[FFmpeg-devel] [PATCH] Optimization of AC3 floating point decoder for MIPS
Babic, Nedeljko
nbabic at mips.com
Mon Aug 6 13:48:57 CEST 2012
Hi,
I was on vacation, so I didn’t respond last week (sorry).
>>>> diff --git a/libavcodec/fft.c b/libavcodec/fft.c
>>>> index 6b93a5c..8463bfb 100644
>>>> --- a/libavcodec/fft.c
>>>> +++ b/libavcodec/fft.c
>>>> @@ -31,6 +31,7 @@
>>>> #include "libavutil/mathematics.h"
>>>> #include "fft.h"
>>>> #include "fft-internal.h"
>>>> +#include "mips/fft_table.h"
>>>>
>>>> /* cos(2*pi*x/n) for 0<=x<=n/4, followed by its reverse */
>>>> #if !CONFIG_HARDCODED_TABLES
>>>> @@ -157,11 +158,13 @@ av_cold int ff_fft_init(FFTContext *s, int nbits, int inverse)
>>>> s->mdct_calc = ff_mdct_calc_c;
>>>> #endif
>>>>
>>>> + if (ARCH_MIPS) ff_fft_lut_init();
>>>> #if CONFIG_FFT_FLOAT
>>>> if (ARCH_ARM) ff_fft_init_arm(s);
>>>> if (HAVE_ALTIVEC) ff_fft_init_altivec(s);
>>>> if (HAVE_MMX) ff_fft_init_mmx(s);
>>>> if (CONFIG_MDCT) s->mdct_calcw = s->mdct_calc;
>>>> + if (HAVE_MIPSFPU) ff_fft_init_mips(s);
>>>> #else
>>>> if (CONFIG_MDCT) s->mdct_calcw = ff_mdct_calcw_c;
>>>> if (ARCH_ARM) ff_fft_fixed_init_arm(s);
>>>
>>>I think that you can do one single call here like for all the other archs.
>>>
>> In the next patch that we are preparing implementation and optimization of AC3
>> fixed point decoder will be delivered. The same LUT is used
>> in this patch so I moved initialization of LUT in separate call in order for it
>> to be usable for both floating and fixed point code.
>
>It will still be MIPS specific, so can still be done in ff_fft_init_mips(),
I moved initialization of LUT out of ff_fft_init_mips because LUT is also used in
fixed point AC3 decoder (please see patch for it) and it has its own
initialization function (ff_fft_fixed_init_mips) that is called in "#else" branch
of "#if CONFIG_FFT_FLOAT" directive.
However I will change these initialization functions so there is no call for
ff_fft_lut_init outside of them.
>
>>>> diff --git a/libavcodec/mips/fft_init_table.c b/libavcodec/mips/fft_init_table.c
>>>> new file mode 100644
>>>> index 0000000..2e729e1
>>>> --- /dev/null
>>>> +++ b/libavcodec/mips/fft_init_table.c
>>>> @@ -0,0 +1,78 @@
>>>> +/*
>>>> + * Copyright (c) 2012
>>>> + * MIPS Technologies, Inc., California.
>>>> + *
>>>> + * Redistribution and use in source and binary forms, with or without
>>>> + * modification, are permitted provided that the following conditions
>>>> + * are met:
>>>> + * 1. Redistributions of source code must retain the above copyright
>>>> + * notice, this list of conditions and the following disclaimer.
>>>> + * 2. Redistributions in binary form must reproduce the above copyright
>>>> + * notice, this list of conditions and the following disclaimer in the
>>>> + * documentation and/or other materials provided with the distribution.
>>>> + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its
>>>> + * contributors may be used to endorse or promote products derived from
>>>> + * this software without specific prior written permission.
>>>> + *
>>>> + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND
>>>> + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
>>>> + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
>>>> + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE
>>>> + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
>>>> + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
>>>> + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
>>>> + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
>>>> + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
>>>> + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
>>>> + * SUCH DAMAGE.
>>>> + *
>>>> + * Author: Stanislav Ocovaj (socovaj at mips.com)
>>>> + *
>>>> + * 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
>>>> + */
>>>> +
>>>> +/**
>>>> + * @file
>>>> + * definitions and initialization of LUT table for MIPS FFT
>>>> + */
>>>> +#include "fft_table.h"
>>>> +
>>>> +short * fft_offsets_lut;
>>>
>>>Why not just
>>>
>>>static uint16_t fft_offsets_lut[0x2aab];
>>>
>>>?
>>>
>>>This way, it'll never leak and can eventually be shared between threads.
>>>
>> If I do that fft_offset_lut will have scope only in fft_init_table.c file and
>> will not be visible from fft where it is used (in fft_mips.c).
>> Since the same LUT table will be used also for fixed point fft I had to make it
>> global.
>> I can put:
>> static uint16_t fft_offsets_lut[0x2aab];
>> in fft_table.h but I wanted to avoid using static definition in header file.
>
>Sorry, that should be then
>
> uint16_t fft_offsets_lut[0x2aab];
Ok. I will use this instead of dynamic allocation in next patch.
-Nedeljko
________________________________________
From: ffmpeg-devel-bounces at ffmpeg.org [ffmpeg-devel-bounces at ffmpeg.org] on behalf of Vitor Sessak [vitor1001 at gmail.com]
Sent: Saturday, July 28, 2012 18:33
To: FFmpeg development discussions and patches
Subject: Re: [FFmpeg-devel] [PATCH] Optimization of AC3 floating point decoder for MIPS
On Wed, Jul 25, 2012 at 1:34 PM, Babic, Nedeljko <nbabic at mips.com> wrote:
>>> diff --git a/libavcodec/fft.c b/libavcodec/fft.c
>>> index 6b93a5c..8463bfb 100644
>>> --- a/libavcodec/fft.c
>>> +++ b/libavcodec/fft.c
>>> @@ -31,6 +31,7 @@
>>> #include "libavutil/mathematics.h"
>>> #include "fft.h"
>>> #include "fft-internal.h"
>>> +#include "mips/fft_table.h"
>>>
>>> /* cos(2*pi*x/n) for 0<=x<=n/4, followed by its reverse */
>>> #if !CONFIG_HARDCODED_TABLES
>>> @@ -157,11 +158,13 @@ av_cold int ff_fft_init(FFTContext *s, int nbits, int inverse)
>>> s->mdct_calc = ff_mdct_calc_c;
>>> #endif
>>>
>>> + if (ARCH_MIPS) ff_fft_lut_init();
>>> #if CONFIG_FFT_FLOAT
>>> if (ARCH_ARM) ff_fft_init_arm(s);
>>> if (HAVE_ALTIVEC) ff_fft_init_altivec(s);
>>> if (HAVE_MMX) ff_fft_init_mmx(s);
>>> if (CONFIG_MDCT) s->mdct_calcw = s->mdct_calc;
>>> + if (HAVE_MIPSFPU) ff_fft_init_mips(s);
>>> #else
>>> if (CONFIG_MDCT) s->mdct_calcw = ff_mdct_calcw_c;
>>> if (ARCH_ARM) ff_fft_fixed_init_arm(s);
>>
>>I think that you can do one single call here like for all the other archs.
>>
> In the next patch that we are preparing implementation and optimization of AC3
> fixed point decoder will be delivered. The same LUT is used
> in this patch so I moved initialization of LUT in separate call in order for it
> to be usable for both floating and fixed point code.
It will still be MIPS specific, so can still be done in ff_fft_init_mips(),
>>> diff --git a/libavcodec/mips/fft_init_table.c b/libavcodec/mips/fft_init_table.c
>>> new file mode 100644
>>> index 0000000..2e729e1
>>> --- /dev/null
>>> +++ b/libavcodec/mips/fft_init_table.c
>>> @@ -0,0 +1,78 @@
>>> +/*
>>> + * Copyright (c) 2012
>>> + * MIPS Technologies, Inc., California.
>>> + *
>>> + * Redistribution and use in source and binary forms, with or without
>>> + * modification, are permitted provided that the following conditions
>>> + * are met:
>>> + * 1. Redistributions of source code must retain the above copyright
>>> + * notice, this list of conditions and the following disclaimer.
>>> + * 2. Redistributions in binary form must reproduce the above copyright
>>> + * notice, this list of conditions and the following disclaimer in the
>>> + * documentation and/or other materials provided with the distribution.
>>> + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its
>>> + * contributors may be used to endorse or promote products derived from
>>> + * this software without specific prior written permission.
>>> + *
>>> + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND
>>> + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
>>> + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
>>> + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE
>>> + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
>>> + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
>>> + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
>>> + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
>>> + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
>>> + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
>>> + * SUCH DAMAGE.
>>> + *
>>> + * Author: Stanislav Ocovaj (socovaj at mips.com)
>>> + *
>>> + * 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
>>> + */
>>> +
>>> +/**
>>> + * @file
>>> + * definitions and initialization of LUT table for MIPS FFT
>>> + */
>>> +#include "fft_table.h"
>>> +
>>> +short * fft_offsets_lut;
>>
>>Why not just
>>
>>static uint16_t fft_offsets_lut[0x2aab];
>>
>>?
>>
>>This way, it'll never leak and can eventually be shared between threads.
>>
> If I do that fft_offset_lut will have scope only in fft_init_table.c file and
> will not be visible from fft where it is used (in fft_mips.c).
> Since the same LUT table will be used also for fixed point fft I had to make it
> global.
> I can put:
> static uint16_t fft_offsets_lut[0x2aab];
> in fft_table.h but I wanted to avoid using static definition in header file.
Sorry, that should be then
uint16_t fft_offsets_lut[0x2aab];
-Vitor
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel at ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
More information about the ffmpeg-devel
mailing list