T-Head C908 (cycles): vc1dsp.vc1_inv_trans_4x4_dc_c: 113.7 vc1dsp.vc1_inv_trans_4x4_dc_rvv_i32: 46.5 (before) vc1dsp.vc1_inv_trans_4x4_dc_rvv_i32: 45.5 (after) vc1dsp.vc1_inv_trans_4x8_dc_c: 230.7 vc1dsp.vc1_inv_trans_4x8_dc_rvv_i32: 65.7 (before) vc1dsp.vc1_inv_trans_4x8_dc_rvv_i32: 52.5 (after) vc1dsp.vc1_inv_trans_8x4_dc_c: 246.7 vc1dsp.vc1_inv_trans_8x4_dc_rvv_i64: 56.7 (before) vc1dsp.vc1_inv_trans_8x4_dc_rvv_i64: 45.5 (after) vc1dsp.vc1_inv_trans_8x8_dc_c: 419.7 vc1dsp.vc1_inv_trans_8x8_dc_rvv_i64: 81.2 (before) vc1dsp.vc1_inv_trans_8x8_dc_rvv_i64: 53.5 (after)
Showing the changes to libavcodec (1 of 1 file). Show all
Unified · Split
| @@ -38,18 +38,20 @@ func ff_vc1_inv_trans_\cols\()x\rows\()_dc_rvv, zve64x, zba · 37 unmodified lines | |||
| 38 | .endif | 38 | .endif |
| 39 | addi t2, t2, 64 | 39 | addi t2, t2, 64 |
| 40 | srai t2, t2, 7 | 40 | srai t2, t2, 7 |
| 41 | .if \rows * \cols == 64 | 41 | .if \rows * \cols >= 32 |
| 42 | vsetvli zero, t0, e16, m8, ta, ma | 42 | vsetvli zero, t0, e8, m\mat_lmul, ta, ma |
| 43 | .elseif \rows * \cols == 32 | ||
| 44 | vsetvli zero, t0, e16, m4, ta, ma | ||
| 45 | .else | 43 | .else |
| 46 | vsetivli zero, \rows * \cols, e16, m2, ta, ma | 44 | vsetivli zero, \rows * \cols, e8, m\mat_lmul, ta, ma |
| 47 | .endif | 45 | .endif |
| 48 | vzext.vf2 v8, v0 | 46 | bgez t2, 1f |
| 49 | vadd.vx v8, v8, t2 | 47 | |
| 50 | vmax.vx v8, v8, zero | 48 | neg t2, t2 |
| 51 | vsetvli zero, zero, e8, m\mat_lmul, ta, ma | 49 | vssubu.vx v0, v0, t2 |
| 52 | vnclipu.wi v0, v8, 0 | 50 | vsetivli zero, \rows, e8, m\row_lmul, ta, ma |
| 51 | vsse\w\().v v0, (a0), a1 | ||
| 52 | ret | ||
| 53 | 1: | ||
| 54 | vsaddu.vx v0, v0, t2 | ||
| 53 | vsetivli zero, \rows, e8, m\row_lmul, ta, ma | 55 | vsetivli zero, \rows, e8, m\row_lmul, ta, ma |
| 54 | vsse\w\().v v0, (a0), a1 | 56 | vsse\w\().v v0, (a0), a1 |
| 55 | ret | 57 | ret |