:        1637.3 ( 8.85x)
inv_txfm_add_16x16_dct_flipadst_2_8bpc_c:         14486.4 ( 1.00x)
inv_txfm_add_16x16_dct_flipadst_2_8bpc_rvv:        1637.3 ( 8.85x)
inv_txfm_add_16x16_flipadst_adst_0_8bpc_c:        15307.7 ( 1.00x)
inv_txfm_add_16x16_flipadst_adst_0_8bpc_rvv:       1808.0 ( 8.47x)
inv_txfm_add_16x16_flipadst_adst_1_8bpc_c:        15341.0 ( 1.00x)
inv_txfm_add_16x16_flipadst_adst_1_8bpc_rvv:       1808.1 ( 8.48x)
inv_txfm_add_16x16_flipadst_adst_2_8bpc_c:        15333.5 ( 1.00x)
inv_txfm_add_16x16_flipadst_adst_2_8bpc_rvv:       1808.1 ( 8.48x)
inv_txfm_add_16x16_flipadst_dct_0_8bpc_c:         14530.0 ( 1.00x)
inv_txfm_add_16x16_flipadst_dct_0_8bpc_rvv:        1636.4 ( 8.88x)
inv_txfm_add_16x16_flipadst_dct_1_8bpc_c:         14510.3 ( 1.00x)
inv_txfm_add_16x16_flipadst_dct_1_8bpc_rvv:        1636.3 ( 8.87x)
inv_txfm_add_16x16_flipadst_dct_2_8bpc_c:         14504.7 ( 1.00x)
inv_txfm_add_16x16_flipadst_dct_2_8bpc_rvv:        1636.3 ( 8.86x)
inv_txfm_add_16x16_flipadst_flipadst_0_8bpc_c:    15315.5 ( 1.00x)
inv_txfm_add_16x16_flipadst_flipadst_0_8bpc_rvv:   1823.5 ( 8.40x)
inv_txfm_add_16x16_flipadst_flipadst_1_8bpc_c:    15324.1 ( 1.00x)
inv_txfm_add_16x16_flipadst_flipadst_1_8bpc_rvv:   1823.3 ( 8.40x)
inv_txfm_add_16x16_flipadst_flipadst_2_8bpc_c:    15315.6 ( 1.00x)
inv_txfm_add_16x16_flipadst_flipadst_2_8bpc_rvv:   1823.5 ( 8.40x)
This commit is contained in:
Nathan E. Egge
2024-02-12 18:43:40 -05:00
parent b981bc9c3e
commit 7b15ca1375
2 changed files with 11 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/riscv/64/itx.S
+10
View File
@@ -1188,6 +1188,11 @@ function inv_adst_e16_x16_rvv, export=1, ext=v
jr t0
endfunc
function inv_flipadst_e16_x16_rvv, export=1, ext=v
iadst_16 v15, v14, v13, v12, v11, v10, v9, v8, v7, v6, v5, v4, v3, v2, v1, v0
jr t0
endfunc
.macro def_horz_16 variant
function inv_txfm_horz\variant\()_16x8_rvv, export=1, ext=v
vmv.v.x v16, zero
@@ -1323,7 +1328,12 @@ endfunc
def_fn_16x16 dct, dct
def_fn_16x16 identity, identity
def_fn_16x16 dct, adst
def_fn_16x16 dct, flipadst
def_fn_16x16 dct, identity
def_fn_16x16 adst, dct
def_fn_16x16 adst, adst
def_fn_16x16 adst, flipadst
def_fn_16x16 flipadst, dct
def_fn_16x16 flipadst, adst
def_fn_16x16 flipadst, flipadst
def_fn_16x16 identity, dct
src/riscv/itx.h
+1 -6
View File
@@ -106,11 +106,6 @@ static ALWAYS_INLINE void itx_dsp_init_riscv(Dav1dInvTxfmDSPContext *const c, in
#if BITDEPTH == 8
assign_itx17_fn( , 4, 4, rvv);
assign_itx16_fn( , 8, 8, rvv);
assign_itx2_fn( , 16, 16, rvv);
assign_itx_fn( , 16, 16, dct_adst, ADST_DCT, rvv);
assign_itx_fn( , 16, 16, dct_identity, H_DCT, rvv);
assign_itx_fn( , 16, 16, adst_dct, DCT_ADST, rvv);
assign_itx_fn( , 16, 16, adst_adst, ADST_ADST, rvv);
assign_itx_fn( , 16, 16, identity_dct, V_DCT, rvv);
assign_itx12_fn( , 16, 16, rvv);
#endif
}