search for: celt_lpc

.... > We write directly in ASM since typically neither compilers do what you want. > > Cliff On 6/11/2013 1:00 PM, opus-request at xiph.org wrote: > Date: Tue, 11 Jun 2013 09:31:31 +0200 > From: Aur?lien Zanelli<aurelien.zanelli at parrot.com> > Subject: Re: [opus] Bug fix in celt_lpc.c and some xcorr_kernel > optimizations > To:<opus at xiph.org> > Message-ID:<51B6D253.9030505 at parrot.com> > Content-Type: text/plain; charset="ISO-8859-1"; format=flowed > > Hi, > > I compared C version, John's versions and azanelli's versi...

...for the NEON code, it would be good to compare the performance with the code Aur?lien Zanelli posted at http://darkosphere.fr/public/0002-Add-optimized-NEON-version-of-celt_fir-celt_iir-and-.patch Cheers, Jean-Marc On 06/06/2013 08:07 PM, John Ridges wrote: > Hi JM, > > At line 221 in celt_lpc.c (the celt_iir function) I think you really > want the RESTORE_STACK statement to be before the #endif instead of > after it. Also, I couldn't help notice that your SSE code for > xcorr_kernel reads more than "len" elements of "_x". I don't know if > tha...

Hi JM, At line 221 in celt_lpc.c (the celt_iir function) I think you really want the RESTORE_STACK statement to be before the #endif instead of after it. Also, I couldn't help notice that your SSE code for xcorr_kernel reads more than "len" elements of "_x". I don't know if that's really a pro...

On 06/07/2013 02:33 PM, John Ridges wrote: > I have no doubt that Mr. Zanelli's NEON code is faster, since hand tuned > assembly is bound to be faster than using intrinsics. I was mostly curious about comparing vectorization approaches (assuming the two are different) than exact code. > However I notice > that his code can also read past the y buffer. Yeah we'd need to

Hi all, This is Linfeng Zhang from Google. I'll work on ARM NEON optimization in the next few months. I'm submitting 2 patches in the following couple of emails, which have the new created celt_fir_neon(). I revised celt_fir_c() to not pass in argument "mem" in Patch 1. If there are concerns to this change, please let me know. Many thanks to your comments. Linfeng Zhang

Unfortunately I don't have a setup that lets me easily profile ARM code, so I really can't tell which method is faster (though I suspect Mr. Zanelli's code is). Let me offer up another intrinsic version of the NEON xcorr_kernel that is almost identical to the SSE version, and more in line with Mr. Zanelli's code: static inline void xcorr_kernel_neon(const opus_val16 *x, const

...h_buf, COMBFILTER_MAXPERIOD+N, CC, st->arch); /* Don't search for the fir last 1.5 octave of the range because there's too many false-positives due to short-term correlation */ pitch_search(pitch_buf+(COMBFILTER_MAXPERIOD>>1), pitch_buf, N, diff --git a/celt/celt_lpc.c b/celt/celt_lpc.c index c75c25b..1eac65b 100644 --- a/celt/celt_lpc.c +++ b/celt/celt_lpc.c @@ -32,9 +32,21 @@ #include "celt_lpc.h" #include "stack_alloc.h" #include "mathops.h" +#include "cpu_support.h" -#ifdef ARM_HAVE_NEON +#ifdef ARM_ASM #include...

Hi JM, I have no doubt that Mr. Zanelli's NEON code is faster, since hand tuned assembly is bound to be faster than using intrinsics. However I notice that his code can also read past the y buffer. Cheers, --John On 6/6/2013 9:22 PM, Jean-Marc Valin wrote: > Hi John, > > Thanks for the two fixes. They're in git now. Your SSE version seems to > also be slightly faster than

...-point intrinsics optimization celt_fir_neon() for ARM NEON. Create test tests/test_unit_optimization to unit test the optimization. --- .gitignore | 1 + Makefile.am | 39 ++++- celt/arm/arm_celt_map.c | 17 +++ celt/arm/celt_lpc_arm.h | 65 ++++++++ celt/arm/celt_lpc_neon_intr.c | 254 ++++++++++++++++++++++++++++++++ celt/celt_lpc.h | 5 + celt/tests/test_unit_dft.c | 1 + celt/tests/test_unit_mathops.c | 1 + celt/tests/test_unit_mdct.c...

...'arch' [-Wunused-parameter] void opus_fft_free(const kiss_fft_state *cfg, int arch) ^ CCLD celt/tests/test_unit_dft celt/tests/test_unit_dft.o: In function `_celt_autocorr': /home/tterribe/src/xiph/git/xiph/opus/build/../celt/celt_lpc.c:285: undefined reference to `CELT_PITCH_XCORR_IMPL' /home/tterribe/src/xiph/git/xiph/opus/build/../celt/celt_lpc.c:285: undefined reference to `CELT_PITCH_XCORR_IMPL' celt/tests/test_unit_dft.o: In function `pitch_search': /home/tterribe/src/xiph/git/xiph/opus/build/../celt/pitch.c:...

...US_X86_MAY_HAVE_SSE4_1) # include "x86/x86cpu.c" -#elif defined(OPUS_HAVE_RTCD) && \ - (defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)) +#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR) # include "arm/armcpu.c" # include "celt_lpc.c" # include "pitch.c" diff --git a/celt/tests/test_unit_mathops.c b/celt/tests/test_unit_mathops.c index 5b446b7..fd3319d 100644 --- a/celt/tests/test_unit_mathops.c +++ b/celt/tests/test_unit_mathops.c @@ -63,8 +63,7 @@ # include "x86/celt_lpc_sse.c" # endif # includ...

...US_X86_MAY_HAVE_SSE4_1) # include "x86/x86cpu.c" -#elif defined(OPUS_HAVE_RTCD) && \ - (defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)) +#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR) # include "arm/armcpu.c" # include "celt_lpc.c" # include "pitch.c" diff --git a/celt/tests/test_unit_mathops.c b/celt/tests/test_unit_mathops.c index 5b446b7..fd3319d 100644 --- a/celt/tests/test_unit_mathops.c +++ b/celt/tests/test_unit_mathops.c @@ -63,8 +63,7 @@ # include "x86/celt_lpc_sse.c" # endif # includ...

...ise_shape_quantizer_del_dec() function. 1. I am trying to convert this entire function to assembly. I have tried to incoporate .asm file but I get linking error as shown below undefined first referenced symbol in file --------- ---------------- _celt_fir_asm vocalopus.lib<celt_lpc.obj> I have added the files in celt_sources.mk . But they are not getting compiled at all. Has any one had the same behavior. If so could you please help me. 2. I also know TI compiler is very efficient , I would like to know if writing assembly code with help in any way. Thanks and Regar...

Fixes build failure on platforms with MAY_HAVE_SSE4_1 (but not PRESUME_SSE4_1) with --enable-intrinsics. --- silk/fixed/main_FIX.h | 11 +++++++++++ silk/x86/x86_silk_map.c | 2 ++ 2 files changed, 13 insertions(+) diff --git a/silk/fixed/main_FIX.h b/silk/fixed/main_FIX.h index ffeb4f3..375b5eb 100644 --- a/silk/fixed/main_FIX.h +++ b/silk/fixed/main_FIX.h @@ -97,6 +97,17 @@ void

Fixes build failure on platforms with MAY_HAVE_SSE4_1 (but not PRESUME_SSE4_1) with --enable-intrinsics. --- silk/fixed/main_FIX.h | 11 +++++++++++ silk/x86/x86_silk_map.c | 2 ++ 2 files changed, 13 insertions(+) diff --git a/silk/fixed/main_FIX.h b/silk/fixed/main_FIX.h index ffeb4f3..375b5eb 100644 --- a/silk/fixed/main_FIX.h +++ b/silk/fixed/main_FIX.h @@ -97,6 +97,17 @@ void

...sed-parameter] > void opus_fft_free(const kiss_fft_state *cfg, int arch) > ^ > CCLD celt/tests/test_unit_dft > celt/tests/test_unit_dft.o: In function `_celt_autocorr': > /home/tterribe/src/xiph/git/xiph/opus/build/../celt/celt_lpc.c:285: > undefined reference to `CELT_PITCH_XCORR_IMPL' > /home/tterribe/src/xiph/git/xiph/opus/build/../celt/celt_lpc.c:285: > undefined reference to `CELT_PITCH_XCORR_IMPL' > celt/tests/test_unit_dft.o: In function `pitch_search': > /home/tterribe/src/xiph/git/xiph/opus...

...ing optimized macros (ex: ARMv5E). A possible side effect of loop unroll is that i don't check for odd length here. - Add NEON version of FIR filter and autocorr -- Aur?lien Zanelli Parrot SA 174, quai de Jemmapes 75010 Paris France -------------- next part -------------- diff --git a/celt/celt_lpc.c b/celt/celt_lpc.c index d2addbf..14a7839 100644 --- a/celt/celt_lpc.c +++ b/celt/celt_lpc.c @@ -33,6 +33,10 @@ #include "stack_alloc.h" #include "mathops.h" +#ifdef ARM_HAVE_NEON +#include "celt_lpc_neon.h" +#endif + void _celt_lpc( opus_val16 *_lpc...

...tch_arm.h | 4 +- celt/bands.c | 6 +- celt/celt.c | 16 +- celt/celt.h | 12 +- celt/celt_decoder.c | 6 +- celt/celt_encoder.c | 4 +- celt/celt_lpc.h | 2 +- celt/cpu_support.h | 15 +- celt/mips/celt_mipsr1.h | 2 +- celt/pitch.c | 4 +- celt/pitch.h | 19 +- celt/tests/test_unit_dft.c | 4 +- celt/te...

I rebased my previous 3 patches to the current master with minor changes. Patches 1 to 3 replace all my previous submitted patches. Patches 4 and 5 are new. Thanks, Linfeng Zhang

...tch_arm.h | 4 +- celt/bands.c | 6 +- celt/celt.c | 16 +- celt/celt.h | 12 +- celt/celt_decoder.c | 6 +- celt/celt_encoder.c | 4 +- celt/celt_lpc.h | 2 +- celt/cpu_support.h | 15 +- celt/mips/celt_mipsr1.h | 2 +- celt/pitch.c | 4 +- celt/pitch.h | 19 +- celt/tests/test_unit_dft.c | 4 +- celt/te...