History

lavc/idctdsp: improve R-V V add_pixels_clamped

Rémi Denis-Courmont committed 3 years ago

Commit d48810f1 parent 600c6f11 file · +14 −14Browse files

Unified · Split

libavcodec/riscv/idctdsp_rvv.S+14−14
@@ -23,7 +23,6 @@ · 22 unmodified lines
2323 func ff_put_pixels_clamped_rvv, zve32x
2424 vsetivli zero, 8, e16, m1, ta, ma
2525 vlseg8e16.v v24, (a0)
26−1:
2726 /* RVV only has signed-signed and unsigned-unsigned clipping.
2827 * We need two steps for signed-to-unsigned clipping. */
2928 vsetvli t0, zero, e16, m8, ta, ma
@@ -54,17 +53,18 @@ func ff_put_signed_pixels_clamped_rvv, zve64x · 24 unmodified lines
5453 ret
5554 endfunc
5655
57−func ff_add_pixels_clamped_rvv, zve32x
58− vsetivli zero, 8, e8, mf2, ta, ma
59− vlseg8e16.v v24, (a0)
60− vlsseg8e8.v v16, (a1), a2
61− vwaddu.wv v24, v24, v16
62− vwaddu.wv v25, v25, v17
63− vwaddu.wv v26, v26, v18
64− vwaddu.wv v27, v27, v19
65− vwaddu.wv v28, v28, v20
66− vwaddu.wv v29, v29, v21
67− vwaddu.wv v30, v30, v22
68− vwaddu.wv v31, v31, v23
69− j 1b
56+func ff_add_pixels_clamped_rvv, zve64x
57+ vsetivli zero, 8, e8, mf2, ta, ma
58+ li t0, 8 * 8
59+ vlse64.v v16, (a1), a2
60+ vsetvli zero, t0, e8, m4, ta, ma
61+ vle16.v v24, (a0)
62+ vwaddu.wv v24, v24, v16
63+ vsetvli zero, zero, e16, m8, ta, ma
64+ vmax.vx v24, v24, zero
65+ vsetvli zero, zero, e8, m4, ta, ma
66+ vnclipu.wi v16, v24, 0
67+ vsetivli zero, 8, e8, mf2, ta, ma
68+ vsse64.v v16, (a1), a2
69+ ret
7070 endfunc

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t