search for: sigproc_fix

--- silk/SigProc_FIX.h | 4 +++ silk/arm/SigProc_FIX_arm64.h | 46 ++++++++++++++++++++++++++++++ silk/arm/macros_arm64.h | 66 ++++++++++++++++++++++++++++++++++++++++++++ silk/macros.h | 4 +++ silk_headers.mk | 2 ++ 5 files changed, 122 insertions(+) create mode 100644...

Created corresponding unit test, and the optimization is bit exact with C function. --- silk/SigProc_FIX.h | 7 ++- silk/arm/arm_silk_map.c | 12 ++++ silk/arm/inner_prod_aligned_arm.h | 58 +++++++++++++++++++ silk/arm/inner_prod_aligned_neon_intr.c | 66 ++++++++++++++++++++++ silk/enc_API.c...

Created corresponding unit test. --- silk/LPC_analysis_filter.c | 8 +- silk/SigProc_FIX.h | 8 +- silk/arm/LPC_analysis_filter_arm.h | 60 +++++++ silk/arm/LPC_analysis_filter_neon_intr.c | 176 +++++++++++++++++++++ silk/arm/arm_silk_map.c | 14 ++ .../test_unit_optimization_LPC_analysis_filter.c...

...4:49:41 +0800 +++ opus-1.3.1/celt/modes.c 2019-05-27 17:18:26 +0800 @@ -71,7 +71,7 @@ static const unsigned char band_allocati #endif /* CUSTOM_MODES_ONLY */ #ifndef M_PI -#define M_PI 3.141592653 +#define M_PI 3.14159265358979f #endif #ifdef CUSTOM_MODES diff -Naupr opus-1.3.1-vanilla/silk/SigProc_FIX.h opus-1.3.1/silk/SigProc_FIX.h --- opus-1.3.1-vanilla/silk/SigProc_FIX.h 2018-09-26 14:49:41 +0800 +++ opus-1.3.1/silk/SigProc_FIX.h 2019-05-27 17:18:59 +0800 @@ -540,7 +540,7 @@ static OPUS_INLINE opus_int32 silk_ROR32 #define silk_max(a, b) (((a) > (b)) ? (a) : (b)) /...

...celt/arm/pitch_arm.h | 15 +++++++- celt/pitch.h | 17 ++++++-- celt/tests/test_unit_mathops.c | 2 +- celt/tests/test_unit_rotation.c | 2 +- celt_sources.mk | 3 ++ configure.ac | 56 ++++++++++++++++++++++----- silk/SigProc_FIX.h | 4 +- silk/macros.h | 4 +- 13 files changed, 205 insertions(+), 25 deletions(-) create mode 100644 celt/arm/celt_neon_intr.c diff --git a/Makefile.am b/Makefile.am index e20f7b4..0e9e120 100644 --- a/Makefile.am +++ b/Makefile.am @@ -38,6 +38,12 @@ CELT_S...

...ts in 32 bits */ +#undef MULT16_16 +static inline opus_val32 MULT16_16(opus_val16 a, opus_val16 b) +{ + int res; + __asm__( + "smulbb %0, %1, %2;\n" + : "=&r"(res) + : "r"(a), "r"(b) + ); + return res; +} + +#endif diff --git a/silk/SigProc_FIX.h b/silk/SigProc_FIX.h index cf1ab36..bccd73d 100644 --- a/silk/SigProc_FIX.h +++ b/silk/SigProc_FIX.h @@ -576,6 +576,10 @@ static inline opus_int64 silk_max_64(opus_int64 a, opus_int64 b) #include "MacroCount.h" #include "MacroDebug.h" +#ifdef ARM5E_ASM +#include "SigP...

...celt/arm/pitch_arm.h | 15 +++++++- celt/pitch.h | 17 ++++++-- celt/tests/test_unit_mathops.c | 2 +- celt/tests/test_unit_rotation.c | 2 +- celt_sources.mk | 3 ++ configure.ac | 56 ++++++++++++++++++++++----- silk/SigProc_FIX.h | 4 +- silk/macros.h | 4 +- 13 files changed, 205 insertions(+), 25 deletions(-) create mode 100644 celt/arm/celt_neon_intr.c -- 1.7.9.5

...h | 19 ------ celt_headers.mk | 1 + configure.ac | 19 ++++++ silk/NSQ.c | 55 +++++------------ silk/NSQ.h | 97 ++++++++++++++++++++++++++++++ silk/NSQ_del_dec.c | 37 +++++------- silk/SigProc_FIX.h | 4 ++ silk/arm/NSQ_neon.c | 130 +++++++++++++++++++++++++++++++++++++++++ silk/arm/NSQ_neon.h | 101 ++++++++++++++++++++++++++++++++ silk/arm/SigProc_FIX_arm64.h | 46 +++++++++++++++ silk/arm/macros_arm64.h | 66 +++++++++++++++++++++ silk/macro...

...| 20 ------- celt_headers.mk | 1 + configure.ac | 23 +++++++- silk/NSQ.c | 55 +++++------------ silk/NSQ.h | 97 ++++++++++++++++++++++++++++++ silk/NSQ_del_dec.c | 37 +++++------- silk/SigProc_FIX.h | 4 ++ silk/arm/NSQ_neon.c | 130 +++++++++++++++++++++++++++++++++++++++++ silk/arm/NSQ_neon.h | 101 ++++++++++++++++++++++++++++++++ silk/arm/SigProc_FIX_arm64.h | 46 +++++++++++++++ silk/arm/macros_arm64.h | 66 +++++++++++++++++++++ silk/macro...

...celt/x86/x86cpu.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/celt/x86/x86cpu.c b/celt/x86/x86cpu.c index 9f570af..b901bd9 100644 --- a/celt/x86/x86cpu.c +++ b/celt/x86/x86cpu.c @@ -77,9 +77,6 @@ static void cpuid(unsigned int CPUInfo[4], unsigned int InfoType) #endif -#include "SigProc_FIX.h" -#include "celt_lpc.h" - typedef struct CPU_Feature{ /* SIMD: 128-bit */ int HW_SSE2; -- 2.3.2 (Apple Git-55)

...macros_armv5e.h" #endif +#ifdef OPUS_ARM_PRESUME_AARCH64_NEON_INTR +#include "arm/macros_arm64.h" +#endif + #endif /* SILK_MACROS_H */ diff --git a/silk_headers.mk b/silk_headers.mk index c74ab81..f8bf1d2 100644 --- a/silk_headers.mk +++ b/silk_headers.mk @@ -24,6 +24,7 @@ silk/SigProc_FIX.h \ silk/x86/SigProc_FIX_sse.h \ silk/arm/macros_armv4.h \ silk/arm/macros_armv5e.h \ +silk/arm/macros_arm64.h \ silk/arm/SigProc_FIX_armv4.h \ silk/arm/SigProc_FIX_armv5e.h \ silk/arm/NSQ_neon.h \ -- 2.4.9 (Apple Git-60)

...on is bit exact with C function. This optimization speeds up SILK encoder on NEON as following. Fixed-point: Complexity 0-5: 0% Complexity 6-7: 6% Complexity 8-9: 10% Complexity 10: 8% Got similar results on floating-point. --- silk/NSQ_del_dec.c | 6 +- silk/SigProc_FIX.h | 4 +- silk/arm/NSQ_del_dec_arm.h | 88 ++ silk/arm/NSQ_del_dec_neon_intr.c | 1125 +++++++++++++++++++++++ silk/arm/arm_silk_map.c | 23 + silk/main.h | 6 +- s...

NSQ_LPC_BUF_LENGTH is independent of DECISION_DELAY. --- silk/define.h | 4 ---- 1 file changed, 4 deletions(-) diff --git a/silk/define.h b/silk/define.h index 781cfdc..1286048 100644 --- a/silk/define.h +++ b/silk/define.h @@ -173,11 +173,7 @@ extern "C" #define MAX_MATRIX_SIZE MAX_LPC_ORDER /* Max of LPC Order and LTP order */ -#if( MAX_LPC_ORDER >

--- configure.ac | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/configure.ac b/configure.ac index 90a06c8..adcb969 100644 --- a/configure.ac +++ b/configure.ac @@ -503,6 +503,26 @@ AS_IF([test x"$enable_intrinsics" = x"yes"],[ [rtcd_support="$rtcd_support (NE10)"]) ]) + OPUS_CHECK_INTRINSICS( +

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

--- configure.ac | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/configure.ac b/configure.ac index f52d2c2..e1a6e9b 100644 --- a/configure.ac +++ b/configure.ac @@ -190,7 +190,7 @@ AC_ARG_ENABLE([rtcd], [enable_rtcd=yes]) AC_ARG_ENABLE([intrinsics], - [AS_HELP_STRING([--disable-intrinsics], [Disable intrinsics optimizations for ARM(float) X86(fixed)])],, +

...+- celt/x86/x86cpu.c | 47 ++++- celt/x86/x86cpu.h | 26 ++- celt_sources.mk | 5 +- configure.ac | 313 ++++++++++++++++++----------- m4/opus-intrinsics.m4 | 29 +++ silk/x86/SigProc_FIX_sse.h | 17 ++ silk/x86/main_sse.h | 48 +++++ silk/x86/x86_silk_map.c | 25 ++- win32/VS2010/celt.vcxproj | 17 +- win32/VS2010/celt.vcxproj.filters | 27 +++ win32/VS2010/silk_common.vcxproj | 17 +- win32/VS2...

...+- celt/x86/x86cpu.c | 47 ++++- celt/x86/x86cpu.h | 26 ++- celt_sources.mk | 5 +- configure.ac | 312 ++++++++++++++++++----------- m4/opus-intrinsics.m4 | 29 +++ silk/x86/SigProc_FIX_sse.h | 17 ++ silk/x86/main_sse.h | 48 +++++ silk/x86/x86_silk_map.c | 25 ++- win32/VS2010/celt.vcxproj | 17 +- win32/VS2010/celt.vcxproj.filters | 27 +++ win32/VS2010/silk_common.vcxproj | 17 +- win32/VS2...

The attached patch cleans up Opus's x86 intrinsics configury. It: * Makes ?enable-intrinsics work with clang and other non-GCC compilers * Enables RTCD for the floating-point-mode SSE code in Celt. * Disables use of RTCD in cases where the compiler targets an instruction set by default. * Enables the SSE4.1 Silk optimizations that apply to the common parts of Silk when Opus is built in

Following Tim's comments, here are my reworked patches for the Neon intrinsic function patches of of my Aarch64 patchset, i.e. replacing patches 5-8 of the v2 series. Patches 1-4 and 9-18 of the old series still apply unmodified. The one new (as opposed to changed) patch is the first one in this series, to add named constants for the ARM architecture variants. There are also some minor code