The CORE-MATH implementation is correctly rounded (for any rounding mode)
and shows better performance compared to the generic exp2m1f.
The code was adapted to glibc style and to use the definition of
math_config.h (to handle errno, overflow, and underflow). The
only change is to handle FLT_MAX_EXP for FE_DOWNWARD or FE_TOWARDZERO.
The benchmark inputs are based on exp2f ones.
Benchtest on x64_64 (Ryzen 9 5900X, gcc 14.2.1), aarch64 (Neoverse-N1,
gcc 13.3.1), and powerpc (POWER10, gcc 13.2.1):
Latency master patched improvement
x86_64 40.6042 48.7104 -19.96%
x86_64v2 40.7506 35.9032 11.90%
x86_64v3 35.2301 31.7956 9.75%
i686 102.094 94.6657 7.28%
aarch64 18.2704 15.1387 17.14%
power10 11.9444 8.2402 31.01%
reciprocal-throughput master patched improvement
x86_64 20.8683 16.1428 22.64%
x86_64v2 19.5076 10.4474 46.44%
x86_64v3 19.2106 10.4014 45.86%
i686 56.4054 59.3004 -5.13%
aarch64 12.0781 7.3953 38.77%
power10 6.5306 5.9388 9.06%
The generic implementation calls __ieee754_exp2f and x86_64 provides
an optimized ifunc version (built with -mfma -mavx2, not correctly
rounded). This explains the performance difference for x86_64.
Same for i686, where the ABI provides an optimized __ieee754_exp2f
version built with '-msse2 -mfpmath=sse'. When built wth same
flags, the new algorithm shows a better performance:
master patched improvement
latency 102.094 91.2823 10.59%
reciprocal-throughput 56.4054 52.7984 6.39%
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
The CORE-MATH implementation is correctly rounded (for any rounding mode)
and shows better performance compared to the generic exp10m1f.
The code was adapted to glibc style and to use the definition of
math_config.h (to handle errno, overflow, and underflow). I mostly
fixed some small issues in corner cases (sNaN handling, -INFINITY,
a specific overflow check).
Benchtest on x64_64 (Ryzen 9 5900X, gcc 14.2.1), aarch64 (Neoverse-N1,
gcc 13.3.1), and powerpc (POWER10, gcc 13.2.1):
Latency master patched improvement
x86_64 45.4690 49.5845 -9.05%
x86_64v2 46.1604 36.2665 21.43%
x86_64v3 37.8442 31.0359 17.99%
i686 121.367 93.0079 23.37%
aarch64 21.1126 15.0165 28.87%
power10 12.7426 8.4929 33.35%
reciprocal-throughput master patched improvement
x86_64 19.6005 17.4005 11.22%
x86_64v2 19.6008 11.1977 42.87%
x86_64v3 17.5427 10.2898 41.34%
i686 59.4215 60.9675 -2.60%
aarch64 13.9814 7.9173 43.37%
power10 6.7814 6.4258 5.24%
The generic implementation calls __ieee754_exp10f which has an
optimized version, although it is not correctly rounded, which is
the main culprit of the the latency difference for x86_64 and
throughp for i686.
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Signed-off-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: DJ Delorie <dj@redhat.com>
The CORE-MATH implementation is correctly rounded (for any rounding mode).
This can be checked by exhaustive tests in a few minutes since there are
less than 2^32 values to check against for example GNU MPFR.
This patch also adds some bench values for tgammaf.
Tested on x86_64 and x86 (cfarm26).
With the initial GNU libc code it gave on an Intel(R) Core(TM) i7-8700:
"tgammaf": {
"": {
"duration": 3.50188e+09,
"iterations": 2e+07,
"max": 602.891,
"min": 65.1415,
"mean": 175.094
}
}
With the new code:
"tgammaf": {
"": {
"duration": 3.30825e+09,
"iterations": 5e+07,
"max": 211.592,
"min": 32.0325,
"mean": 66.1649
}
}
With the initial GNU libc code it gave on cfarm26 (i686):
"tgammaf": {
"": {
"duration": 3.70505e+09,
"iterations": 6e+06,
"max": 2420.23,
"min": 243.154,
"mean": 617.509
}
}
With the new code:
"tgammaf": {
"": {
"duration": 3.24497e+09,
"iterations": 1.8e+07,
"max": 1238.15,
"min": 101.155,
"mean": 180.276
}
}
Signed-off-by: Alexei Sibidanov <sibid@uvic.ca>
Signed-off-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Changes in v2:
- include <math.h> (fix the linknamespace failures)
- restored original benchtests/strcoll-inputs/filelist#en_US.UTF-8 file
- restored original wrapper code (math/w_tgammaf_compat.c),
except for the dealing with the sign
- removed the tgammaf/float entries in all libm-test-ulps files
- address other comments from Joseph Myers
(https://sourceware.org/pipermail/libc-alpha/2024-July/158736.html)
Changes in v3:
- pass NULL argument for signgam from w_tgammaf_compat.c
- use of math_narrow_eval
- added more comments
Changes in v4:
- initialize local_signgam to 0 in math/w_tgamma_template.c
- replace sysdeps/ieee754/dbl-64/gamma_productf.c by dummy file
Changes in v5:
- do not mention local_signgam any more in math/w_tgammaf_compat.c
- initialize local_signgam to 1 instead of 0 in w_tgamma_template.c
and added comment
Changes in v6:
- pass NULL as 2nd argument of __ieee754_gammaf_r in
w_tgammaf_compat.c, and check for NULL in e_gammaf_r.c
Changes in v7:
- added Signed-off-by line for Alexei Sibidanov (author of the code)
Changes in v8:
- added Signed-off-by line for Paul Zimmermann (submitted of the patch)
Changes in v9:
- address comments from review by Adhemerval Zanella
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Building the s390 specific iconv modules - utf16-utf32-z9.c, utf8-utf32-z9.c
and utf8-utf16-z9.c - with -fno-omit-frame-pointer leads to a build error
"error: %r11 cannot be used in 'asm' here" as r11 is needed as frame-pointer.
The cuXY-instructions need two even-odd register pairs. Therefore the register
pinning is used. This patch just uses a different register pair.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
The z13/vector-optimized wcsncmp implementation segfaults if n=1
and there is only one character (equal on both strings) before
the page end. Then it loads and compares one character and misses
to check n again. The following load fails.
This patch removes the extra load and compare of the first character
and just start with the loop which uses vector-load-to-block-boundary.
This code-path also checks n.
With this patch both tests are passing:
- the simplified one mentioned in the bugzilla 31934
- the full one in Florian Weimer's patch:
"manual: Document a GNU extension for strncmp/wcsncmp"
(https://patchwork.sourceware.org/project/glibc/patch/874j9eml6y.fsf@oldenburg.str.redhat.com/):
On s390x-linux-gnu (z16), the new wcsncmp test fails due to bug 31934.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
The test is not a run-time check, so update the description.
Also use readelf -W for a more stable output format and fix
an LC_ALL typo.
This avoids garbled configure messages:
checking for s390-specific static PIE requirements (runtime check)... 0x0000000000000017 (JMPREL) 0x280
yes
Remove the definitions of HWCAP_IMPORTANT after removal of
LD_HWCAP_MASK / tunable glibc.cpu.hwcap_mask. There HWCAP_IMPORTANT
was used as default value.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Remove the definitions of _DL_PLATFORMS_COUNT as those are not used
anymore after removal in elf/dl-cache.c:search_cache().
Note: On x86, we can also get rid of the definitions
HWCAP_PLATFORMS_START and HWCAP_PLATFORMS_COUNT.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Remove the definitions of _DL_FIRST_PLATFORM as those were only used
in the _DL_HWCAP_PLATFORM definitions and in _dl_string_platform().
Both were removed.
Note: Removed on every architecture despite of powerpc, where
_dl_string_platform() is still used.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Remove the definitions of _DL_HWCAP_PLATFORM as those are not used
anymore after removal in elf/dl-cache.c:search_cache().
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Remove the platform strings in dl-procinfo.c where also
the implementation of _dl_string_platform() was removed.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Despite of powerpc where the returned integer is stored in tcb,
and the diagnostics output, there is no user anymore.
Thus this patch removes the diagnostics output and
_dl_string_platform for all other platforms.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
As discussed at the patch review meeting
Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org>
Reviewed-by: Simon Chopin <simon.chopin@canonical.com>
C23 adds various <math.h> function families originally defined in TS
18661-4. Add the logp1 functions (aliases for log1p functions - the
name is intended to be more consistent with the new log2p1 and
log10p1, where clearly it would have been very confusing to name those
functions log21p and log101p). As aliases rather than new functions,
the content of this patch is somewhat different from those actually
adding new functions.
Tests are shared with log1p, so this patch *does* mechanically update
all affected libm-test-ulps files to expect the same errors for both
functions.
The vector versions of log1p on aarch64 and x86_64 are *not* updated
to have logp1 aliases (and thus there are no corresponding header,
tests, abilist or ulps changes for vector functions either). It would
be reasonable for such vector aliases and corresponding changes to
other files to be made separately. For now, the log1p tests instead
avoid testing logp1 in the vector case (a Makefile change is needed to
avoid problems with grep, used in generating the .c files for vector
function tests, matching more than one ALL_RM_TEST line in a file
testing multiple functions with the same inputs, when it assumes that
the .inc file only has a single such line).
Tested for x86_64 and x86, and with build-many-glibcs.py.
Instead of tying based on the linker name and version, check for the
required support:
* whether it does not generate dynamic TLS relocations in PIE
(binutils PR ld/22263);
* if it accepts --no-dynamic-linker (by using -static-pie);
* and if it adds a DT_JMPREL pointing to .rela.iplt with static pie.
The patch also trims the comments, for binutils one of the tests should
already cover it. The kernel ones are not clear which version should
have the backport, nor it is something that glibc can do much about
it. Finally, the glibc is somewhat confusing, since it refers
to commits not related to s390x.
Checked with a build for s390x-linux-gnu.
Reviewed-by: Stefan Liebler <stli@linux.ibm.com>
WG14 decided to use the name C23 as the informal name of the next
revision of the C standard (notwithstanding the publication date in
2024). Update references to C2X in glibc to use the C23 name.
This is intended to update everything *except* where it involves
renaming files (the changes involving renaming tests are intended to
be done separately). In the case of the _ISOC2X_SOURCE feature test
macro - the only user-visible interface involved - support for that
macro is kept for backwards compatibility, while adding
_ISOC23_SOURCE.
Tested for x86_64.
Starting with commits
- 7ea510127e
string: Add libc_hidden_proto for strchrnul
- 22999b2f0f
string: Add libc_hidden_proto for memrchr
building glibc on s390x with --disable-multi-arch fails if only
the C-variant of strchrnul / memrchr is used. This is the case
if gcc uses -march < z13.
The build fails with:
../sysdeps/s390/strchrnul-c.c:28:49: error: ‘__strchrnul_c’ undeclared here (not in a function); did you mean ‘__strchrnul’?
28 | __hidden_ver1 (__strchrnul_c, __GI___strchrnul, __strchrnul_c);
With --disable-multi-arch, __strchrnul_c is not available as string/strchrnul.c
is just included without defining STRCHRNUL and thus we also don't have to create
the internal hidden symbol.
Tested-by: Andreas K. Hüttel <dilfridge@gentoo.org>
The tunable parsing duplicates the tunable environment variable so it
null-terminates each one since it simplifies the later parsing. It has
the drawback of adding another point of failure (__minimal_malloc
failing), and the memory copy requires tuning the compiler to avoid mem
operations calls.
The parsing now tracks the tunable start and its size. The
dl-tunable-parse.h adds helper functions to help parsing, like a strcmp
that also checks for size and an iterator for suboptions that are
comma-separated (used on hwcap parsing by x86, powerpc, and s390x).
Since the environment variable is allocated on the stack by the kernel,
it is safe to keep the references to the suboptions for later parsing
of string tunables (as done by set_hwcaps by multiple architectures).
Checked on x86_64-linux-gnu, powerpc64le-linux-gnu, and
aarch64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The linker just concatenates the .init and .fini sections which
results in the complete _init and _fini functions. If needed the
linker adds padding bytes due to an alignment. GNU ld is adding
NOPs, which is fine. But e.g. mold is adding traps which results
in broken _init and _fini functions.
Thus this patch removes the alignment in .init and .fini sections
in crtn.S files.
We keep the 4 byte function alignment in crti.S files. As the
assembler now also outputs the start of _init and _fini functions
as multiples of 4 byte, it perhaps has to fill it. Although GNU as
is using NOPs here, to be sure, we just keep the alignment with
0x07 (=NOPs) at the end of crti.S.
In order to avoid an obvious NOP slide in _fini, this patch also
uses an lg instead of lgr instruction. Then the emitted instructions
needs a multiple of 4 bytes.
The _dl_non_dynamic_init does not parse LD_PROFILE, which does not
enable profile for dlopen objects. Since dlopen is deprecated for
static objects, it is better to remove the support.
It also allows to trim down libc.a of profile support.
Checked on x86_64-linux-gnu.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Using the memcmp symbol directly allows the compile to inline the
memcmp calls (especially because _dl_tunable_set_hwcaps uses constants
values), generating better code.
Checked with tst-tunables on s390x-linux-gnu (qemu system).
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
If feenableexcept or fedisableexcept gets excepts=FE_INVALID=0x80
as input, we have a signed left shift: 0x80 << 24 which is not
representable as int and thus is undefined behaviour according to
C standard.
This patch casts excepts as unsigned int before shifting, which is
defined.
For me, the observed undefined behaviour is that the shift is done
with "unsigned"-instructions, which is exactly what we want.
Furthermore, I don't get any exception-flags.
After the fix, the code is using the same instruction sequence as
before.
The static PIE configure check uses link tests. When bootstrapping
a cross-toolchain, the link tests fail due to missing crt-files /
libc.so. As we explicitely want to test an issue in binutils (ld),
we now also explicitely check for known linker versions.
See also commit 368b7c614b
S390: Use compile-only instead of also link-tests in configure.
Bump autoconf requirement to 2.71 to allow regenerating configure on
more recent distributions. autoconf 2.71 has been in Fedora since F36
and is the current version in Debian stable (bookworm). It appears to
be current in Gentoo as well.
All sysdeps configure and preconfigure scripts have also been
regenerated; all changes are trivial transformations that do not affect
functionality.
Signed-off-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Depending on build configuration, the [routine]-c.c files may be chosen
to provide fortified routines implementation. While [routines].c
implementation were automatically excluded, the [routines]-c.c ones were
not. This patch fixes that by adding these file to the list to be
filtered.
Some of the s390-specific configure checks are using compile and
link configure tests. Now use only compile tests as the link
tests fails when e.g. bootstrapping a cross-toolchain due to
missing crt-files/libc.so. This is achieved by using
AC_COMPILE_IFELSE in configure.ac file.
This is observable e.g. when using buildroot which builds glibc
only once or the build-many-glibcs.py script. Note that the latter
one is building glibc twice in the compilers-step (configure-checks
fails) and in the glibcs-step (configure-checks succeed).
Note, that the s390 specific configure tests for static PIE have to
link an executable to test binutils support. Thus we can't fix
those tests.
And make always supported. The configure option was added on glibc 2.25
and some features require it (such as hwcap mask, huge pages support, and
lock elisition tuning). It also simplifies the build permutations.
Changes from v1:
* Remove glibc.rtld.dynamic_sort changes, it is orthogonal and needs
more discussion.
* Cleanup more code.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The _FPU_SETCW and _FPU_GETCW macros are defined with inline assemblies.
They use the sfpc and efpc instructions, respectively. But both contain
a spurious second operand that leads to a compile error with Clang.
Removing this operand works both with gcc/gas (since binutils 2.18) as
well as with clang/llvm.
Needed due to recent commits:
- "added pair of inputs for hypotf in binary32"
commit ID cf7ffdd8a5
- "update auto-libm-test-out-hypot"
commit ID 3efbf11fdf
It fixes the build after 7ea510127e and 22999b2f0f.
Checked with build for s390x-linux-gnu with -march=z13.
Reviewed-by: Arjun Shankar <arjun@redhat.com>
And use a packed structure instead. The compiler generates optimized
unaligned code if the architecture supports it.
Checked on x86_64-linux-gnu and i686-linux-gnu.
Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
Builds for s390 recently started failing with:
../sysdeps/s390/multiarch/ifunc-impl-list.c: In function '__libc_ifunc_impl_list':
../sysdeps/s390/multiarch/ifunc-impl-list.c:83:21: error: unused variable 'dl_hwcap' [-Werror=unused-variable]
83 | unsigned long int dl_hwcap = features->hwcap;
| ^~~~~~~~
https://sourceware.org/pipermail/libc-testresults/2023q1/010855.html
Add __attribute__ ((unused)) as already done for another variable
there.
Tested with build-many-glibcs.py (compilers and glibcs) for
s390x-linux-gnu and s390-linux-gnu.
Note: s390x-linux-gnu-O3 started failing with a different error
earlier; that problem may still need to be fixed after this fix is in.
https://sourceware.org/pipermail/libc-testresults/2023q1/010829.html
Although static linker can optimize it to local call, it follows the
internal scheme to provide hidden proto and definitions.
Reviewed-by: Carlos Eduardo Seo <carlos.seo@linaro.org>
Although static linker can optimize it to local call, it follows the
internal scheme to provide hidden proto and definitions.
Reviewed-by: Carlos Eduardo Seo <carlos.seo@linaro.org>
This patch enables the option to influence hwcaps and stfle bits used
by the s390 specific ifunc-resolvers. The currently x86-specific
tunable glibc.cpu.hwcaps is also used on s390x to achieve the task. In
addition the user can also set a CPU arch-level like z13 instead of
single HWCAP and STFLE features.
Note that the tunable only handles the features which are really used
in the IFUNC-resolvers. All others are ignored as the values are only
used inside glibc. Thus we can influence:
- HWCAP_S390_VXRS (z13)
- HWCAP_S390_VXRS_EXT (z14)
- HWCAP_S390_VXRS_EXT2 (z15)
- STFLE_MIE3 (z15)
The influenced hwcap/stfle-bits are stored in the s390-specific
cpu_features struct which also contains reserved fields for future
usage.
The ifunc-resolvers and users of stfle bits are adjusted to use the
information from cpu_features struct.
On 31bit, the ELF_MACHINE_IRELATIVE macro is now also defined.
Otherwise the new ifunc-resolvers segfaults as they depend on
the not yet processed_rtld_global_ro@GLIBC_PRIVATE relocation.
It also cleanups the multiple inclusion by leaving the ifunc
implementation to undef the weak_alias and libc_hidden_def.
Co-authored-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
New algorithm now calls strchrnul.
Checked on x86_64-linux-gnu, i686-linux-gnu, powerpc-linux-gnu,
and powerpc64-linux-gnu by removing the arch-specific assembly
implementation and disabling multi-arch (it covers both LE and BE
for 64 and 32 bits).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
New algorithm read the first aligned address and mask off the unwanted
bytes (this strategy is similar to arch-specific implementations used
on powerpc, sparc, and sh).
The loop now read word-aligned address and check using the has_zero_eq
function.
Checked on x86_64-linux-gnu, i686-linux-gnu, powerpc64-linux-gnu,
and powerpc-linux-gnu by removing the arch-specific assembly
implementation and disabling multi-arch (it covers both LE and BE
for 64 and 32 bits).
Co-authored-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
