similar to: [PATCH 2/2] bitmath: Finish up optimizations

Hi, attached are patches that improve decoding speed a bit. The first patch improves the bit scan macro used for decoding unary values, the second one adds a GCC inline assembly for bswap and the third patch replaces the read_rice_block function. In my testing it turned out to be even faster than the _ia32_bswap function. If the code produced by MSVC is faster as well, I'd suggest to remove

More or less detailed explanation of this patch. 1. The first parameter of _BitScanReverse() and _BitScanReverse64() is a pointer to unsigned long (4-byte int). However _BitScanReverse64() is called with a pointer to FLAC__uint64 (8-byte int). IMHO it's a bug and this patch changes the type of idx variable inside FLAC__bitmath_ilog2_wide() from FLAC__uint64 to unsigned long. The type of idx

1) FLAC supports only MSVS 2005 and newer, so (_MSC_VER >= 1400) is always true and can be removed. 2) The argument for FLAC__clz_uint32() is of FLAC__uint32 type, so FLAC__clz_soft_uint32() should have the same argument type. 3) The patch removes unnecessary parentheses around 'word' variable. 4) It also replaces "sizeof(FLAC__uint32) * CHAR_BIT - 1 -

--- src/libFLAC/bitreader.c | 445 +++++++++++------------------------------------ 1 files changed, 105 insertions(+), 340 deletions(-) diff --git a/src/libFLAC/bitreader.c b/src/libFLAC/bitreader.c index ae515a0..7ae086d 100644 --- a/src/libFLAC/bitreader.c +++ b/src/libFLAC/bitreader.c @@ -755,379 +755,144 @@ FLAC__bool FLAC__bitreader_read_rice_signed(FLAC__BitReader *br, int *val, unsig }

On Fri, Mar 14, 2008 at 07:36:31PM +0100, Miroslav Lichvar wrote: > attached are patches that improve decoding speed a bit. The first > patch improves the bit scan macro used for decoding unary values, the > second one adds a GCC inline assembly for bswap and the third patch > replaces the read_rice_block function. The third patch has a bug causing reading past input buffer, attaching

--- src/libFLAC/include/private/bitmath.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/libFLAC/include/private/bitmath.h b/src/libFLAC/include/private/bitmath.h index 61b0e03..d32b1a7 100644 --- a/src/libFLAC/include/private/bitmath.h +++ b/src/libFLAC/include/private/bitmath.h @@ -42,7 +42,7 @@ #endif /* Will never be emitted for MSVC, GCC, Intel compilers */

Josh Coalson <xflac@yahoo.com> wrote: > thanks for the patch. No prob :) > also, if you have miroslav's patch again a more updated version > of bitbuffer.c that would be great. I have been meaning to get > around to applying it for a long time. This is Miroslav's patch, from the mailing list post I dug up in the archives: --- orig/src/libFLAC/bitbuffer.c +++

lvqcl wrote: > More or less detailed explanation of this patch. I've applied part of that patch. > 1. The first parameter of _BitScanReverse() and _BitScanReverse64() is > a pointer to unsigned long (4-byte int). On windows, yes, unsigned long is 4 bytes for both 32 and 64 bit versions. This is not true for most Unices. > However _BitScanReverse64() is called with a pointer to

Ok, here is a patch waiting for new CVS :). It works fine for me, but please check it before commiting... -- Miroslav Lichvar -------------- next part -------------- --- src/libFLAC/bitbuffer.c.orig 2003-01-30 17:36:01.000000000 +0100 +++ src/libFLAC/bitbuffer.c 2003-01-30 21:53:18.000000000 +0100 @@ -51,6 +51,25 @@ */ static const unsigned FLAC__BITBUFFER_DEFAULT_CAPACITY = ((65536 - 64) *

There is a function FLAC__clz_soft_uint32 in bitmath.h: static inline unsigned int FLAC__clz_soft_uint32(unsigned int word) { static const unsigned char byte_to_unary_table[] = { 8, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, .................................... 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }; return (word) > 0xffffff ?

Hello, I had some difficulties compiling the current git ( http://git.xiph.org/?p=flac.git;a=commit;h=a7e3705d051bafd1cae90f6605287cc1d9f2a18d ) using the Ubuntu 12.04 supplied MinGW32 cross compiler: I configured the FLAC build with --host=i586-mingw32msvc --target=i586-mingw32msvc --build=i586-linux but ran into several linker problems. Are these options somehow wrong? It worked fine when

- INCLUDES is deprecated, and CPPFLAGS is an user-defined variable, use the proper AM_CPPFLAGS instead - Remove FLAC__INLINE definition, providing proper replacement for MSVC compilers. - Detect if we have C99 's lround and provide a replacement for windows... --- configure.ac | 32 ++++++++-------------------- examples/c/decode/file/Makefile.am

--- configure.ac | 7 +++++ src/libFLAC/bitreader.c | 12 ++------- src/libFLAC/bitwriter.c | 8 ++---- src/libFLAC/fixed.c | 18 +++++-------- src/libFLAC/format.c | 8 ++---- src/libFLAC/include/private/macros.h | 29 ++++++++++++++++++++ src/libFLAC/metadata_iterators.c | 17 +++---------

--- src/libFLAC/include/private/bitmath.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/libFLAC/include/private/bitmath.h b/src/libFLAC/include/private/bitmath.h index 42ce639..e5c7695 100644 --- a/src/libFLAC/include/private/bitmath.h +++ b/src/libFLAC/include/private/bitmath.h @@ -74,7 +74,7 @@ static inline unsigned int FLAC__clz_uint32(FLAC__uint32 v) { /* Never

Found this code: ftp://ftp.samba.org/pub/unpacked/ntdb/lib/ccan/ilog/ilog.c Tests show that it's faster to use the following code in FLAC__bitmath_ilog2_wide(): static const unsigned char DEBRUIJN_IDX32[32]={ 0, 1,28, 2,29,14,24, 3,30,22,20,15,25,17, 4, 8, 31,27,13,23,21,19,16, 7,26,12,18, 6,11, 5,10, 9 }; FLAC__uint32 v; int m;

rutine -> routine Also MSVC complains that FLAC__uint32* (unsigned int*) is not of the same type as unsigned long* --- a\src\libFLAC\include\private\bitmath.h 2013-08-13 13:30:24.000000000 +0400 +++ b\src\libFLAC\include\private\bitmath.h 2013-08-14 10:20:51.484053700 +0400 @@ -78,12 +78,12 @@ return _bit_scan_reverse(v) ^ 31U; #elif defined(__GNUC__) && (__GNUC__ >= 4 ||

In the precompute_partition_info_sums_ function, instead of selecting 64-bit accumulator when the signal bps is larger than 16, revert to the original approach based on partition size, but make room for few extra bits to not overflow with unusual signals where the average residual magnitude may be larger than bps. It slightly improves the performance with standard encoding levels and 16-bit files

Cristian Rodr?guez wrote: > The following changes since commit b78d8e4db10e57b8d82bb82e4e3662d5dedd7255: > > FLAC__bitmath_ilog2,FLAC__bitmath_ilog2_wide,COUNT_ZERO_MSBS: add gcc > specific optimizations (2012-03-28 15:43:48 -0300) > > are available in the git repository at: > > git://github.com/crrodriguez/flac.git master > > for you to fetch changes up

There are preprocessor definitions in bitreader.c and bitwriter.c: /* Things should be fastest when this matches the machine word size */ /* WATCHOUT: if you change this you must also change the following #defines down to SWAP_BE_WORD_TO_HOST below to match */ /* WATCHOUT: there are a few places where the code will not work unless uint32_t is >= 32 bits wide */ #define

Ben Allison wrote: > There's a more few issues for compiling on MSVC. I put together a patch. > > Here's a quick rundown: > * must use __inline keyword with static inline functions (bitmath.h) If you include "share/compath.h" you will get a inline #defined as __inline when _MSC_VER is defined. I'd prefer a single diffinition of that in one place and one place