similar to: [PATCH] for precompute_partition_info_sums_* functions

Here I attach a preliminary version of a patch for precompute_partition_info_sums_() function that should accelerate encoding of 24-bit input data. 1) SSE2, SSSE3 and AVX2 versions of this function should be updated, too 2) the patch also changes if(FLAC__bitmath_ilog2(default_partition_samples) + bps + FLAC__MAX_EXTRA_RESIDUAL_BPS < 32) into (in effect)

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

All(?) non-static functions have FLAC__ prefix. But precompute_partition_info_sums_32bit_asm_ia32_() and ..._intrin_sse2() and ..._intrin_ssse3() don't have it. This patch adds the prefix to them. -------------- next part -------------- A non-text attachment was scrubbed... Name: FLAC__prefix.patch Type: application/octet-stream Size: 6396 bytes Desc: not available Url :

Martijn van Beurden wrote: > You've exposed at least two very serious FLAC bugs, > namely a malfunctioning RICE2-partition encoder and a bug concerning > choosing verbatim frames over fixed/lpc frames. The second, not choosing verbatim frames over fixed/lpc frames is almost certainly a direct result of the first problem. The fix was changing one local variable from FLAC_uint32 to

Miroslav Lichvar wrote: > On Wed, Jul 17, 2013 at 07:45:53PM +1000, Erik de Castro Lopo wrote: > > The fix was changing one local variable from FLAC_uint32 to FLAC_uint64 > > in function precompute_partition_info_sums_(). > > > > https://git.xiph.org/?p=flac.git;a=commit;h=6f7ec60c7e7f05f5ab0b1cf6b7b0945e44afcd4b > > I don't like this fix. It will

Erik de Castro Lopo wrote: >> It slightly improves the performance with standard encoding levels and >> 16-bit files as the 17-bit side channel can still be processed with the >> 32-bit accumulator and correctly selects the 64-bit accumulator with >> very large 16-bit partitions. >> >> This is related to commits 6f7ec60c and 187e596e. > > Sorry I

Previously I wrote that precompute_partition_info_sums_32bit_asm_ia32_() only makes encoding slower. Now I managed to compile flac with GCC 4.8.1, with this function enabled and disabled. NASM was enabled, SSE intrinsics disabled. Then I added -msse option (so that all C code was compiled with -msse), then -msse2 and so on. Input file for test: 44.1kHz/16bit/stereo; best compression mode (flac

lvqcl wrote: > It doesn't know about uint32_t type, so the definition of cpu_xgetbv_x86() fails. > It can be fixed by adding "#include share/compat.h" to cpu.c (or by using > FLAC__uint32 from FLAC/ordinals.h). Ok, added share/compat.h. > When I fix this, the following problem occurs: > > error LNK2019: unresolved external symbol ___cpuidex referenced in

It's not possible to use ia32/*.nasm code in 64-bit compiles. There's still no 64-bit asm code in FLAC. I'm not familiar with asm too, so I wrote SSE-accelerated code using intrinsics. This code uses two new preprocessor macros: FLAC__CPU_X86_64 (analogous to FLAC__CPU_IA32) and FLAC__HAS_X86INTRIN (analogous to FLAC__HAS_NASM) Patch for cpu.c/cpu.h adds CPU features (sse3, ssse3)

Erik de Castro Lopo wrote: > Unfortunately I've lost/forgotten my Xiph SVN password. While I get > that sorted out for the real release, the pre-releases are here: > > http://mega-nerd.com/tmp/flac-1.3.2pre1-win.zip > http://mega-nerd.com/tmp/flac-1.3.2pre1.tar.xz Forgot to mention that I have tested this on x86_64/linux, armhf/linux and powerpc/linux. Erik --

On 16/07/13 8:10 PM, Erik de Castro Lopo wrote: > Leigh Dyer wrote: > >> Certainly -- I've uploaded the analysis files for both the -6 and -7 >> encodes, in case you wanted to compare: >> >> http://wootangent.net/~lsd/blah/6.ana >> http://wootangent.net/~lsd/blah/7.ana >> >> The encode seems to proceed normally until 59% of the way through the

Patch is here: https://github.com/openwrt/packages/commit/1263599f96f13f11d719ce336dfb6a639b32de98 . Probably needs to be modified for inclusion into mainline.

In stream_encoder.c there's the following code: #if defined FLAC__CPU_X86_64 /* and other 64-bit arch, too */ if(mean <= 0x80000000/512) { /* 512: more or less optimal for both 16- and 24-bit input */ #else if(mean <= 0x80000000/8) { /* 32-bit arch: use 32-bit math if possible */ #endif A) How to properly check for 64-bit architectures? I can check for "defined

This set of patches removes PPC/Altivec code from FLAC. I decided to split the patch into 5 parts to make it more simple: 1) removes FLAC__lpc_restore_signal_asm_ppc_altivec_16* from lpc.h and stream_decoder.c 2) removes PPC-specific code from cpu.c and cpu.h 3) removes PPC stuff from libFLAC/Makefile.lite and build/*.mk 4) removes as/gas/PPC-specific stuff from configure.ac and

Martijn van Beurden wrote: > Like I reported just before the release of 1.3.0 (mail of Fri, > 05 Apr 2013 08:25:10 +0200, to be specific), compiling on > Raspbian (Debian Wheezy, GCC 4.6) returns quite some warnings of > the type -Wcast-align. > > > CC lpc_intrin_sse2.lo > > CC lpc_intrin_sse41.lo > > CC md5.lo > > md5.c: In function

The patch does the following: 1. splits lpc_x86intrin.c to lpc_intrin_sse.c and lpc_intrin_sse2.c 2. adds FLAC__lpc_compute_residual_from_qlp_coefficients_intrin_sse2() function to lpc_intrin_sse2.c 3. adds lpc_intrin_sse41.c with two ..._wide_intrin_sse41() functions (useful for 24-bit en-/decoding) 4. adds precompute_partition_info_sums_intrin_sse2() / ...ssse3() and disables

lvqcl wrote: > > These versions of Visual Studio don't have stdint.h and > > all [u]intNN_t types. But now these types are everywhere > > in FLAC codebase. > > Here is the patch that fixes the problem: > it moves definitions of all [u]intNN_t types from share/compat.h > into new file share/msvc2005_int.h and then includes this file > into all files (via

libFLAC has several places like this: if(0 == (ptr = realloc(ptr, size))) return false; which results in memory leaks if realloc fails (the old value of ptr is lost). The patch should fix this. -------------- next part -------------- A non-text attachment was scrubbed... Name: realloc_memleak_fix.patch Type: application/octet-stream Size: 6272 bytes Desc: not available Url :

Currently cpu.h lacks FLAC__SSE_TARGET and FLAC__SSEnn_SUPPORTED macros for clang. I added them, but I cannot properly test them as I can't get compiled flac.exe under Windows (don't know how to setup clang under MSYS2). If somebody has working clang, please test this patch. Does it affect en/decoding speed? Or at least, dows it affect disassembly of functions such as

Currently apodization function is hardcoded, see commit <http://git.xiph.org/?p=flac.git;a=commitdiff;h=4e8fe85bceb245dfce07d1956b144e1cb6587c9f> FLAC is locale-dependent so "tukey(0.5)" doesn't work for locales with decimal comma. But "tukey(5e-1)" should be locale-independent so it is possible to re-add 'const char *apodization' member into