armv7(float): Optimize encode usecase using NE10 library

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>

1 files changed, 24 insertions, 13 deletions

diff --git a/Makefile.am b/Makefile.am
index 4094e33e..4d3a8880 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -10,7 +10,7 @@ lib_LTLIBRARIES = libopus.la
 DIST_SUBDIRS = doc
 
 AM_CPPFLAGS = -I$(top_srcdir)/include -I$(top_srcdir)/celt -I$(top_srcdir)/silk \
-              -I$(top_srcdir)/silk/float -I$(top_srcdir)/silk/fixed
+              -I$(top_srcdir)/silk/float -I$(top_srcdir)/silk/fixed $(NE10_CFLAGS)
 
 include celt_sources.mk
 include silk_sources.mk
@@ -51,6 +51,10 @@ if OPUS_ARM_NEON_INTR
 CELT_SOURCES += $(CELT_SOURCES_ARM_NEON_INTR)
 endif
 
+if HAVE_ARM_NE10
+CELT_SOURCES += $(CELT_SOURCES_ARM_NE10)
+endif
+
 if OPUS_ARM_EXTERNAL_ASM
 noinst_LTLIBRARIES = libarmasm.la
 libarmasm_la_SOURCES = $(CELT_SOURCES_ARM_ASM:.s=-gnu.S)
@@ -69,7 +73,7 @@ include opus_headers.mk
 
 libopus_la_SOURCES = $(CELT_SOURCES) $(SILK_SOURCES) $(OPUS_SOURCES)
 libopus_la_LDFLAGS = -no-undefined -version-info @OPUS_LT_CURRENT@:@OPUS_LT_REVISION@:@OPUS_LT_AGE@
-libopus_la_LIBADD = $(LIBM)
+libopus_la_LIBADD = $(NE10_LIBS) $(LIBM)
 if OPUS_ARM_EXTERNAL_ASM
 libopus_la_LIBADD += libarmasm.la
 endif
@@ -85,32 +89,35 @@ TESTS = celt/tests/test_unit_types celt/tests/test_unit_mathops celt/tests/test_
 
 opus_demo_SOURCES = src/opus_demo.c
 
-opus_demo_LDADD = libopus.la $(LIBM)
+opus_demo_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 repacketizer_demo_SOURCES = src/repacketizer_demo.c
 
-repacketizer_demo_LDADD = libopus.la $(LIBM)
+repacketizer_demo_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 opus_compare_SOURCES = src/opus_compare.c
 opus_compare_LDADD = $(LIBM)
 
 tests_test_opus_api_SOURCES = tests/test_opus_api.c tests/test_opus_common.h
-tests_test_opus_api_LDADD = libopus.la $(LIBM)
+tests_test_opus_api_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 tests_test_opus_encode_SOURCES = tests/test_opus_encode.c tests/test_opus_common.h
-tests_test_opus_encode_LDADD = libopus.la $(LIBM)
+tests_test_opus_encode_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 tests_test_opus_decode_SOURCES = tests/test_opus_decode.c tests/test_opus_common.h
-tests_test_opus_decode_LDADD = libopus.la $(LIBM)
+tests_test_opus_decode_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 tests_test_opus_padding_SOURCES = tests/test_opus_padding.c tests/test_opus_common.h
-tests_test_opus_padding_LDADD = libopus.la $(LIBM)
+tests_test_opus_padding_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
 
 celt_tests_test_unit_cwrs32_SOURCES = celt/tests/test_unit_cwrs32.c
 celt_tests_test_unit_cwrs32_LDADD = $(LIBM)
 
 celt_tests_test_unit_dft_SOURCES = celt/tests/test_unit_dft.c
-celt_tests_test_unit_dft_LDADD = $(LIBM)
+celt_tests_test_unit_dft_LDADD = $(NE10_LIBS) $(LIBM)
+if OPUS_ARM_EXTERNAL_ASM
+celt_tests_test_unit_dft_LDADD += libarmasm.la
+endif
 
 celt_tests_test_unit_entropy_SOURCES = celt/tests/test_unit_entropy.c
 celt_tests_test_unit_entropy_LDADD = $(LIBM)
@@ -119,16 +126,19 @@ celt_tests_test_unit_laplace_SOURCES = celt/tests/test_unit_laplace.c
 celt_tests_test_unit_laplace_LDADD = $(LIBM)
 
 celt_tests_test_unit_mathops_SOURCES = celt/tests/test_unit_mathops.c
-celt_tests_test_unit_mathops_LDADD = $(LIBM)
+celt_tests_test_unit_mathops_LDADD = $(NE10_LIBS) $(LIBM)
 if OPUS_ARM_EXTERNAL_ASM
 celt_tests_test_unit_mathops_LDADD += libarmasm.la
 endif
 
 celt_tests_test_unit_mdct_SOURCES = celt/tests/test_unit_mdct.c
-celt_tests_test_unit_mdct_LDADD = $(LIBM)
+celt_tests_test_unit_mdct_LDADD = $(NE10_LIBS) $(LIBM)
+if OPUS_ARM_EXTERNAL_ASM
+celt_tests_test_unit_mdct_LDADD += libarmasm.la
+endif
 
 celt_tests_test_unit_rotation_SOURCES = celt/tests/test_unit_rotation.c
-celt_tests_test_unit_rotation_LDADD = $(LIBM)
+celt_tests_test_unit_rotation_LDADD = $(NE10_LIBS) $(LIBM)
 if OPUS_ARM_EXTERNAL_ASM
 celt_tests_test_unit_rotation_LDADD += libarmasm.la
 endif
@@ -286,5 +296,6 @@ endif
 
 if OPUS_ARM_NEON_INTR
 CELT_ARM_NEON_INTR_OBJ = $(CELT_SOURCES_ARM_NEON_INTR:.c=.lo)
-$(CELT_ARM_NEON_INTR_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += $(OPUS_ARM_NEON_INTR_CFLAGS)
+$(CELT_ARM_NEON_INTR_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += \
+ $(OPUS_ARM_NEON_INTR_CFLAGS)  $(NE10_CFLAGS)
 endif