Josh Coalson
|
0e3576e771
|
more changes to get conditionals/defines working
|
2001-05-25 00:07:51 +00:00 |
|
Josh Coalson
|
c8ef8353bf
|
add memory.c, fix to work with i386 convenience library (thanks mdz)
|
2001-05-25 00:06:33 +00:00 |
|
Josh Coalson
|
45d7bfc891
|
change .nasm extension to .s
|
2001-05-25 00:05:47 +00:00 |
|
Josh Coalson
|
5469112ee3
|
oops, fix bug with error_scale
|
2001-05-25 00:05:15 +00:00 |
|
Josh Coalson
|
a3f7c2c76c
|
gate cpu-specific code with FLAC__NO_ASM
|
2001-05-25 00:04:45 +00:00 |
|
Josh Coalson
|
febdca986a
|
remove config.h
|
2001-05-25 00:04:08 +00:00 |
|
Josh Coalson
|
a2fd3b8390
|
renamed .nasm file to .s to be compatible with libtool
|
2001-05-25 00:02:53 +00:00 |
|
Josh Coalson
|
6a0283a009
|
config.h no longer needed
|
2001-05-24 23:57:08 +00:00 |
|
Josh Coalson
|
8090b3610a
|
kludge in nasm support
|
2001-05-24 19:30:08 +00:00 |
|
Josh Coalson
|
034d38e73a
|
gate i386 code with FLAC__HAS_NASM
|
2001-05-24 19:29:30 +00:00 |
|
Josh Coalson
|
2d071a6b6e
|
remove debug printouts
|
2001-05-24 19:27:31 +00:00 |
|
Josh Coalson
|
a1b53c4dfb
|
be more lenient with negative lpc_errors due to accuracy problems with float
|
2001-05-24 19:27:08 +00:00 |
|
Josh Coalson
|
acc7fe5742
|
add cident macro for defining function entry points
|
2001-05-24 19:26:02 +00:00 |
|
Josh Coalson
|
51e65a9e31
|
add FLAC__lpc_compute_expected_bits_per_residual_sample_with_error_scale()
|
2001-05-24 19:25:33 +00:00 |
|
Josh Coalson
|
def8eb85b3
|
switch from AC_CANONICAL_TARGET to AC_CANONICAL_HOST
|
2001-05-23 22:08:27 +00:00 |
|
Josh Coalson
|
b985bbc4a9
|
add cpuinfo + runtime asm routine selection
|
2001-05-23 21:59:52 +00:00 |
|
Josh Coalson
|
126fecc2f7
|
include miroslav's mmx lpc_restore routine
|
2001-05-23 21:42:56 +00:00 |
|
Josh Coalson
|
cf030c8576
|
discriminate by cpu, check for nasm, add option for disabling asm optimizations
|
2001-05-23 20:59:48 +00:00 |
|
Josh Coalson
|
a2e49a99cb
|
use new conditional names
|
2001-05-23 20:58:57 +00:00 |
|
Josh Coalson
|
e02e8babf2
|
add asm routines
|
2001-05-23 20:58:14 +00:00 |
|
Josh Coalson
|
cf30f50db1
|
use cpu info to select which routine to execute
|
2001-05-23 20:57:44 +00:00 |
|
Josh Coalson
|
44258bf30c
|
update to use new asm routine naming
|
2001-05-23 20:56:30 +00:00 |
|
Josh Coalson
|
24ef83a0fa
|
add cpu.h
|
2001-05-23 20:56:09 +00:00 |
|
Josh Coalson
|
45df9ca89d
|
make the asm routine names more descriptive
|
2001-05-23 20:54:41 +00:00 |
|
Josh Coalson
|
fac1c0b294
|
add config.h
|
2001-05-23 20:53:51 +00:00 |
|
Josh Coalson
|
edade2ced1
|
initial import
|
2001-05-23 20:53:31 +00:00 |
|
Josh Coalson
|
8d1676c001
|
soup up FLAC__fixed_compute_best_predictor_asm
|
2001-05-22 23:33:40 +00:00 |
|
Josh Coalson
|
7d297060b8
|
oops, MMX moves are MOVQ not MOVAPS
|
2001-05-21 23:49:11 +00:00 |
|
Josh Coalson
|
a16d8ad55d
|
fix compiler warnings about comparing signed with unsigned
|
2001-05-21 23:43:35 +00:00 |
|
Josh Coalson
|
f5490343ba
|
soup up, fix bugs, replace PSHUFW with pre-Katmai-friendly PSRLQ mm,32
|
2001-05-21 23:40:16 +00:00 |
|
Josh Coalson
|
d1299d9626
|
tighten loops using "Jcc NEAR"
|
2001-05-21 22:12:55 +00:00 |
|
Josh Coalson
|
2cfb9ee952
|
remove some redundant instructions
|
2001-05-21 22:03:32 +00:00 |
|
Josh Coalson
|
820bbd63b8
|
initial import
|
2001-05-21 21:51:10 +00:00 |
|
Josh Coalson
|
81241a9fb9
|
syntax fixes to the SSE version
|
2001-05-21 21:50:50 +00:00 |
|
Josh Coalson
|
2d9d31e854
|
fix syntax error in FLAC__fixed_compute_best_predictor_slow()
|
2001-05-21 21:28:12 +00:00 |
|
Josh Coalson
|
1fd9b17163
|
incorporate mdz patches to fix log(0) and unsigned-goes-negative bugs
|
2001-05-21 18:03:14 +00:00 |
|
Josh Coalson
|
eee20a5fdd
|
streamline FLAC__fixed_compute_best_predictor_slow()
|
2001-05-18 18:49:19 +00:00 |
|
Josh Coalson
|
d80c18e4ac
|
streamline FLAC__fixed_compute_best_predictor()
|
2001-05-18 18:47:55 +00:00 |
|
Josh Coalson
|
675effa90e
|
add miroslav's version of FLAC__lpc_restore_signal, add SSE version of FLAC__lpc_compute_autocorrelation
|
2001-05-18 18:46:40 +00:00 |
|
Josh Coalson
|
7e40e6dad9
|
use float instead of double for basic FP type
|
2001-05-16 19:28:54 +00:00 |
|
Josh Coalson
|
5157f7c1c3
|
fix bug where gcc gets shifting wrong
|
2001-05-16 19:28:12 +00:00 |
|
Josh Coalson
|
183019474f
|
fix deallocation bugs
|
2001-05-16 19:25:33 +00:00 |
|
Josh Coalson
|
0c0fed693c
|
revert to original FLAC__fixed_compute_residual()
|
2001-05-16 19:24:59 +00:00 |
|
Josh Coalson
|
376807d911
|
fix LOCAL_FABS bug
|
2001-05-16 19:23:35 +00:00 |
|
Josh Coalson
|
9de0942d28
|
fix section defs
|
2001-05-16 19:23:13 +00:00 |
|
Josh Coalson
|
a151f2c1d4
|
minor wording
|
2001-05-14 21:44:29 +00:00 |
|
Josh Coalson
|
b787e2afeb
|
don't try and outsmart the compiler on constant multplies
|
2001-05-14 21:31:50 +00:00 |
|
Josh Coalson
|
c59638ea49
|
use local fabs() routine
|
2001-05-14 21:25:34 +00:00 |
|
Josh Coalson
|
2d5aae0af7
|
remove debug printouts
|
2001-05-14 19:34:12 +00:00 |
|
Josh Coalson
|
9f5073872b
|
minor tweaks
|
2001-05-14 19:33:42 +00:00 |
|