PREFER_FLOAT_COMPARISON setting was not correct as it could raise
spurious exceptions. Fixing it is easy: just use ISLESS(x, y) instead
of abstop12(x) < abstop12(y) with appropriate non-signaling definition
for ISLESS. However it seems this setting is not very useful (there is
only minor performance difference on various architectures), so remove
this option for now.
The !HAVE_FAST_FMA code path split r = z/c - 1 into r = rhi + rlo such
that when z = 1-tiny and c = 1 then rlo and rhi could have much larger
magnitude than r which later caused large rounding errors.
So do a nearest rounding instead of truncation at the split.
In newlib with default settings this was observable on some arm targets
that enable the new math code but has no fma.
The roundtoint and converttoint internal functions are only called with small
values, so 32 bit result is enough for converttoint and it is a signed int
conversion so the natural return type is int32_t.
The original idea was to help the compiler keeping the result in uint64_t,
then it's clear that no sign extension is needed and there is no accidental
undefined or implementation defined signed int arithmetics.
But it turns out gcc does a good job with inlining so changing the type has
no overhead and the semantics of the conversion is less surprising this way.
Since we want to allow the asuint64 (x + 0x1.8p52) style conversion, the top
bits were never usable and the existing code ensures that only the bottom
32 bits of the conversion result are used.
In newlib with default settings only aarch64 is affected and there is no
significant code generation change with gcc after the patch.
Synchronize code style and comments with Arm Optimized Routines, there
are no code changes in this patch. This ensures different projects using
the same code have consistent code style so bug fix patches can be applied
more easily.
The new implementation is provided under !__OBSOLETE_MATH, it uses
ISO C99 code. With default settings the worst case error in nearest
rounding mode is 0.54 ULP with inlined fma and fma contraction. It uses
a 4 KB lookup table in addition to the table in exp_data.c, on aarch64
.text+.rodata size of libm.a is increased by 2295 bytes.
Improvements on Cortex-A72:
latency: 3.3x
thruput: 4.9x
The new implementation is provided under !__OBSOLETE_MATH, it uses
ISO C99 code. With default settings the worst case error in nearest
rounding mode is 0.547 ULP with inlined fma and fma contraction. It uses
a 1 KB lookup table, on aarch64 .text+.rodata size of libm.a is increased
by 1584 bytes.
Note that the math.h header defines log2(x) to be log(x)/Ln2, this is
not changed, so the new code is only used if that macro is suppressed.
Improvements on Cortex-A72:
latency: 2.0x
thruput: 2.2x
The new implementations are provided under !__OBSOLETE_MATH, it uses
ISO C99 code. With default settings the worst case error in nearest
rounding mode is 0.519 ULP with inlined fma and fma contraction. It uses
a 2 KB lookup table, on aarch64 .text+.rodata size of libm.a is increased
by 1703 bytes. The w_log.c wrapper is disabled since error handling is
inline in the new code.
New __HAVE_FAST_FMA and __HAVE_FAST_FMA_DEFAULT feature macros were
added to enable selecting between the code path that uses fma and the
one that does not. Targets supposed to set __HAVE_FAST_FMA_DEFAULT
if they have single instruction fma and the compiler can actually
inline it (gcc has __FP_FAST_FMA macro but that does not guarantee
inlining with -fno-builtin-fma).
Improvements on Cortex-A72:
latency: 1.9x
thruput: 2.3x
The new implementations are provided under !__OBSOLETE_MATH, they use
ISO C99 code. There are several settings, with the default one the
worst case error in nearest rounding mode is 0.509 ULP for exp and
0.507 ULP for exp2 when a multiply and add is contracted into an fma.
They use a shared 2 KB lookup table, on aarch64 .text+.rodata size
of libm.a is increased by 1868 bytes. The w_*.c wrappers are disabled
for the new code as it takes care of error handling inline.
The old exp2(x) code used to be just pow(2,x) so the speedup there
is more significant.
The file name has no special prefix to avoid any name collision with
existing files.
Improvements on Cortex-A72:
exp latency: 3.2x
exp thruput: 4.1x
exp2 latency: 7.8x
exp2 thruput: 18.8x
This change is equivalent to the commit
c65db17340
and only affects code that is from the Arm optimized-routines project.
It does not affect the observable behaviour, but the code generation
can be different on 64bit targets. The intention is to make the
portable semantics of the code obvious by using a fixed size type.
Here is the correct patch with both filenames and int cast fixed:
This patch is a complete rewrite of sinf, cosf and sincosf. The new version
is significantly faster, as well as simple and accurate.
The worst-case ULP is 0.56072, maximum relative error is 0.5303p-23 over all
4 billion inputs. In non-nearest rounding modes the error is 1ULP.
The algorithm uses 3 main cases: small inputs which don't need argument
reduction, small inputs which need a simple range reduction and large inputs
requiring complex range reduction. The code uses approximate integer
comparisons to quickly decide between these cases - on some targets this may
be slow, so this can be configured to use floating point comparisons.
The small range reducer uses a single reduction step to handle values up to
120.0. It is fastest on targets which support inlined round instructions.
The large range reducer uses integer arithmetic for simplicity. It does a
32x96 bit multiply to compute a 64-bit modulo result. This is more than
accurate enough to handle the worst-case cancellation for values close to
an integer multiple of PI/4. It could be further optimized, however it is
already much faster than necessary.
Simple benchmark showing speedup factor on AArch64 for various ranges:
range 0.7853982 sinf 1.7 cosf 2.2 sincosf 2.8
range 1.570796 sinf 1.9 cosf 1.9 sincosf 2.7
range 3.141593 sinf 2.0 cosf 2.0 sincosf 3.5
range 6.283185 sinf 2.3 cosf 2.3 sincosf 4.2
range 125.6637 sinf 2.9 cosf 3.0 sincosf 5.1
range 1.1259e15 sinf 26.8 cosf 26.8 sincosf 45.2
ChangeLog:
2018-05-18 Wilco Dijkstra <wdijkstr@arm.com>
* newlib/libm/common/Makefile.in: Regenerated.
* newlib/libm/common/Makefile.am: Add sinf.c, cosf.c, sincosf.c
sincosf.h, sincosf_data.c. Add -fbuiltin -fno-math-errno to CFLAGS.
* newlib/libm/common/math_config.h: Add HAVE_FAST_ROUND, HAVE_FAST_LROUND,
roundtoint, converttoint, force_eval_float, force_eval_double, eval_as_float,
eval_as_double, likely, unlikely.
* newlib/libm/common/cosf.c: New file.
* newlib/libm/common/sinf.c: Likewise.
* newlib/libm/common/sincosf.h: Likewise.
* newlib/libm/common/sincosf.c: Likewise.
* newlib/libm/common/sincosf_data.c: Likewise.
* newlib/libm/math/sf_cos.c: Add #if to build conditionally.
* newlib/libm/math/sf_sin.c: Likewise.
* newlib/libm/math/wf_sincos.c: Likewise.
--
This patch is a complete rewrite of sinf, cosf and sincosf. The new version
is significantly faster, as well as simple and accurate.
The worst-case ULP is 0.56072, maximum relative error is 0.5303p-23 over all
4 billion inputs. In non-nearest rounding modes the error is 1ULP.
The algorithm uses 3 main cases: small inputs which don't need argument
reduction, small inputs which need a simple range reduction and large inputs
requiring complex range reduction. The code uses approximate integer
comparisons to quickly decide between these cases - on some targets this may
be slow, so this can be configured to use floating point comparisons.
The small range reducer uses a single reduction step to handle values up to
120.0. It is fastest on targets which support inlined round instructions.
The large range reducer uses integer arithmetic for simplicity. It does a
32x96 bit multiply to compute a 64-bit modulo result. This is more than
accurate enough to handle the worst-case cancellation for values close to
an integer multiple of PI/4. It could be further optimized, however it is
already much faster than necessary.
Simple benchmark showing speedup factor on AArch64 for various ranges:
range 0.7853982 sinf 1.7 cosf 2.2 sincosf 2.8
range 1.570796 sinf 1.9 cosf 1.9 sincosf 2.7
range 3.141593 sinf 2.0 cosf 2.0 sincosf 3.5
range 6.283185 sinf 2.3 cosf 2.3 sincosf 4.2
range 125.6637 sinf 2.9 cosf 3.0 sincosf 5.1
range 1.1259e15 sinf 26.8 cosf 26.8 sincosf 45.2
ChangeLog:
2018-06-18 Wilco Dijkstra <wdijkstr@arm.com>
* newlib/libm/common/Makefile.in: Regenerated.
* newlib/libm/common/Makefile.am: Add sinf.c, cosf.c, sincosf.c
sincosf.h, sincosf_data.c. Add -fbuiltin -fno-math-errno to CFLAGS.
* newlib/libm/common/math_config.h: Add HAVE_FAST_ROUND, HAVE_FAST_LROUND,
roundtoint, converttoint, force_eval_float, force_eval_double, eval_as_float,
eval_as_double, likely, unlikely.
* newlib/libm/common/cosf.c: New file.
* newlib/libm/common/sinf.c: Likewise.
* newlib/libm/common/sincosf.h: Likewise.
* newlib/libm/common/sincosf.c: Likewise.
* newlib/libm/common/sincosf_data.c: Likewise.
* newlib/libm/math/sf_cos.c: Add #if to build conditionally.
* newlib/libm/math/sf_sin.c: Likewise.
* newlib/libm/math/wf_sincos.c: Likewise.
--
- From: Cesar Philippidis <cesar@codesourcery.com>
Date: Tue, 10 Apr 2018 14:43:42 -0700
Subject: [PATCH] nvptx port
This port adds support for Nvidia GPU's, which are primarily used as
offload accelerators in OpenACC and OpenMP.
Updated patch to use 0.0f in addition to calling rintf.
Tested same way as before, with a testcase that triggers the code and
make check.
OK?
newlib/
* libm/math/wf_pow.c (powf): Call rintf instead of rint. Use 0.0f
for compare.
Discard QUICKREF sections, rather than writing them to stderr
Discard MATHREF sections, rather than discarding as an error
Pass NOTES sections through to texinfo, rather than discarding as an error
Don't redirect makedoc stderr to .ref file
Remove makedoc output on error
Remove .ref files from CLEANFILES
Regenerate Makefile.ins
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
The recently added new math code inlines error handling instead of using
error handling wrappers around __ieee754* internal symbols, and thus the
__ieee754* symbols are no longer provided.
However __ieee754_expf and __ieee754_logf are used in the implementation
of a number of other math functions. These symbols are safe to redirect
to the external expf and logf symbols, because those names are always
reserved when single precision math functions are reserved and the
additional error handling code is either not reached or there will be
an error in the final result that will override an internal spurious
errno setting.
For consistency all of __ieee754_expf, __ieee754_logf and __ieee754_powf
are redirected using a macro.
Based on code from https://github.com/ARM-software/optimized-routines/
This patch adds a highly optimized generic implementation of expf,
exp2f, logf, log2f and powf. The new functions are not only
faster (6x for powf!), but are also smaller and more accurate.
In order to achieve this, the algorithm uses double precision
arithmetic for accuracy, avoids divisions and uses small table
lookups to minimize the polynomials. Special cases are handled
inline to avoid the unnecessary overhead of wrapper functions and
set errno to POSIX requirements.
The new functions are added under newlib/libm/common, but the old
implementations are kept (in newlib/libm/math) for non-IEEE or
pre-C99 systems. Targets can enable the new math code by defining
__OBSOLETE_MATH_DEFAULT to 0 in newlib/libc/include/machine/ieeefp.h,
users can override the default by defining __OBSOLETE_MATH.
Currently the new code is enabled for AArch64 and AArch32 with VFP.
Targets with a single precision FPU may still prefer the old
implementation.
libm.a size changes:
arm: -1692
arm/thumb/v7-a/nofp: -878
arm/thumb/v7-a+fp/hard: -864
arm/thumb/v7-a+fp/softfp: -908
aarch64: -1476
previous commit 4c90db7bc8 introduced
a compile time error because libm/common/s_infconst.c used the remove
__fmath, __dmath, and __ldmath union types.
Since this is very old, and unused for a very long time, just drop the
file and thus the __infinity constants entirely.
Exception: Cygwin exports __infinity from the beginning. There's a very,
VERY low probability that any existing executable or lib still uses this
constant, but we just keep it in for backward compat, nevertheless.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Hi,
make pdf on Ubuntu 16.04 fail with:
newlib/libc/libc.texinfo:9: Missing @endcsname inserted.
After a lot of fiddling the reason appears to be the combination of concept
and function index despite a lack of concept index entries. Arguably texinfo
should not error in that case but here we are, newlib will fail to build its
documentation on some systems because of this. Since libc.texinfo only
contains function index entries this patch simply removes the combination of
indices. It does the same for libm.texinfo which has concept index entries but
no function index entries.
Tested by running make pdf, make dvi, make info and make html successfully.
libc.pdf appears to have only one index as expected.
== Proposed commit message ==
Fix pdf build failure with texinfo 6.1.0 as provided in Ubuntu 16.04. Index
combination in libc.texinfo and libm.texinfo fails because both file have only
one type of index entries. Removing index combination is thus harmless and
solves the problem.
Is this ok for master?
Best regards,
Thomas
Add makedocbook, a tool to process makedoc markup and output DocBook XML
refentries.
Process all the source files which are processed with makedoc with
makedocbook as well
Add chapter-texi2docbook, a tool to automatically generate DocBook XML
chapter files from the chapter .texi files. For generating man pages all we
care about is the content of the refentries, so all this needs to do is
convert the @include of the makedoc generated .def files to xi:include of
the makedocbook generated .xml files.
Add skeleton Docbook XML book files, lib[cm].in.xml which include these
generated chapters, which in turn include the generated files containing
refentries, which is processed with xsltproc to generate the lib[cm].xml
Add new make targets to generate and install man pages from lib[cm].xml
Add makedocbook, a tool to process makedoc markup and output DocBook XML
refentries.
Process all the source files which are processed with makedoc with
makedocbook as well
Add chapter-texi2docbook, a tool to automatically generate DocBook XML
chapter files from the chapter .texi files. For generating man pages all we
care about is the content of the refentries, so all this needs to do is
convert the @include of the makedoc generated .def files to xi:include of
the makedocbook generated .xml files.
Add skeleton Docbook XML book files, lib[cm].in.xml which include these
generated chapters, which in turn include the generated files containing
refentries, which is processed with xsltproc to generate the lib[cm].xml
Add new make targets to generate and install man pages from lib[cm].xml
__builtin_isinf_sign doesn't reflect the correct return value for NaN
to emulate finite function. Use __builtin_isfinite instead whichg is available
since GCC 4.4 just as __builtin_isinf_sign.
* libm/common/sl_finite.c (finitel): Use __builtin_isfinite.
Signed-off-by: Corinna Vinschen <corinna@vinschen.de>
Currently, the newlib version information needs to be updated in two places:
- newlib/acinclude.m4
- newlib/libc/include/sys/features.h
The goal of this patch is to:
- supply a single location for defining the newlib version
information: newlib/acinclude.m4
- define __NEWLIB__, __NEWLIB_MINOR__ and __NEWLIB_PATCHLEVEL__
This is in line with what gcc does for its version macros. See:
https://gcc.gnu.org/onlinedocs/cpp/Common-Predefined-Macros.html
This patch moves the definition of the _NEWLIB_VERSION, __NEWLIB__
and __NEWLIB_MINOR__ macros from newlib/libc/include/sys/features.h,
to the newly generated newlib/_newlib_version.h file. Additionally,
the __NEWLIB_PATCHLEVEL__ macro was created, for completeness.
In order to stay backwards compatible, newlib/_newlib_version.h gets
included by newlib/newlib.h and newlib/libc/include/sys/features.h.
Note: This patch does _not_ include the modifications to the following
files, as these should all be generated any way.
*Makefile.in,
*aclocal.m4,
*configure
stamp-* files
Signed-off-by: Pieter du Preez <pdupreez@gmail.com>
Currently, double precision math functions in newlib/libm/machine/arm detect whether neon instructions can be used for double precision computation by checking the architecture version (>= 8) and the availability of floating-point instructions. However, these instructions would not be available if targeting fpv5-sp-d16 as the FPU, which by definition does not have 64bit float instructions. This patch adds a check that __ARM_FP advertises 64bit float instructions.
* libm/machine/arm/s_ceil.c: Also check that 64bit FP instructions are
available in the guard.
* libm/machine/arm/s_floor.c: Likewise.
* libm/machine/arm/s_nearbyint.c: Likewise.
* libm/machine/arm/s_rint.c: Likewise.
* libm/machine/arm/s_round.c: Likewise.
* libm/machine/arm/s_trunc.c: Likewise.
makedoc defines a command as 'all upper case, and alone on a line'.
A few QUICKREF lines currently violate this by having some additional text after
the QUICKREF.
So, currently, these lines are treated as an unknown command.
This is benign as QUICKREF currently does nothing but produce some ignored
output on stderr. I'm not sure what the intent of QUICKREF is.
2015-11-06 Jon Turney <jon.turney@dronecode.org.uk>
* libm/mathfp/s_acos.c: Fix QUICKREF.
* libm/mathfp/e_acosh.c: Ditto.
* libm/math/w_asin.c: Ditto.
* libm/mathfp/e_acosh.c: Ditto.
* libm/mathfp/s_acos.c: Ditto.
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
This got removed everywhere else in commit 139f923b, so I'm assuming whatever
ancient bug this was has been fixed long since.
There are plenty of other uses of texinfo with a filename containing an
underscore now.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
I think these are accidental omissions, as these source files are listed to be
chewed by makedoc, but the result is not included by any texinfo source file.
Future work: Nothing in libc/reent/ which is processed by makedoc is included by
reent.tex
2015-06-23 Jon Turney <jon.turney@dronecode.org.uk>
* libc/stdlib/stdlib.tex: Include itoa and utoa, and add to menu.
* libc/string/strings.tex: Include memrchr and rawmemchr, and add
to menu.
* libm/math/math.tex: Include exp10 and pow10, and add to menu.
* libm/common/s_exp10.c: Improve one-line description.
* libm/common/s_exp10.c: Ditto.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
2015-06-23 Jon Turney <jon.turney@dronecode.org.uk>
* libc/stdlib/llabs.c: Mark up parameter name in ANSI_SYNOPSIS.
* libc/time/tzset.c: Add and mark up parameter in SYNOPSIS.
* libm/common/s_nan.c: Ditto.
Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
Corinna Vinschen
Szabolcs Nagy
Jon Beniston
Wilco Dijkstra
Matthias Kannwischer
Jeff Johnston
Yaakov Selkowitz
Jim Wilson
Jon Turney
Brian Inglis
Kito Cheng
Aditya Upadhyay
Sebastian Huber
Thomas Preudhomme
Pieter du Preez
Thomas Preud'homme
DJ Delorie