RepoMirrors/ffmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2024-12-23 15:53:08 +00:00
Code Issues Projects Releases Wiki Activity
8aed3911fc
ffmpeg/libavcodec/riscv/lpc_init.c

44 lines
1.5 KiB
C
Raw Normal View History

lavc/lpc: R-V V apply_welch_window apply_welch_window_even_c: 617.5 apply_welch_window_even_rvv_f64: 235.0 apply_welch_window_odd_c: 709.0 apply_welch_window_odd_rvv_f64: 256.5
2023-12-08 15:42:22 +00:00
/*
* Copyright © 2022 Rémi Denis-Courmont.
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "config.h"
#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
lavc/lpc: R-V V compute_autocorr The loop iterates over the length of the vector, not the order. This is to avoid reloading the same data for each lag value. However this means the loop only works if the maximum order is no larger than VLENB. The loop is roughly equivalent to: for (size_t j = 0; j < lag; j++) autoc[j] = 1.; while (len > lag) { for (ptrdiff_t j = 0; j < lag; j++) autoc[j] += data[j] * *data; data++; len--; } while (len > 0) { for (ptrdiff_t j = 0; j < len; j++) autoc[j] += data[j] * *data; data++; len--; } Since register pressure is only at 50%, it should be possible to implement the same loop for order up to 2xVLENB. But this is left for future work. Performance numbers are all over the place from ~1.25x to ~4x speedups, but at least they are always noticeably better than nothing.
2023-12-08 19:38:20 +00:00
#include "libavutil/riscv/cpu.h"
lavc/lpc: R-V V apply_welch_window apply_welch_window_even_c: 617.5 apply_welch_window_even_rvv_f64: 235.0 apply_welch_window_odd_c: 709.0 apply_welch_window_odd_rvv_f64: 256.5
2023-12-08 15:42:22 +00:00
#include "libavcodec/lpc.h"
void ff_lpc_apply_welch_window_rvv(const int32_t *, ptrdiff_t, double *);
lavc/lpc: R-V V compute_autocorr The loop iterates over the length of the vector, not the order. This is to avoid reloading the same data for each lag value. However this means the loop only works if the maximum order is no larger than VLENB. The loop is roughly equivalent to: for (size_t j = 0; j < lag; j++) autoc[j] = 1.; while (len > lag) { for (ptrdiff_t j = 0; j < lag; j++) autoc[j] += data[j] * *data; data++; len--; } while (len > 0) { for (ptrdiff_t j = 0; j < len; j++) autoc[j] += data[j] * *data; data++; len--; } Since register pressure is only at 50%, it should be possible to implement the same loop for order up to 2xVLENB. But this is left for future work. Performance numbers are all over the place from ~1.25x to ~4x speedups, but at least they are always noticeably better than nothing.
2023-12-08 19:38:20 +00:00
void ff_lpc_compute_autocorr_rvv(const double *, ptrdiff_t, int, double *);
lavc/lpc: R-V V apply_welch_window apply_welch_window_even_c: 617.5 apply_welch_window_even_rvv_f64: 235.0 apply_welch_window_odd_c: 709.0 apply_welch_window_odd_rvv_f64: 256.5
2023-12-08 15:42:22 +00:00
av_cold void ff_lpc_init_riscv(LPCContext *c)
{
#if HAVE_RVV && (__riscv_xlen >= 64)
int flags = av_get_cpu_flags();
lavc/lpc: R-V V compute_autocorr The loop iterates over the length of the vector, not the order. This is to avoid reloading the same data for each lag value. However this means the loop only works if the maximum order is no larger than VLENB. The loop is roughly equivalent to: for (size_t j = 0; j < lag; j++) autoc[j] = 1.; while (len > lag) { for (ptrdiff_t j = 0; j < lag; j++) autoc[j] += data[j] * *data; data++; len--; } while (len > 0) { for (ptrdiff_t j = 0; j < len; j++) autoc[j] += data[j] * *data; data++; len--; } Since register pressure is only at 50%, it should be possible to implement the same loop for order up to 2xVLENB. But this is left for future work. Performance numbers are all over the place from ~1.25x to ~4x speedups, but at least they are always noticeably better than nothing.
2023-12-08 19:38:20 +00:00
if ((flags & AV_CPU_FLAG_RVV_F64) && (flags & AV_CPU_FLAG_RVB_ADDR)) {
lavc/lpc: R-V V apply_welch_window apply_welch_window_even_c: 617.5 apply_welch_window_even_rvv_f64: 235.0 apply_welch_window_odd_c: 709.0 apply_welch_window_odd_rvv_f64: 256.5
2023-12-08 15:42:22 +00:00
c->lpc_apply_welch_window = ff_lpc_apply_welch_window_rvv;
lavc/lpc: R-V V compute_autocorr The loop iterates over the length of the vector, not the order. This is to avoid reloading the same data for each lag value. However this means the loop only works if the maximum order is no larger than VLENB. The loop is roughly equivalent to: for (size_t j = 0; j < lag; j++) autoc[j] = 1.; while (len > lag) { for (ptrdiff_t j = 0; j < lag; j++) autoc[j] += data[j] * *data; data++; len--; } while (len > 0) { for (ptrdiff_t j = 0; j < len; j++) autoc[j] += data[j] * *data; data++; len--; } Since register pressure is only at 50%, it should be possible to implement the same loop for order up to 2xVLENB. But this is left for future work. Performance numbers are all over the place from ~1.25x to ~4x speedups, but at least they are always noticeably better than nothing.
2023-12-08 19:38:20 +00:00
if (ff_get_rv_vlenb() >= c->max_order)
c->lpc_compute_autocorr = ff_lpc_compute_autocorr_rvv;
}
lavc/lpc: R-V V apply_welch_window apply_welch_window_even_c: 617.5 apply_welch_window_even_rvv_f64: 235.0 apply_welch_window_odd_c: 709.0 apply_welch_window_odd_rvv_f64: 256.5
2023-12-08 15:42:22 +00:00
#endif
}
Reference in New Issue Copy Permalink