Rich Felker [Thu, 12 Apr 2012 04:25:52 +0000 (00:25 -0400)]
remove magic numbers from floatscan
Rich Felker [Thu, 12 Apr 2012 04:16:01 +0000 (00:16 -0400)]
optimize more integer cases in floatscan; comment the whole procedure
Rich Felker [Thu, 12 Apr 2012 03:08:50 +0000 (23:08 -0400)]
revert invalid optimization in floatscan
Rich Felker [Thu, 12 Apr 2012 01:29:12 +0000 (21:29 -0400)]
fix stupid typo in floatscan that caused excess rounding of some values
Rich Felker [Thu, 12 Apr 2012 00:25:42 +0000 (20:25 -0400)]
add some more useful suggested options to config.mak template
Rich Felker [Wed, 11 Apr 2012 18:59:36 +0000 (14:59 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Wed, 11 Apr 2012 18:51:08 +0000 (14:51 -0400)]
optimize floatscan downscaler to skip results that won't be needed
when upscaling, even the very last digit is needed in cases where the
input is exact; no digits can be discarded. but when downscaling, any
digits less significant than the mantissa bits are destined for the
great bitbucket; the only influence they can have is their presence
(being nonzero). thus, we simply throw them away early. the result is
nearly a 4x performance improvement for processing huge values.
the particular threshold LD_B1B_DIG+3 is not chosen sharply; it's
simply a "safe" distance past the significant bits. it would be nice
to replace it with a sharp bound, but i suspect performance will be
comparable (within a few percent) anyway.
Rich Felker [Wed, 11 Apr 2012 18:20:45 +0000 (14:20 -0400)]
simplify/debloat radix point alignment code in floatscan
now that this is the first operation, it can rely on the circular
buffer contents not being wrapped when it begins. we limit the number
of digits read slightly in the initial parsing loops too so that this
code does not have to consider the case where it might cause the
circular buffer to wrap; this is perfectly fine because KMAX is chosen
as a power of two for circular-buffer purposes and is much larger than
it otherwise needs to be, anyway.
these changes should not affect performance at all.
Rich Felker [Wed, 11 Apr 2012 18:11:47 +0000 (14:11 -0400)]
optimize floatscan: avoid excessive upscaling
upscaling by even one step too much creates 3-29 extra iterations for
the next loop. this is still suboptimal since it always goes by 2^29
rather than using a smaller upscale factor when nearing the target,
but performance on common, small-magnitude, few-digit values has
already more than doubled with this change.
more optimizations on the way...
Rich Felker [Wed, 11 Apr 2012 04:26:41 +0000 (00:26 -0400)]
fix incorrect initial count in shgetc when data is already buffered
Rich Felker [Wed, 11 Apr 2012 04:18:57 +0000 (00:18 -0400)]
fix bug parsing lone zero followed by junk, and hex float over-reading
Rich Felker [Wed, 11 Apr 2012 03:41:54 +0000 (23:41 -0400)]
fix float scanning of certain values ending in zeros
for example, "
1000000000" was being read as "1" due to this loop
exiting early. it's necessary to actually update z and zero the
entries so that the subsequent rounding code does not get confused;
before i did that, spurious inexact exceptions were being raised.
Rich Felker [Wed, 11 Apr 2012 03:05:16 +0000 (23:05 -0400)]
fix potential overflow in exponent reading
note that there's no need for a precise cutoff, because exponents this
large will always result in overflow or underflow (it's impossible to
read enough digits to compensate for the exponent magnitude; even at a
few nanoseconds per digit it would take hundreds of years).
Rich Felker [Wed, 11 Apr 2012 02:38:21 +0000 (22:38 -0400)]
set errno properly when parsing floating point
Rich Felker [Wed, 11 Apr 2012 01:47:37 +0000 (21:47 -0400)]
add "scan helper getc" and rework strtod, etc. to use it
the immediate benefit is a significant debloating of the float parsing
code by moving the responsibility for keeping track of the number of
characters read to a different module.
by linking shgetc with the stdio buffer logic, counting logic is
defered to buffer refill time, keeping the calls to shgetc fast and
light.
in the future, shgetc will also be useful for integrating the new
float code with scanf, which needs to not only count the characters
consumed, but also limit the number of characters read based on field
width specifiers.
shgetc may also become a useful tool for simplifying the integer
parsing code.
Rich Felker [Wed, 11 Apr 2012 00:25:06 +0000 (20:25 -0400)]
unify strtof/strtod/strtold wrappers and fix initial whitespace issue
Rich Felker [Tue, 10 Apr 2012 15:52:55 +0000 (11:52 -0400)]
new floating point parser/converter
this version is intended to be fully conformant to the ISO C, POSIX,
and IEEE standards for conversion of decimal/hex floating point
strings to float, double, and long double (ld64 or ld80 only at
present) values. in particular, all results are intended to be rounded
correctly according to the current rounding mode. further, this
implementation aims to set the floating point underflow, overflow, and
inexact flags to reflect the conversion performed.
a moderate amount of testing has been performed (by nsz and myself)
prior to integration of the code in musl, but it still may have bugs.
so far, only strto(d|ld|f) use the new code. scanf integration will be
done as a separate commit, and i will add implementations of the wide
character functions later.
Rich Felker [Mon, 9 Apr 2012 20:22:05 +0000 (16:22 -0400)]
fix alloca issue in stdlib.h too
I forgot _GNU_SOURCE also has it declared here...
Rich Felker [Mon, 9 Apr 2012 19:06:58 +0000 (15:06 -0400)]
alloca cannot be a function. #define it to the gcc builtin if possible
gcc makes this mapping by default anyway, but it will be disabled by
-fno-builtin (and presumably by -std=c99 or similar). for the main
program the error will be reported by the linker, and the issue can
easily be fixed, but for dynamic-loaded so files, the error cannot be
detected until dlopen time, at which point it has become very obscure.
nsz [Wed, 4 Apr 2012 15:34:28 +0000 (17:34 +0200)]
math: fix x86 asin accuracy
use (1-x)*(1+x) instead of (1-x*x) in asin.s
the later can be inaccurate with upward rounding when x is close to 1
Rich Felker [Wed, 4 Apr 2012 04:37:33 +0000 (00:37 -0400)]
work around nasty gcc bug in the i386 syscall asm
when the "r" (register) constraint is used to let gcc choose a
register, gcc will sometimes assign the same register that was used
for one of the other fixed-register operands, if it knows the values
are the same. one common case is multiple zero arguments to a syscall.
this horribly breaks the intended usage, which is swapping the GOT
pointer from ebx into the temp register and back to perform the
syscall.
presumably there is a way to fix this with advanced usage of register
constaints on the inline asm, but having bad memories about hellish
compatibility issues with different gcc versions, for the time being
i'm just going to hard-code specific registers to be used. this may
hurt the compiler's ability to optimize, but it will fix serious
miscompilation issues.
so far the only function i know what compiled incorrectly is
getrlimit.c, and naturally the bug only applies to shared (PIC)
builds, but it may be more extensive and may have gone undetected..
Rich Felker [Tue, 3 Apr 2012 23:16:51 +0000 (19:16 -0400)]
remove useless (at best, harmful) feature test checks in aio.h
Rich Felker [Mon, 2 Apr 2012 03:22:16 +0000 (23:22 -0400)]
improve name lookup performance in corner cases
the buffer in getaddrinfo really only matters when /etc/hosts is huge,
but in that case, the huge number of syscalls resulting from a tiny
buffer would seriously impact the performance of every name lookup.
the buffer in __dns.c has also been enlarged a bit so that typical
resolv.conf files will fit fully in the buffer. there's no need to
make it so large as to dominate the syscall overhead for large files,
because resolv.conf should never be large.
Rich Felker [Sat, 31 Mar 2012 03:41:43 +0000 (23:41 -0400)]
optimize signbit macro
Rich Felker [Sat, 31 Mar 2012 03:33:00 +0000 (23:33 -0400)]
make math.h more c++-friendly
nsz [Thu, 29 Mar 2012 12:09:57 +0000 (14:09 +0200)]
math: minor cleanups in ceil and floor
nsz [Thu, 29 Mar 2012 12:05:16 +0000 (14:05 +0200)]
math: remove x86 modf asm
the int part was wrong when -1 < x <= -0 (+0.0 instead of -0.0)
and the size and performace gain of the asm version was negligible
nsz [Thu, 29 Mar 2012 12:03:18 +0000 (14:03 +0200)]
math: rewrite modf.c and clean up modff.c
cleaner implementation with unions and unsigned arithmetic
nsz [Wed, 28 Mar 2012 21:51:09 +0000 (23:51 +0200)]
math: fix modfl.c bug
modfl(+-inf) was wrong on ld80 because the explicit msb
was not taken into account during inf vs nan check
nsz [Tue, 27 Mar 2012 20:49:37 +0000 (22:49 +0200)]
math: fix a regression in powl and do some cleanups
previously a division was accidentally turned into integer div
(w = -i/NXT;) instead of long double div (w = -i; w /= NXT;)
nsz [Tue, 27 Mar 2012 20:17:36 +0000 (22:17 +0200)]
math: add dummy tgamma and tgammaf implementations
nsz [Tue, 27 Mar 2012 20:12:20 +0000 (22:12 +0200)]
math: remove comment about aliasing lgamma as gamma
It is probably not worth supporting gamma.
(it was already deprecated in 4.3BSD)
nsz [Tue, 27 Mar 2012 20:01:21 +0000 (22:01 +0200)]
math: fix typo in i386 remquof and remquol asm
(fldl instruction was used instead of flds and fldt)
nsz [Mon, 26 Mar 2012 11:47:31 +0000 (13:47 +0200)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Sun, 25 Mar 2012 04:42:51 +0000 (00:42 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Sun, 25 Mar 2012 04:21:20 +0000 (00:21 -0400)]
add strfmon_l variant (still mostly incomplete)
Rich Felker [Sat, 24 Mar 2012 21:46:42 +0000 (17:46 -0400)]
update COPYRIGHT status of TRE regex code
Rich Felker [Sat, 24 Mar 2012 21:43:07 +0000 (17:43 -0400)]
update README to remove information no longer relevant as of 0.8.7
nsz [Fri, 23 Mar 2012 10:16:56 +0000 (11:16 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Fri, 23 Mar 2012 05:52:49 +0000 (01:52 -0400)]
asm for hypot and hypotf
special care is made to avoid any inexact computations when either arg
is zero (in which case the exact absolute value of the other arg
should be returned) and to support the special condition that
hypot(±inf,nan) yields inf.
hypotl is not yet implemented since avoiding overflow is nontrivial.
Rich Felker [Fri, 23 Mar 2012 04:28:20 +0000 (00:28 -0400)]
make dlerror conform to posix
the error status is required to be sticky after failure of dlopen or
dlsym until cleared by dlerror. applications and especially libraries
should never rely on this since it is not thread-safe and subject to
race conditions, but glib does anyway.
nsz [Fri, 23 Mar 2012 00:26:04 +0000 (01:26 +0100)]
minor rintl.c fix: remove unsupported ldbl format message
nsz [Fri, 23 Mar 2012 00:18:12 +0000 (01:18 +0100)]
fix tgammal: don't set the signgam global
(tgamma must be thread-safe, signgam is for lgamma* functions)
Rich Felker [Fri, 23 Mar 2012 00:00:58 +0000 (20:00 -0400)]
simplify creal and cimag macros
Rich Felker [Thu, 22 Mar 2012 19:54:55 +0000 (15:54 -0400)]
add creal/cimag macros in complex.h (and use them in the functions defs)
Rich Felker [Thu, 22 Mar 2012 19:36:56 +0000 (15:36 -0400)]
tgmath.h: suppress any existing macro definitions before defining macros
this is necessary so that we can freely add macro versions of some of
the math/complex functions without worrying about breaking tgmath.
nsz [Thu, 22 Mar 2012 13:54:47 +0000 (14:54 +0100)]
acos.s fix: use the formula acos(x) = atan2(sqrt(1-x),sqrt(1+x))
the old formula atan2(1,sqrt((1+x)/(1-x))) was faster but
could give nan result at x=1 when the rounding mode is
FE_DOWNWARD (so 1-1 == -0 and 2/-0 == -inf), the new formula
gives -0 at x=+-1 with downward rounding.
Rich Felker [Thu, 22 Mar 2012 05:00:35 +0000 (01:00 -0400)]
sysconf support for dynamic limits (open files and processes)
Rich Felker [Wed, 21 Mar 2012 16:42:48 +0000 (12:42 -0400)]
fix DECIMAL_DIG definitions
DECIMAL_DIG is not the same as LDBL_DIG
type_DIG is the maximimum number of decimal digits that can survive a
round trip from decimal to type and back to decimal.
DECIMAL_DIG is the minimum number of decimal digits required in order
for any floating point type to survive the round trip to decimal and
back, and it is generally larger than LDBL_DIG. since the exact
formula is non-trivial, and defining it larger than necessary may be
legal but wasteful, just define the right value in bits/float.h.
Rich Felker [Wed, 21 Mar 2012 04:47:37 +0000 (00:47 -0400)]
initial, very primitive strfmon
Rich Felker [Wed, 21 Mar 2012 03:29:24 +0000 (23:29 -0400)]
x86_64 math asm, long double functions only
this has not been tested heavily, but it's known to at least assemble
and run in basic usage cases. it's nearly identical to the
corresponding i386 code, and thus expected to be just as correct or
just as incorrect.
Rich Felker [Wed, 21 Mar 2012 01:10:06 +0000 (21:10 -0400)]
limits.h: support gcc's -funsigned-char
some software apparently uses this and breaks with musl due to
mismatching definitions...
Rich Felker [Tue, 20 Mar 2012 23:51:11 +0000 (19:51 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Tue, 20 Mar 2012 23:44:05 +0000 (19:44 -0400)]
upgrade to latest upstream TRE regex code (0.8.0)
the main practical results of this change are
1. the regex code is no longer subject to LGPL; it's now 2-clause BSD
2. most (all?) popular nonstandard regex extensions are supported
I hesitate to call this a "sync" since both the old and new code are
heavily modified. in one sense, the old code was "more severely"
modified, in that it was actively hostile to non-strictly-conforming
expressions. on the other hand, the new code has eliminated the
useless translation of the entire regex string to wchar_t prior to
compiling, and now only converts multibyte character literals as
needed.
in the future i may use this modified TRE as a basis for writing the
long-planned new regex engine that will avoid multibyte-to-wide
character conversion entirely by compiling multibyte bracket
expressions specific to UTF-8.
nsz [Tue, 20 Mar 2012 21:49:19 +0000 (22:49 +0100)]
nearbyint optimization (only clear inexact when necessary)
old code saved/restored the fenv (the new code is only as slow
as that when inexact is not set before the call, but some other
flag is set and the rounding is inexact, which is rare)
before:
bench_nearbyint_exact
5000000 N 261 ns/op
bench_nearbyint_inexact_set
5000000 N 262 ns/op
bench_nearbyint_inexact_unset
5000000 N 261 ns/op
after:
bench_nearbyint_exact
10000000 N 94.99 ns/op
bench_nearbyint_inexact_set
25000000 N 65.81 ns/op
bench_nearbyint_inexact_unset
10000000 N 94.97 ns/op
nsz [Tue, 20 Mar 2012 19:08:35 +0000 (20:08 +0100)]
remove a fixme comment
nsz [Tue, 20 Mar 2012 19:04:53 +0000 (20:04 +0100)]
clean up pow.c and powf.c
fix comments about special cases
nsz [Tue, 20 Mar 2012 18:59:50 +0000 (19:59 +0100)]
clean up powl.c
fix special cases, use multiplication instead of scalbnl
nsz [Tue, 20 Mar 2012 14:17:15 +0000 (15:17 +0100)]
fix a cbrtl.c regression and remove x87 precision setting
nsz [Tue, 20 Mar 2012 09:25:02 +0000 (10:25 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Tue, 20 Mar 2012 04:51:32 +0000 (00:51 -0400)]
optimize scalbn family
the fscale instruction is slow everywhere, probably because it
involves a costly and unnecessary integer truncation operation that
ends up being a no-op in common usages. instead, construct a floating
point scale value with integer arithmetic and simply multiply by it,
when possible.
for float and double, this is always possible by going to the
next-larger type. we use some cheap but effective saturating
arithmetic tricks to make sure even very large-magnitude exponents
fit. for long double, if the scaling exponent is too large to fit in
the exponent of a long double value, we simply fallback to the
expensive fscale method.
on atom cpu, these changes speed up scalbn by over 30%. (min rdtsc
timing dropped from 110 cycles to 70 cycles.)
Rich Felker [Tue, 20 Mar 2012 03:53:52 +0000 (23:53 -0400)]
remquo asm: return quotient mod 8, as intended by the spec
this is a lot more efficient and also what is generally wanted.
perhaps the bit shuffling could be more efficient...
nsz [Tue, 20 Mar 2012 03:09:14 +0000 (04:09 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Tue, 20 Mar 2012 03:08:49 +0000 (23:08 -0400)]
use alternate formula for acos asm to avoid loss of precision
nsz [Tue, 20 Mar 2012 02:13:26 +0000 (03:13 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Tue, 20 Mar 2012 02:07:43 +0000 (22:07 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Tue, 20 Mar 2012 01:55:53 +0000 (21:55 -0400)]
fix exp asm
exponents (base 2) near 16383 were broken due to (1) wrong cutoff, and
(2) inability to fit the necessary range of scalings into a long
double value.
as a solution, we fall back to using frndint/fscale for insanely large
exponents, and also have to special-case infinities here to avoid
inf-inf generating nan.
thankfully the costly code never runs in normal usage cases.
nsz [Mon, 19 Mar 2012 22:55:06 +0000 (23:55 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
nsz [Mon, 19 Mar 2012 22:41:19 +0000 (23:41 +0100)]
code cleanup of named constants
zero, one, two, half are replaced by const literals
The policy was to use the f suffix for float consts (1.0f),
but don't use suffix for long double consts (these consts
can be exactly represented as double).
nsz [Mon, 19 Mar 2012 22:39:47 +0000 (23:39 +0100)]
fix remainder*.c: remove useless long double cast
nsz [Mon, 19 Mar 2012 22:30:45 +0000 (23:30 +0100)]
don't try to create non-standard denormalization signal
Underflow exception is only raised when the result is
invalid, but fmod is always exact. x87 has a denormalization
exception, but that's nonstandard. And the superflous *1.0
will be optimized away by any compiler that does not honor
signaling nans.
Rich Felker [Mon, 19 Mar 2012 22:30:35 +0000 (18:30 -0400)]
Merge remote branch 'nsz/master'
nsz [Mon, 19 Mar 2012 22:27:45 +0000 (23:27 +0100)]
new modff.c code, fix nan handling in modfl
nsz [Mon, 19 Mar 2012 21:57:58 +0000 (22:57 +0100)]
use scalbn or *2.0 instead of ldexp, fix fmal
Some code assumed ldexp(x, 1) is faster than 2.0*x,
but ldexp is a wrapper around scalbn which uses
multiplications inside, so this optimization is
wrong.
This commit also fixes fmal which accidentally
used ldexp instead of ldexpl loosing precision.
There are various additional changes from the
work-in-progress const cleanups.
nsz [Mon, 19 Mar 2012 21:49:03 +0000 (22:49 +0100)]
fix long double const workaround in cbrtl
nsz [Mon, 19 Mar 2012 18:26:31 +0000 (19:26 +0100)]
don't inline __rem_pio2l so the code size is smaller
nsz [Mon, 19 Mar 2012 18:14:32 +0000 (19:14 +0100)]
minor fix in __tanl (get sign properly)
Rich Felker [Mon, 19 Mar 2012 17:58:47 +0000 (13:58 -0400)]
bug fix: wrong opcode for writing long long
nsz [Mon, 19 Mar 2012 17:52:17 +0000 (18:52 +0100)]
remove long double const workarounds
Some long double consts were stored in two doubles as a workaround
for x86_64 and i386 with the following comment:
/* Long double constants are slow on these arches, and broken on i386. */
This is most likely old gcc bug related to the default x87 fpu
precision setting (it's double instead of double extended on BSD).
nsz [Mon, 19 Mar 2012 17:06:06 +0000 (18:06 +0100)]
fix erfl wrapper for long double==double case
Rich Felker [Mon, 19 Mar 2012 14:59:41 +0000 (10:59 -0400)]
asm for log1p
Rich Felker [Mon, 19 Mar 2012 14:43:28 +0000 (10:43 -0400)]
asm for log2
Rich Felker [Mon, 19 Mar 2012 13:42:51 +0000 (09:42 -0400)]
asm for remquo
this could perhaps use some additional testing for corner cases, but
it seems to be correct.
Rich Felker [Mon, 19 Mar 2012 13:00:30 +0000 (09:00 -0400)]
optimize exponential asm for i386
up to 30% faster exp2 by avoiding slow frndint and fscale functions.
expm1 also takes a much more direct path for small arguments (the
expected usage case).
Rich Felker [Mon, 19 Mar 2012 10:28:22 +0000 (06:28 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Mon, 19 Mar 2012 10:22:54 +0000 (06:22 -0400)]
fix broken modf family functions
Rich Felker [Mon, 19 Mar 2012 09:56:41 +0000 (05:56 -0400)]
asm for modf functions
nsz [Mon, 19 Mar 2012 09:54:07 +0000 (10:54 +0100)]
simplify scalbn*.c implementations
The old scalbn.c was wrong and slow, the new one is just slow.
(scalbn(0x1p+1023,-2097) should give 0x1p-1074, but the old code gave 0)
nsz [Mon, 19 Mar 2012 09:50:42 +0000 (10:50 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Mon, 19 Mar 2012 09:42:04 +0000 (05:42 -0400)]
asm for floor/ceil/trunc
nsz [Mon, 19 Mar 2012 09:20:24 +0000 (10:20 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Mon, 19 Mar 2012 09:15:30 +0000 (05:15 -0400)]
asm for scalbn family
unlike some implementations, these functions perform the equivalent of
gcc's -ffloat-store on the result before returning. this is necessary
to raise underflow/overflow/inexact exceptions, perform the correct
rounding with denormals, etc.
Rich Felker [Mon, 19 Mar 2012 08:56:07 +0000 (04:56 -0400)]
asm for inverse trig functions
unlike trig functions, these are easy to do in asm because they do not
involve (arbitrary-precision) argument reduction. fpatan automatically
takes care of domain issues, and in asin and acos, fsqrt takes care of
them for us.
nsz [Mon, 19 Mar 2012 07:01:21 +0000 (08:01 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Mon, 19 Mar 2012 03:50:54 +0000 (23:50 -0400)]
asm for log functions
Rich Felker [Mon, 19 Mar 2012 03:17:28 +0000 (23:17 -0400)]
fix broken exponential asm
infinities were getting converted into nans. the new code simply tests
for infinity and replaces it with a large magnitude value of the same
sign.
also, the fcomi instruction is apparently not part of the i387
instruction set, so avoid using it.
Rich Felker [Mon, 19 Mar 2012 02:05:20 +0000 (22:05 -0400)]
asm for lrint family on i386
nsz [Mon, 19 Mar 2012 01:55:31 +0000 (02:55 +0100)]
Merge branch 'master' of git://git.etalabs.net/musl
Rich Felker [Mon, 19 Mar 2012 01:43:43 +0000 (21:43 -0400)]
asm exponential functions for i386
nsz [Mon, 19 Mar 2012 01:05:57 +0000 (02:05 +0100)]
remove unnecessary TODO comments from fma.c