similar to: [PATCH 2/4]

Erik de Castro Lopo wrote: >> This patch adds new SSE code: it's simpler but not faster so >> this new version is disabled by default. Maybe it will be faster >> on newer CPUs though... > > I'm a little sceptical about adding code that by default is not > being compiled. FLAC__fixed_compute_residual() has the following code: #if 1 /* OPT: may be faster with

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

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

Hi all, People watching the git commits might have noticed that I have been fixing a number of issues around undefined behaviour. Why you ask? * Some forms of undefined behaviour have potential for security exploits. * Compiler writers are free to replace anything which invokes UB with a NOP or even, nothing at all. * Having large numbers of UB warnings makes it difficult (or rather time

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: >> part 1: fixes >> >> part 2: new code >> >> Any comments? > > Applied cpu_part2.patch. Thanks! Thanks. I hope that this code will help to avoid "bug" reports such as http://sourceforge.net/p/flac/bugs/409/ (it seems that the author of this report compiled FLAC without --enable-sse option). OTOH, SSE support is

First off, this code is horrible to read and work on. The recent commits are the first of what I hope is a massive clean up of this code. lvqcl wrote: > So if I understand things correctly, the current meaning of --(en|dis)able-sse is: > > on Linux: > --enable-sse: > add -msse2 to the compiler switches > do not test SSE OS support (assume that SSE is

This patch increases speed of FLAC__lpc_restore_signal_wide_intrin_sse41 (decoding of 24-bit FLAC files for 32-bit platform). -------------- next part -------------- A non-text attachment was scrubbed... Name: lpc_sse4.zip Type: application/zip Size: 3310 bytes Desc: not available Url : http://lists.xiph.org/pipermail/flac-dev/attachments/20140920/a3d8efb4/attachment.zip

lvqcl wrote: > Found a bug in Android OS SSE test. Sorry, what is the bug and how does this fix it? Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/

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

lvqcl wrote: > lvqcl wrote: > > > 1) in config.h FLAC__HAS_X86INTRIN macro is always defined and empty, > > even if x86intrin.h is not available. > > > > 2) sse_os is defined as 'yes' or 'no', but AM_CONDITIONAL tests it for 'true': > > The patch is attached. Please check it. Looks good. I need to do a little testing. > > 3)

lvqcl wrote: > FLAC__fixed_restore_signal() is a part of flac *de*coder, so it makes > sense to test it on different architectures (ARM, MIPS?). But I have no > idea how to make it. Yes, but the code in FLAC__fixed_restore_signal() is (ignoring identifier name differences) identical to FLAC__fixed_compute_residual(), Maybe I should just publish my micro benchmarking code as part of the

This patch removes FLAC__lpc_restore_signal_16_intrin_sse2(). It's faster than C code, but not faster than MMX-accelerated ASM functions. It's also slower than the new SSE4.1 functions that were added by the previous patch. So this function wasn't very useful before, and now it's even less useful. I don't see a reason to keep it. -------------- next part -------------- A

lvqcl wrote: > 1) if FLAC__SSE_OS==1 then MSVC cannot compile current cpu.c: > "#include <windows.h>" line is inactive and MSVC complains: > cpu.c(278): error C2065: 'EXCEPTION_EXECUTE_HANDLER': undeclared identifier > cpu.c(279): error C2065: 'STATUS_ILLEGAL_INSTRUCTION': undeclared identifier > > 2) the current code (simplified): > >

Erik de Castro Lopo wrote: > I actually think my change is correct. It should be easier to see if you > look at the github version of the change: > > https://github.com/xiph/flac/commit/e120037f3c67b23fd9eef7ccd04d2df57fa1a9a6#diff-9f048b83ff55071de36263cf0f403b2eL209 > FLAC__NO_SSE_OS was never defined anywhere, so I think that the following changes should be made:

Erik de Castro Lopo wrote: > lvqcl.mail wrote: >> now. Removing OS check will greatly simplify src/libFLAC/cpu.c. > > That makes sense. Should I post a patch that removes OS check and keeps only CPU check? >> 2. >> "configure" build system adds -msse2 option by default. It means that >> x86 (32-bit) library won't work on older, non-SSE2

1. A program can use SSE instructions only if both CPU and OS support SSE. Currently libFLAC tests both CPU and OS for this support, but is it really necessary? Maybe CPU check is enough? Operating systems that don't support SSE (Win95, WinNT 4.0, Linux kernel 2.2 (iirc), ...) are really outdated now. Removing OS check will greatly simplify src/libFLAC/cpu.c. 2. "configure" build

Currently the only way to compile FLAC using GCC w/o SSE support is to disable asm optimizations (see configure.ac): if test "x$asm_optimisation" = "xyes" ; then XIPH_ADD_CFLAGS([-msse2]) fi Also it's not possible to enable SSE4.1 intrinsic functions even with -msse4.1 option. The patch fixes both problems. --------------- BTW: I'm not sure that share/compat.h

1) in config.h FLAC__HAS_X86INTRIN macro is always defined and empty, even if x86intrin.h is not available. 2) sse_os is defined as 'yes' or 'no', but AM_CONDITIONAL tests it for 'true': AM_CONDITIONAL(FLaC__SSE_OS, test "x$sse_os" = xtrue) It seems that it should be changed to AM_CONDITIONAL(FLaC__SSE_OS, test "x$sse_os" = xyes) 3) configure

lvqcl wrote: > I have a couple of questions: > > 1) Do you plan to release 1.3.1 pre1, pre2 etc or just 1.3.1 w/o any pre-releases? I had not planned to do a pre-release. > 2) Do you plan to release any official binaries (flac, metaflac, maybe something else)? Nor had I planned to release binaries. The source code tarball ends up here: https://svn.xiph.org/releases/flac/ I