Implements the functions `roundeven()`, `roundevenf()`, `roundevenl()`
from the roundeven family of functions introduced in C23. Also
implements `roundevenf128()`.
We compute atan2f(y, x) in 2 stages:
- Fast step: perform computations in double precision , with relative
errors < 2^-50
- Accurate step: if the result from the Fast step fails Ziv's rounding
test, then we perform computations in double-double precision, with
relative errors < 2^-100.
On Ryzen 5900X, worst-case latency is ~ 200 clocks, compared to average
latency ~ 60 clocks, and average reciprocal throughput ~ 20 clocks.
Towards the goal of getting `ninja libc-lint` back to green, fix the numerous
instances of:
warning: header guard does not follow preferred style [llvm-header-guard]
This is because many of our header guards start with `__LLVM` rather than
`LLVM`.
To filter just these warnings:
$ ninja -k2000 libc-lint 2>&1 | grep llvm-header-guard
To automatically apply fixits:
$ find libc/src libc/include libc/test -name \*.h | \
xargs -n1 -I {} clang-tidy {} -p build/compile_commands.json \
-checks='-*,llvm-header-guard' --fix --quiet
Some manual cleanup is still necessary as headers that were missing header
guards outright will have them inserted before the license block (we prefer
them after).
We compute `pow(x, y)` using the formula
```
pow(x, y) = x^y = 2^(y * log2(x))
```
We follow similar steps as in `log2f(x)` and `exp2f(x)`, by breaking
down into `hi + mid + lo` parts, in which `hi` parts are computed using
the exponent field directly, `mid` parts will use look-up tables, and
`lo` parts are approximated by polynomials.
We add some speedup for common use-cases:
```
pow(2, y) = exp2(y)
pow(10, y) = exp10(y)
pow(x, 2) = x * x
pow(x, 1/2) = sqrt(x)
pow(x, -1/2) = rsqrt(x) - to be added
```
Explicit braces were added to fix the "suggest explicit braces to avoid
ambiguous ‘else’" warning since the current solution (switch (0) case 0:
default:) doesn't work since gcc 7 (see
https://github.com/google/googletest/issues/1119)
gcc 13 generates about 5000 of these warnings when building libc without
this patch.
Fixing an issue with LLVM libc's fenv.h defined rounding mode macros
differently from system libc, making get_round() return different values from
fegetround(). Also letting math tests to skip rounding modes that cannot be
set. This should allow math tests to be run on platforms in which fenv.h is not
implemented yet.
This allows us to re-enable hermatic floating point tests in
https://reviews.llvm.org/D151123 and reverting https://reviews.llvm.org/D152742.
Reviewed By: jhuber6
Differential Revision: https://reviews.llvm.org/D152873
This patch mimics the behavior of Google Test and allow users to log custom messages after all flavors of ASSERT_ / EXPECT_.
Reviewed By: sivachandra, lntue
Differential Revision: https://reviews.llvm.org/D152630
This patch mimics the behavior of Google Test and allow users to log custom messages after all flavors of ASSERT_ / EXPECT_.
Reviewed By: sivachandra, lntue
Differential Revision: https://reviews.llvm.org/D152630
This part of the effort to make all test related pieces into the `test`
directory. This helps is excluding test related pieces in a straight
forward manner if LLVM_INCLUDE_TESTS is OFF. Future patches will also move
the MPFR wrapper and testutils into the 'test' directory.
Adding `EXPECT_MPFR_MATCH_ROUNDING_SILENTLY` macro that does not call
`explainError` when the tests fail. This is useful to check the passing or
failing rates, such as hitting percentages of fast passes in math
implementations.
Reviewed By: michaelrj, sivachandra
Differential Revision: https://reviews.llvm.org/D136731
Implement exp10f function correctly rounded to all rounding modes.
Algorithm: perform range reduction to reduce
```
10^x = 2^(hi + mid) * 10^lo
```
where:
```
hi is an integer,
0 <= mid * 2^5 < 2^5
-log10(2) / 2^6 <= lo <= log10(2) / 2^6
```
Then `2^mid` is stored in a table of 32 entries and the product `2^hi * 2^mid` is
performed by adding `hi` into the exponent field of `2^mid`.
`10^lo` is then approximated by a degree-5 minimax polynomials generated by Sollya with:
```
> P = fpminimax((10^x - 1)/x, 4, [|D...|], [-log10(2)/64. log10(2)/64]);
```
Performance benchmark using perf tool from the CORE-MATH project on Ryzen 1700:
```
$ CORE_MATH_PERF_MODE="rdtsc" ./perf.sh exp10f
GNU libc version: 2.35
GNU libc release: stable
CORE-MATH reciprocal throughput : 10.215
System LIBC reciprocal throughput : 7.944
LIBC reciprocal throughput : 38.538
LIBC reciprocal throughput : 12.175 (with `-msse4.2` flag)
LIBC reciprocal throughput : 9.862 (with `-mfma` flag)
$ CORE_MATH_PERF_MODE="rdtsc" ./perf.sh exp10f --latency
GNU libc version: 2.35
GNU libc release: stable
CORE-MATH latency : 40.744
System LIBC latency : 37.546
BEFORE
LIBC latency : 48.989
LIBC latency : 44.486 (with `-msse4.2` flag)
LIBC latency : 40.221 (with `-mfma` flag)
```
This patch relies on https://reviews.llvm.org/D134002
Reviewed By: orex, zimmermann6
Differential Revision: https://reviews.llvm.org/D134104
Implement acosf function correctly rounded for all rounding modes.
We perform range reduction as follows:
- When `|x| < 2^(-10)`, we use cubic Taylor polynomial:
```
acos(x) = pi/2 - asin(x) ~ pi/2 - x - x^3 / 6.
```
- When `2^(-10) <= |x| <= 0.5`, we use the same approximation that is used for `asinf(x)` when `|x| <= 0.5`:
```
acos(x) = pi/2 - asin(x) ~ pi/2 - x - x^3 * P(x^2).
```
- When `0.5 < x <= 1`, we use the double angle formula: `cos(2y) = 1 - 2 * sin^2 (y)` to reduce to:
```
acos(x) = 2 * asin( sqrt( (1 - x)/2 ) )
```
- When `-1 <= x < -0.5`, we reduce to the positive case above using the formula:
```
acos(x) = pi - acos(-x)
```
Performance benchmark using perf tool from the CORE-MATH project on Ryzen 1700:
```
$ CORE_MATH_PERF_MODE="rdtsc" ./perf.sh acosf
GNU libc version: 2.35
GNU libc release: stable
CORE-MATH reciprocal throughput : 28.613
System LIBC reciprocal throughput : 29.204
LIBC reciprocal throughput : 24.271
$ CORE_MATH_PERF_MODE="rdtsc" ./perf.sh asinf --latency
GNU libc version: 2.35
GNU libc release: stable
CORE-MATH latency : 55.554
System LIBC latency : 76.879
LIBC latency : 62.118
```
Reviewed By: orex, zimmermann6
Differential Revision: https://reviews.llvm.org/D133550
Performance by core-math (core-math/glibc 2.31/current llvm-14):
10.845/43.174/13.467
The review is done on top of D132809.
Differential Revision: https://reviews.llvm.org/D132811
Migrating all private STL code to the standard STL case but keeping it under the CPP namespace to avoid confusion. Starting with the type_traits header.
Differential Revision: https://reviews.llvm.org/D130727
The specified rounding mode will be used and restored
to what it was before the test ran.
Additionally, it moves ForceRoundingMode and RoundingMode
out of MPFRUtils to be used in more places.
Differential Revision: https://reviews.llvm.org/D129685
This is a implementation of find remainder fmod function from standard libm.
The underline algorithm is developed by myself, but probably it was first
invented before.
Some features of the implementation:
1. The code is written on more-or-less modern C++.
2. One general implementation for both float and double precision numbers.
3. Spitted platform/architecture dependent and independent code and tests.
4. Tests covers 100% of the code for both float and double numbers. Tests cases with NaN/Inf etc is copied from glibc.
5. The new implementation in general 2-4 times faster for “regular” x,y values. It can be 20 times faster for x/y huge value, but can also be 2 times slower for double denormalized range (according to perf tests provided).
6. Two different implementation of division loop are provided. In some platforms division can be very time consuming operation. Depend on platform it can be 3-10 times slower than multiplication.
Performance tests:
The test is based on core-math project (https://gitlab.inria.fr/core-math/core-math). By Tue Ly suggestion I took hypot function and use it as template for fmod. Preserving all test cases.
`./check.sh <--special|--worst> fmodf` passed.
`CORE_MATH_PERF_MODE=rdtsc ./perf.sh fmodf` results are
```
GNU libc version: 2.35
GNU libc release: stable
21.166 <-- FPU
51.031 <-- current glibc
37.659 <-- this fmod version.
```
Based on RLIBM implementation similar to logf and log2f. Most of the exceptional inputs are the exact powers of 10.
Reviewed By: sivachandra, zimmermann6, santoshn, jpl169
Differential Revision: https://reviews.llvm.org/D118093
Add an extra argument for rounding mode to EXPECT_MPFR_MATCH and ASSERT_MPFR_MATCH macros.
Reviewed By: sivachandra, michaelrj
Differential Revision: https://reviews.llvm.org/D116777
The idea is to move all pieces related to the actual libc sources to the
"src" directory. This allows downstream users to ship and build just the
"src" directory.
Reviewed By: michaelrj
Differential Revision: https://reviews.llvm.org/D112653
These functions will be used in a future patch to implement
trigonometric functions. Unit tests have been added but to the
libc-long-running-tests suite. The unit tests long running because we
compare against MPFR computations performed at 1280 bits of precision.
Some cleanups or elimination of repeated patterns can be done as follow
up changes.
Differential Revision: https://reviews.llvm.org/D104817
Use expm1f(x) = exp(x) - 1 for |x| > ln(2).
For |x| <= ln(2), divide it into 3 subintervals: [-ln2, -1/8], [-1/8, 1/8], [1/8, ln2]
and use a degree-6 polynomial approximation generated by Sollya's fpminmax for each interval.
Errors < 1.5 ULPs when we use fma to evaluate the polynomials.
Differential Revision: https://reviews.llvm.org/D101134
The implementations use the x86_64 FPU instructions. These instructions
are extremely slow compared to a polynomial based software
implementation. Also, their accuracy falls drastically once the input
goes beyond 2PI. To improve both the speed and accuracy, we will be
taking the following approach going forward:
1. As a follow up to this CL, we will implement a range reduction algorithm
which will expand the accuracy to the entire double precision range.
2. After that, we will replace the HW instructions with a polynomial
implementation to improve the run time.
After step 2, the implementations will be accurate, performant and target
architecture independent.
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D102384
A new function to MPFRWrapper has been added, which is used to set up
the unit tests.
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D93007
Tests for frexp[f|l] now use the new capability. Not all input-output
combinations have been addressed by this change. Support for newer combinations
can be added in future as needed.
Reviewed By: lntue
Differential Revision: https://reviews.llvm.org/D86506
