* [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate
@ 2023-09-29 16:26 Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 2/5] lavc/aacpsdsp: unroll " Rémi Denis-Courmont
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Rémi Denis-Courmont @ 2023-09-29 16:26 UTC (permalink / raw)
To: ffmpeg-devel
Remove some useless vector splat.
---
libavcodec/riscv/aacpsdsp_rvv.S | 12 ++++--------
1 file changed, 4 insertions(+), 8 deletions(-)
diff --git a/libavcodec/riscv/aacpsdsp_rvv.S b/libavcodec/riscv/aacpsdsp_rvv.S
index b581383f77..b85a5cc92c 100644
--- a/libavcodec/riscv/aacpsdsp_rvv.S
+++ b/libavcodec/riscv/aacpsdsp_rvv.S
@@ -238,20 +238,16 @@ func ff_ps_stereo_interpolate_rvv, zve32f
flw ft1, 4(a3)
vfmv.v.f v19, ft3
flw ft2, 8(a3)
- vfmv.v.f v20, ft0
flw ft3, 12(a3)
- vfmv.v.f v21, ft1
fcvt.s.wu ft4, t0 // (float)(vlenb / sizeof (float))
- vfmv.v.f v22, ft2
+ vfmacc.vf v16, ft0, v24 // h0 += (i + 1) * h0_step
fmul.s ft0, ft0, ft4
- vfmv.v.f v23, ft3
+ vfmacc.vf v17, ft1, v24
fmul.s ft1, ft1, ft4
- vfmacc.vv v16, v24, v20 // h0 += (i + 1) * h0_step
+ vfmacc.vf v18, ft2, v24
fmul.s ft2, ft2, ft4
- vfmacc.vv v17, v24, v21
+ vfmacc.vf v19, ft3, v24
fmul.s ft3, ft3, ft4
- vfmacc.vv v18, v24, v22
- vfmacc.vv v19, v24, v23
1:
vsetvli t0, a4, e32, m1, ta, ma
vlseg2e32.v v8, (a0) // v8:l_re, v9:l_im
--
2.40.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 5+ messages in thread
* [FFmpeg-devel] [PATCH 2/5] lavc/aacpsdsp: unroll R-V V stereo interpolate
2023-09-29 16:26 [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate Rémi Denis-Courmont
@ 2023-09-29 16:26 ` Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 3/5] lavc/aacpsdsp: unroll RISC-V V mul_pair_single Rémi Denis-Courmont
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Rémi Denis-Courmont @ 2023-09-29 16:26 UTC (permalink / raw)
To: ffmpeg-devel
---
libavcodec/riscv/aacpsdsp_rvv.S | 46 ++++++++++++++++-----------------
1 file changed, 23 insertions(+), 23 deletions(-)
diff --git a/libavcodec/riscv/aacpsdsp_rvv.S b/libavcodec/riscv/aacpsdsp_rvv.S
index b85a5cc92c..1a92fed515 100644
--- a/libavcodec/riscv/aacpsdsp_rvv.S
+++ b/libavcodec/riscv/aacpsdsp_rvv.S
@@ -223,7 +223,7 @@ func ff_ps_hybrid_synthesis_deint_rvv, zve32x
endfunc
func ff_ps_stereo_interpolate_rvv, zve32f
- vsetvli t0, zero, e32, m1, ta, ma
+ vsetvli t0, zero, e32, m2, ta, ma
vid.v v24
flw ft0, (a2)
vadd.vi v24, v24, 1 // v24[i] = i + 1
@@ -232,43 +232,43 @@ func ff_ps_stereo_interpolate_rvv, zve32f
flw ft2, 8(a2)
vfmv.v.f v16, ft0
flw ft3, 12(a2)
- vfmv.v.f v17, ft1
+ vfmv.v.f v18, ft1
flw ft0, (a3)
- vfmv.v.f v18, ft2
+ vfmv.v.f v20, ft2
flw ft1, 4(a3)
- vfmv.v.f v19, ft3
+ vfmv.v.f v22, ft3
flw ft2, 8(a3)
flw ft3, 12(a3)
fcvt.s.wu ft4, t0 // (float)(vlenb / sizeof (float))
vfmacc.vf v16, ft0, v24 // h0 += (i + 1) * h0_step
fmul.s ft0, ft0, ft4
- vfmacc.vf v17, ft1, v24
+ vfmacc.vf v18, ft1, v24
fmul.s ft1, ft1, ft4
- vfmacc.vf v18, ft2, v24
+ vfmacc.vf v20, ft2, v24
fmul.s ft2, ft2, ft4
- vfmacc.vf v19, ft3, v24
+ vfmacc.vf v22, ft3, v24
fmul.s ft3, ft3, ft4
1:
- vsetvli t0, a4, e32, m1, ta, ma
- vlseg2e32.v v8, (a0) // v8:l_re, v9:l_im
+ vsetvli t0, a4, e32, m2, ta, ma
+ vlseg2e32.v v0, (a0) // v0:l_re, v2:l_im
sub a4, a4, t0
- vlseg2e32.v v10, (a1) // v10:r_re, v11:r_im
- vfmul.vv v12, v8, v16
- vfmul.vv v13, v9, v16
- vfmul.vv v14, v8, v17
- vfmul.vv v15, v9, v17
- vfmacc.vv v12, v10, v18
- vfmacc.vv v13, v11, v18
- vfmacc.vv v14, v10, v19
- vfmacc.vv v15, v11, v19
- vsseg2e32.v v12, (a0)
+ vlseg2e32.v v4, (a1) // v4:r_re, v6:r_im
+ vfmul.vv v8, v0, v16
+ vfmul.vv v10, v2, v16
+ vfmul.vv v12, v0, v18
+ vfmul.vv v14, v2, v18
+ vfmacc.vv v8, v4, v20
+ vfmacc.vv v10, v6, v20
+ vfmacc.vv v12, v4, v22
+ vfmacc.vv v14, v6, v22
+ vsseg2e32.v v8, (a0)
sh3add a0, t0, a0
- vsseg2e32.v v14, (a1)
+ vsseg2e32.v v12, (a1)
sh3add a1, t0, a1
vfadd.vf v16, v16, ft0 // h0 += (vlenb / sizeof (float)) * h0_step
- vfadd.vf v17, v17, ft1
- vfadd.vf v18, v18, ft2
- vfadd.vf v19, v19, ft3
+ vfadd.vf v18, v18, ft1
+ vfadd.vf v20, v20, ft2
+ vfadd.vf v22, v22, ft3
bnez a4, 1b
ret
--
2.40.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 5+ messages in thread
* [FFmpeg-devel] [PATCH 3/5] lavc/aacpsdsp: unroll RISC-V V mul_pair_single
2023-09-29 16:26 [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 2/5] lavc/aacpsdsp: unroll " Rémi Denis-Courmont
@ 2023-09-29 16:26 ` Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 4/5] lavc/aacpsdsp: unroll RISC-V V hybrid_analysis_ileave Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 5/5] lavc/aacpsdsp: unroll RISC-V V hybrid_synthesis_deint Rémi Denis-Courmont
3 siblings, 0 replies; 5+ messages in thread
From: Rémi Denis-Courmont @ 2023-09-29 16:26 UTC (permalink / raw)
To: ffmpeg-devel
---
libavcodec/riscv/aacpsdsp_rvv.S | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavcodec/riscv/aacpsdsp_rvv.S b/libavcodec/riscv/aacpsdsp_rvv.S
index 1a92fed515..f552d50e23 100644
--- a/libavcodec/riscv/aacpsdsp_rvv.S
+++ b/libavcodec/riscv/aacpsdsp_rvv.S
@@ -38,14 +38,14 @@ endfunc
func ff_ps_mul_pair_single_rvv, zve32f
1:
- vsetvli t0, a3, e32, m1, ta, ma
+ vsetvli t0, a3, e32, m4, ta, ma
vlseg2e32.v v24, (a1)
sub a3, a3, t0
vle32.v v16, (a2)
sh3add a1, t0, a1
vfmul.vv v24, v24, v16
sh2add a2, t0, a2
- vfmul.vv v25, v25, v16
+ vfmul.vv v28, v28, v16
vsseg2e32.v v24, (a0)
sh3add a0, t0, a0
bnez a3, 1b
--
2.40.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 5+ messages in thread
* [FFmpeg-devel] [PATCH 4/5] lavc/aacpsdsp: unroll RISC-V V hybrid_analysis_ileave
2023-09-29 16:26 [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 2/5] lavc/aacpsdsp: unroll " Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 3/5] lavc/aacpsdsp: unroll RISC-V V mul_pair_single Rémi Denis-Courmont
@ 2023-09-29 16:26 ` Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 5/5] lavc/aacpsdsp: unroll RISC-V V hybrid_synthesis_deint Rémi Denis-Courmont
3 siblings, 0 replies; 5+ messages in thread
From: Rémi Denis-Courmont @ 2023-09-29 16:26 UTC (permalink / raw)
To: ffmpeg-devel
---
libavcodec/riscv/aacpsdsp_rvv.S | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavcodec/riscv/aacpsdsp_rvv.S b/libavcodec/riscv/aacpsdsp_rvv.S
index f552d50e23..007002de70 100644
--- a/libavcodec/riscv/aacpsdsp_rvv.S
+++ b/libavcodec/riscv/aacpsdsp_rvv.S
@@ -168,10 +168,10 @@ func ff_ps_hybrid_analysis_ileave_rvv, zve32x /* no needs for zve32f here */
mv t4, a4
addi a2, a2, 1
2:
- vsetvli t5, t3, e32, m1, ta, ma
+ vsetvli t5, t3, e32, m4, ta, ma
vlse32.v v16, (t1), t6
sub t3, t3, t5
- vlse32.v v17, (t4), t6
+ vlse32.v v20, (t4), t6
mul t2, t5, t6
vsseg2e32.v v16, (t0)
sh3add t0, t5, t0
--
2.40.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 5+ messages in thread
* [FFmpeg-devel] [PATCH 5/5] lavc/aacpsdsp: unroll RISC-V V hybrid_synthesis_deint
2023-09-29 16:26 [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate Rémi Denis-Courmont
` (2 preceding siblings ...)
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 4/5] lavc/aacpsdsp: unroll RISC-V V hybrid_analysis_ileave Rémi Denis-Courmont
@ 2023-09-29 16:26 ` Rémi Denis-Courmont
3 siblings, 0 replies; 5+ messages in thread
From: Rémi Denis-Courmont @ 2023-09-29 16:26 UTC (permalink / raw)
To: ffmpeg-devel
---
libavcodec/riscv/aacpsdsp_rvv.S | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavcodec/riscv/aacpsdsp_rvv.S b/libavcodec/riscv/aacpsdsp_rvv.S
index 007002de70..fe250cd83b 100644
--- a/libavcodec/riscv/aacpsdsp_rvv.S
+++ b/libavcodec/riscv/aacpsdsp_rvv.S
@@ -203,12 +203,12 @@ func ff_ps_hybrid_synthesis_deint_rvv, zve32x
mv t4, a4
addi a2, a2, 1
2:
- vsetvli t5, t3, e32, m1, ta, ma
+ vsetvli t5, t3, e32, m4, ta, ma
vlseg2e32.v v16, (t1)
sub t3, t3, t5
vsse32.v v16, (t0), t6
mul t2, t5, t6
- vsse32.v v17, (t4), t6
+ vsse32.v v20, (t4), t6
sh3add t1, t5, t1
add t0, t0, t2
add t4, t4, t2
--
2.40.1
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2023-09-29 16:26 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-09-29 16:26 [FFmpeg-devel] [PATCH 1/5] lavc/aacpsdsp: simplify R-V V stereo interpolate Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 2/5] lavc/aacpsdsp: unroll " Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 3/5] lavc/aacpsdsp: unroll RISC-V V mul_pair_single Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 4/5] lavc/aacpsdsp: unroll RISC-V V hybrid_analysis_ileave Rémi Denis-Courmont
2023-09-29 16:26 ` [FFmpeg-devel] [PATCH 5/5] lavc/aacpsdsp: unroll RISC-V V hybrid_synthesis_deint Rémi Denis-Courmont
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git