Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

This makes it match the formatting of the output for ARM assembly
better, and removes some redundant repetition of the word
"intrinsics".  It also fixes the output if a compiler supports RTCD
for Neon intrinsics but not assembly.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

Enables existing Neon intrinsic optimizations to work on aarch64
targets.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Timothy B. Terriberry committed 8 years ago

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Michael Graczyk authored 8 years ago and Jean-Marc Valin committed 8 years ago

The implementation currently only codes each channel independently with no
special allocation rules.

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Jonathan Lennox authored 8 years ago and Jean-Marc Valin committed 8 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Radu Velea authored 9 years ago and Timothy B. Terriberry committed 9 years ago

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

These were causing "syntax error near unexpected token `fi'" in the
 generated configure on some systems, because they produced an

  else

  fi

 with no commands between the two.

Viswanath Puttagunta authored 9 years ago and Jean-Marc Valin committed 9 years ago

Uses NEON optimized fixed point FFT routines in NE10 library.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Viswanath Puttagunta authored 9 years ago and Jean-Marc Valin committed 9 years ago

Optimize opus encode (float only) usecase using ARM NE10
library. Mainly effects opus_fft and ctl_mdct_forward
and related functions.

This optimization can be used for ARM CPUs that have NEON
VFP unit. This patch only enables optimizations for ARMv7.

Official ARM NE10 library page available at
http://projectne10.github.io/Ne10/



To enable this optimization, use
--enable-intrinsics --with-NE10=<install_prefix>
or
--enable-intrinsics --with-NE10-libraries=<NE10_lib_dir> --with-NE10-includes=<NE10_includes_dir>

Compile time checks made during configure process to make sure
optimization option available only when compiler supports NEON
instrinsics.

Runtime checks made to make sure optimized functions only called
on appropriate hardware.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Jonathan Lennox authored 9 years ago and Jean-Marc Valin committed 9 years ago

Enable x86 intrinsics when building in floating-point mode.
Support SSE as an arch value.
Use RTCD to conditionally enable existing floating-point Celt SSE code.
Call functions directly (without RTCD) when their architecture can be presumed.
Use SSE4.1 intrinsics optimized code for Silk even in floating-point mode.

Jonathan Lennox authored 9 years ago and Jean-Marc Valin committed 9 years ago

Move SSE2 and SSE4.1 intrinsics functions to separate files, to be compiled with appropriate compiler flags. Otherwise, compilers are allowed to take advantage of (e.g.) -msse4.1 to generate code that uses SSE4.1 instructions, even when no SSE4.1 intrinsics are explicitly used in the source.

Jonathan Lennox authored 9 years ago and Jean-Marc Valin committed 9 years ago

Actually try to compile intrinsics rather than using the output of --help.
Allow caller of configure script to set custom compiler options to enable intrinsics.
Detect when intrinsics are always available, without needing special compiler options.
Make naming of #defines for detected intrinsics support more systematic.

Viswanath Puttagunta authored 10 years ago and Timothy B. Terriberry committed 10 years ago

Optimize celt_pitch_xcorr function (for floating point)
using ARM NEON intrinsics for SoCs that have NEON VFP unit.

To enable this optimization, use --enable-intrinsics
configure option.

Compile time and runtime checks are also supported to make sure
this optimization is only enabled when the compiler supports
NEON intrinsics.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

It was under-quoted in c95c9a04, which
leads to a flood of warnings when it gets incorrectly expanded, like:

configure.ac:354: warning: AC_LANG_CONFTEST: no AC_LANG_SOURCE call detected in body

xiangmingzhu authored 10 years ago and Jean-Marc Valin committed 10 years ago

1. Only for fixed point on x86 platform (32bit and 64bit, uses SIMD
   intrinsics up to SSE4.2)
2. Use "configure --enable-fixed-point --enable-intrinsics" to enable
   optimization, default is disabled.
3. Official test cases are verified and passed.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Martin Storsjö authored 11 years ago and Timothy B. Terriberry committed 11 years ago

This allows building the arm assembly for iOS.

This checks for the __APPLE__ preprocessor built-in define to
determine whether this extra handling should be enabled.

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Peter Korsgaard authored 11 years ago and Jean-Marc Valin committed 11 years ago

Breaks configure when /bin/sh isn't bash with:

configure: Trying to force-enable ARMv6 media instructions...
checking if assembler supports ARMv6 media instructions on ARM... yes
configure: Trying to force-enable NEON instructions...
checking if assembler supports NEON instructions on ARM... yes
./configure.lineno: 12799: Bad substitution

Fix it by using the %% expansion to remove everything from the first
space instead.

Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

This reverts commit 2446445b.

Timothy B. Terriberry authored 11 years ago and Jean-Marc Valin committed 11 years ago

Optimizing celt_pitch_xcorr()/xcorr_kernel() which also speeds up
FIRs, IIRs and auto-correlations

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

This avoids at least one case where ./autogen.sh && ./configure && make
will re-run configure because the makefile rules updated something that
it depends upon. Pulling a new version from git will change the version
so we should update that at the first step rather than iterating after
the last one.

It only existed to not include -lm in the .pc for fixed point builds,
but that is still needed since the float API is still enabled and
will use at least lrint.

Drop the test for getopt, it's not used anywhere anymore.

Switch the last uses of AC_TRY_COMPILE to AC_COMPILE_IFELSE now.
The former is marked as obsolete, and this will leave no confusion
about which to cut and paste if new tests are added.

Double quote all the parameters to AC_LANG_SOURCE and AC_LANG_PROGRAM.
This is actually required, even if you can get away with not doing it
sometimes, so again set a good example for future changes to follow,
to hopefully avoid people getting bitten harder than they need to be.

Don't bother checking for alloca if we're never going to use it
(ie. if we have C99 variable-size array support).
The test for this is a bit sketchy anyway ...  we separately test for
HAVE_ALLOCA_H and USE_ALLOCA, but the test for USE_ALLOCA depends upon
having alloca.h present, yet the use of these macros in stack_alloc.h
only tests for HAVE_ALLOCA_H inside of a test for USE_ALLOCA.
I'm not going to change this logic right now, since I don't know what
crazy system it was attempting to cater for, though I suspect it was
one that was not using the autoconf build system ... since with the
current test that combination should not be possible to obtain.

Use LT_LIB_M instead of the song and dance with testing for exp().
This should also work for BeOS which is what the exp test was added for.
It also means we don't unconditionally add -lm to everything via LIBS.

Use LIBM now instead of hardcoding -lm everywhere.

Use AS_HELP_STRING to format all option descriptions.

Don't bother to test for doxygen if using it is --disable'd.

Drop the SYMBOL_VISIBILITY export, it isn't used anywhere (we add the
compiler flag to CFLAGS).

These were probably cribbed from libogg, but we don't use them here,
opus_types.h instead has a list of hardcoded arch definitions.

Aurélien Zanelli authored 11 years ago and Timothy B. Terriberry committed 11 years ago

Run-time CPU detection (RTCD) is enabled by default if target platform support
it.
It can be disable at compile time with --disable-rtcd option.

Add RTCD support for ARM architecture.

Thanks to Timothy B. Terriberry for help and code review

Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org>

Define ARMv4_ASM to 1 like the other ARM defines.

Needed by commit 972a34ec.

Use autoreconf in autogen.sh instead of the handwritten version,
it's simpler, and also updates things that we weren't handling.

Drop the hand-written INSTALL file.  Its information content was
~zero, and autotools wants to overwrite it with its own version,
so don't fight that, just .gitignore it.

Original patch by Aurélien Zanelli <aurelien.zanelli@parrot.com>:
 http://lists.xiph.org/pipermail/opus/2013-May/002078.html

Revised version:
- Add autconf detection (ported from libtheora).
- Rename ARM5E to ARMv5E (an ARM5 is not the same thing as ARMv5!).
- Use actual macros so they can still be selectively overridden.
- Split out ARMv4 parts and add a few more ARMv4 macros.
- Label blocks to make them easy to find in generated assembly.
- Fix MULT16_32_Q15() so we can pass make check.
  The MDCT test passes in values larger than 2**30 for b.
  The new version should be just as fast (or faster, since it's
   easier to merge the shift with following instructions), and
   there's no appreciable impact on accuracy (FFT/MDCT SNR actually
   goes up in most cases).
- Fix register constraints.
  We were using early-clobber flags in a bunch of places that
   didn't need them, and commutative-pair flags in a bunch of
   places that weren't actually commutative.
  This was Jean-Marc's fault (the original code came from Speex).
- Simplify silk_CLZ16().
- Port over iFFT C_MULC asm by Andree Buschmann
   <AndreeBuschmann@t-online.de> from Rockbox.
- Speed up the C_MULC asm by using LDRD, allowing more flexible
   addressing, re-ordering instructions to avoid some stalls,
   allowing more flexible register allocation, and getting things
   out of the inline asm block so the compiler can schedule them
   better.
- Add C_MUL and C_MUL4 asm for the FFT to the encoder based, on the
   new C_MULC.

In total, this patch gives a 22.3% speed-up on test_opus_encoder on
 a 600 MHz Cortex A8 using gcc 4.2.1,
When restricted to ARMv4 optimizations, it gives a 9.6% speed-up
 on the same processor/compiler.
On the conformance test vectors:
 Average mono quality is 97.0583 %
 Average stereo quality is 97.775 %

We shouldn't ever have any trailing newlines that need trimming here,
and the _s version wasn't added to m4sugar.m4 until autoconf 2.63b,
so this will let it work with 2.13 again.

This one meets or exceeds the following requirements:

 - Version is checked/updated for every build action when in the git repo.
   Does not require the user to re- ./configure to get the correct version.

 - Version is not updated automatically when using exported tarball source.
   Avoids accidentally getting a wrong version from some other git repo in
   a parent directory of the source, and allows setting the correct version
   for distro package exports.

 - Automatic updating can be manually suppressed.
   For developers doing lots of change/rebuild cycles they don't plan to
   release, when they don't want a full rebuild triggered for every commit,
   and again for every change made immediately after a commit.
   The version will still always be updated if they do a `make dist`.

 - Does not require any manual updating of versions in the mainline git
   repo for each release aside from normal tagging.  The version is
   recorded in one file only, that is automatically generated and will
   never need to be committed.

 - Does not require gnu-make features for the autoconf builds.

It does not currently:

 - Keep a checksum of every source file in tarball releases to mangle the
   version if people modify the tarball source.  Responsible people can
   manually update the version easily though in such cases.

The version.mk file is now only used by the VC project files.  Once they
are updated to use the package_version file too, then it can be deleted
from the repository.