mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-02-01 03:53:33 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
cosmopolitan/test/libc/tinymath/powl_test.c

572 lines
24 KiB
C
Raw Normal View History

Initial import
2020-06-15 14:18:57 +00:00
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
╞══════════════════════════════════════════════════════════════════════════════╡
│ Copyright 2020 Justine Alexandra Roberts Tunney │
│ │
Change license
2020-12-28 01:18:44 +00:00
│ Permission to use, copy, modify, and/or distribute this software for │
│ any purpose with or without fee is hereby granted, provided that the │
│ above copyright notice and this permission notice appear in all copies. │
Initial import
2020-06-15 14:18:57 +00:00
│ │
Change license
2020-12-28 01:18:44 +00:00
│ THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL │
│ WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED │
│ WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE │
│ AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL │
│ DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR │
│ PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER │
│ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR │
│ PERFORMANCE OF THIS SOFTWARE. │
Initial import
2020-06-15 14:18:57 +00:00
╚─────────────────────────────────────────────────────────────────────────────*/
Improve signal handling and math - Polyfill ucontext_t on FreeBSD/OpenBSD/NetBSD - Add tests confirming signals can edit CPU state - Work towards supporting ZIP filesystem on bare metal - Add more tinymath unit tests for POSIX conformance - Add X87 and SSE status flags to crash report - Fix some bugs in blinkenlights - Fix llvm build breakage
2021-02-26 02:30:17 +00:00
#include "libc/calls/struct/sigaction.h"
#include "libc/calls/struct/siginfo.h"
#include "libc/calls/ucontext.h"
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
#include "libc/errno.h"
#include "libc/fmt/fmt.h"
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
#include "libc/macros.internal.h"
Initial import
2020-06-15 14:18:57 +00:00
#include "libc/math.h"
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
#include "libc/mem/gc.h"
Improve signal handling and math - Polyfill ucontext_t on FreeBSD/OpenBSD/NetBSD - Add tests confirming signals can edit CPU state - Work towards supporting ZIP filesystem on bare metal - Add more tinymath unit tests for POSIX conformance - Add X87 and SSE status flags to crash report - Fix some bugs in blinkenlights - Fix llvm build breakage
2021-02-26 02:30:17 +00:00
#include "libc/runtime/pc.internal.h"
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
#include "libc/stdio/rand.h"
Add more libm unit tests and fixes See #61
2021-03-02 21:57:23 +00:00
#include "libc/stdio/stdio.h"
Improve signal handling and math - Polyfill ucontext_t on FreeBSD/OpenBSD/NetBSD - Add tests confirming signals can edit CPU state - Work towards supporting ZIP filesystem on bare metal - Add more tinymath unit tests for POSIX conformance - Add X87 and SSE status flags to crash report - Fix some bugs in blinkenlights - Fix llvm build breakage
2021-02-26 02:30:17 +00:00
#include "libc/sysv/consts/sa.h"
#include "libc/sysv/consts/sig.h"
Initial import
2020-06-15 14:18:57 +00:00
#include "libc/testlib/ezbench.h"
#include "libc/testlib/testlib.h"
#include "libc/x/x.h"
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
#include "libc/x/xasprintf.h"
Initial import
2020-06-15 14:18:57 +00:00
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
int rando;
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
char buf[128];
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
void SetUp(void) {
rando = rand() & 0xffff;
}
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
static double POW(double x, double y) {
return powl(x, y);
}
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
static char *fmtd(double x) {
sprintf(buf, "%.15g", x);
return buf;
}
static char *fmtf(float x) {
sprintf(buf, "%.6g", x);
return buf;
}
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
TEST(powl, test) {
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("27", _gc(xdtoal(powl(3, 3))));
EXPECT_STREQ("-27", _gc(xdtoal(powl(-3, 3))));
EXPECT_STREQ("1e+4932", _gc(xdtoal(powl(10, 4932))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(10, 4933))));
Port a lot more code to AARCH64 - Introduce epoll_pwait() - Rewrite -ftrapv and ffs() libraries in C code - Use more FreeBSD code in math function library - Get significantly more tests passing on qemu-aarch64 - Fix many Musl long double functions that were broken on AARCH64
2023-05-14 16:32:15 +00:00
// EXPECT_STREQ("0", _gc(xdtoal(powl(10, -5000))));
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("1.063382396627933e+37", _gc(xdtoal(powl(2, 123))));
EXPECT_STARTSWITH(".4248496805467504", _gc(xdtoal(powl(.7, 2.4))));
EXPECT_STREQ("1", _gc(xdtoal(powl(1, NAN))));
EXPECT_STREQ("1", _gc(xdtoal(powl(1, rando))));
EXPECT_STREQ("1", _gc(xdtoal(powl(NAN, 0))));
EXPECT_STREQ("1", _gc(xdtoal(powl(rando, 0))));
EXPECT_STREQ("0", _gc(xdtoal(powl(0, 1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(0, 2))));
EXPECT_STREQ("0", _gc(xdtoal(powl(0, 2.1))));
EXPECT_STREQ("1", _gc(xdtoal(powl(-1, INFINITY))));
EXPECT_STREQ("1", _gc(xdtoal(powl(-1, -INFINITY))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(1. / MAX(2, rando), -INFINITY))));
EXPECT_STREQ("0", _gc(xdtoal(powl(1.1, -INFINITY))));
EXPECT_STREQ("0", _gc(xdtoal(powl(MAX(2, rando), -INFINITY))));
EXPECT_STREQ("0", _gc(xdtoal(powl(1. / MAX(2, rando), INFINITY))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(MAX(2, rando), INFINITY))));
EXPECT_STREQ("-0", _gc(xdtoal(powl(-INFINITY, -1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(-INFINITY, -1.1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(-INFINITY, -2))));
EXPECT_STREQ("0", _gc(xdtoal(powl(-INFINITY, -2.1))));
EXPECT_STREQ("-0", _gc(xdtoal(powl(-INFINITY, -3))));
EXPECT_STREQ("0", _gc(xdtoal(powl(-INFINITY, -3.1))));
EXPECT_STREQ("-INFINITY", _gc(xdtoal(powl(-INFINITY, 1))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(-INFINITY, 1.1))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(-INFINITY, 2))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(-INFINITY, 2.1))));
EXPECT_STREQ("-INFINITY", _gc(xdtoal(powl(-INFINITY, 3))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(-INFINITY, 3.1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(INFINITY, -1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(INFINITY, -.1))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(INFINITY, 1))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(INFINITY, .1))));
EXPECT_STREQ("1", _gc(xdtoal(powl(INFINITY, 0))));
EXPECT_STREQ("1", _gc(xdtoal(powl(INFINITY, -0.))));
EXPECT_STREQ("1", _gc(xdtoal(powl(0, 0))));
EXPECT_STREQ("1", _gc(xdtoal(powl(0, -0.))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(0, -(MAX(rando, 1) | 1)))));
EXPECT_STREQ("-INFINITY", _gc(xdtoal(powl(-0., -(MAX(rando, 1) | 1)))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(0, -(rando & -2)))));
EXPECT_STREQ("INFINITY", _gc(xdtoal(powl(-0., -(rando & -2)))));
Handle more pow cases (#61)
2021-03-04 14:13:32 +00:00
EXPECT_TRUE(isnan(powl(-3, 1. / MAX(rando, 2))));
EXPECT_TRUE(isnan(powl(-3, -(1. / MAX(rando, 2)))));
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("-.3333333333333333", _gc(xdtoal(powl(-3, -1))));
EXPECT_STREQ(".1111111111111111", _gc(xdtoal(powl(-3, -2))));
EXPECT_STREQ("-0", _gc(xdtoal(powl(-0., MAX(1, rando) | 1))));
EXPECT_STREQ("0", _gc(xdtoal(powl(-0., MAX(1, rando) & ~1))));
Initial import
2020-06-15 14:18:57 +00:00
}
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
TEST(pow, test) {
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_STREQ("27", fmtd(pow(3, 3)));
EXPECT_STREQ("-27", fmtd(pow(-3, 3)));
EXPECT_STREQ("1e+308", fmtd(pow(10, 308)));
EXPECT_STREQ("inf", fmtd(pow(10, 309)));
EXPECT_STREQ("0", fmtd(pow(10, -5000)));
EXPECT_STREQ("1.06338239662793e+37", fmtd(pow(2, 123)));
EXPECT_STARTSWITH("0.42484968054675", fmtd(pow(.7, 2.4)));
EXPECT_STREQ("1", fmtd(pow(1, NAN)));
EXPECT_STREQ("1", fmtd(pow(1, rando)));
EXPECT_STREQ("1", fmtd(pow(NAN, 0)));
EXPECT_STREQ("1", fmtd(pow(rando, 0)));
EXPECT_STREQ("0", fmtd(pow(0, 1)));
EXPECT_STREQ("0", fmtd(pow(0, 2)));
EXPECT_STREQ("0", fmtd(pow(0, 2.1)));
EXPECT_STREQ("1", fmtd(pow(-1, INFINITY)));
EXPECT_STREQ("1", fmtd(pow(-1, -INFINITY)));
EXPECT_STREQ("inf", fmtd(pow(1. / MAX(2, rando), -INFINITY)));
EXPECT_STREQ("0", fmtd(pow(1.1, -INFINITY)));
EXPECT_STREQ("0", fmtd(pow(MAX(2, rando), -INFINITY)));
EXPECT_STREQ("0", fmtd(pow(1. / MAX(2, rando), INFINITY)));
EXPECT_STREQ("inf", fmtd(pow(MAX(2, rando), INFINITY)));
EXPECT_STREQ("-0", fmtd(pow(-INFINITY, -1)));
EXPECT_STREQ("0", fmtd(pow(-INFINITY, -1.1)));
EXPECT_STREQ("0", fmtd(pow(-INFINITY, -2)));
EXPECT_STREQ("0", fmtd(pow(-INFINITY, -2.1)));
EXPECT_STREQ("-0", fmtd(pow(-INFINITY, -3)));
EXPECT_STREQ("0", fmtd(pow(-INFINITY, -3.1)));
EXPECT_STREQ("-inf", fmtd(pow(-INFINITY, 1)));
EXPECT_STREQ("inf", fmtd(pow(-INFINITY, 1.1)));
EXPECT_STREQ("inf", fmtd(pow(-INFINITY, 2)));
EXPECT_STREQ("inf", fmtd(pow(-INFINITY, 2.1)));
EXPECT_STREQ("-inf", fmtd(pow(-INFINITY, 3)));
EXPECT_STREQ("inf", fmtd(pow(-INFINITY, 3.1)));
EXPECT_STREQ("0", fmtd(pow(INFINITY, -1)));
EXPECT_STREQ("0", fmtd(pow(INFINITY, -.1)));
EXPECT_STREQ("inf", fmtd(pow(INFINITY, 1)));
EXPECT_STREQ("inf", fmtd(pow(INFINITY, .1)));
EXPECT_STREQ("1", fmtd(pow(INFINITY, 0)));
EXPECT_STREQ("1", fmtd(pow(INFINITY, -0.)));
EXPECT_STREQ("1", fmtd(pow(0, 0)));
EXPECT_STREQ("1", fmtd(pow(0, -0.)));
EXPECT_STREQ("inf", fmtd(pow(0, -(MAX(rando, 1) | 1))));
EXPECT_STREQ("-inf", fmtd(pow(-0., -(MAX(rando, 1) | 1))));
EXPECT_STREQ("inf", fmtd(pow(0, -(rando & -2))));
EXPECT_STREQ("inf", fmtd(pow(-0., -(rando & -2))));
EXPECT_STREQ("-0.333333333333333", fmtd(pow(-3, -1)));
EXPECT_STREQ("0.111111111111111", fmtd(pow(-3, -2)));
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("-0", _gc(xdtoa(pow(-0., MAX(1, rando) | 1))));
EXPECT_STREQ("0", _gc(xdtoa(pow(-0., MAX(1, rando) & ~1))));
Initial import
2020-06-15 14:18:57 +00:00
}
Make pow() conform to standard definition
2021-03-03 17:05:21 +00:00
TEST(powf, test) {
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_STREQ("27", fmtf(powf(3, 3)));
EXPECT_STREQ("-27", fmtf(powf(-3, 3)));
EXPECT_STREQ("1e+38", fmtf(powf(10, 38)));
EXPECT_STREQ("inf", fmtf(powf(10, 39)));
EXPECT_STREQ("0", fmtf(powf(10, -5000)));
EXPECT_STREQ("1.06338e+37", fmtf(powf(2, 123)));
EXPECT_STARTSWITH("0.42485", fmtf(powf(.7, 2.4)));
EXPECT_STREQ("1", fmtf(powf(1, NAN)));
EXPECT_STREQ("1", fmtf(powf(1, rando)));
EXPECT_STREQ("1", fmtf(powf(NAN, 0)));
EXPECT_STREQ("1", fmtf(powf(rando, 0)));
EXPECT_STREQ("0", fmtf(powf(0, 1)));
EXPECT_STREQ("0", fmtf(powf(0, 2)));
EXPECT_STREQ("0", fmtf(powf(0, 2.1)));
EXPECT_STREQ("1", fmtf(powf(-1, INFINITY)));
EXPECT_STREQ("1", fmtf(powf(-1, -INFINITY)));
EXPECT_STREQ("inf", fmtf(powf(1. / MAX(2, rando), -INFINITY)));
EXPECT_STREQ("0", fmtf(powf(1.1, -INFINITY)));
EXPECT_STREQ("0", fmtf(powf(MAX(2, rando), -INFINITY)));
EXPECT_STREQ("0", fmtf(powf(1. / MAX(2, rando), INFINITY)));
EXPECT_STREQ("inf", fmtf(powf(MAX(2, rando), INFINITY)));
EXPECT_STREQ("-0", fmtf(powf(-INFINITY, -1)));
EXPECT_STREQ("0", fmtf(powf(-INFINITY, -1.1)));
EXPECT_STREQ("0", fmtf(powf(-INFINITY, -2)));
EXPECT_STREQ("0", fmtf(powf(-INFINITY, -2.1)));
EXPECT_STREQ("-0", fmtf(powf(-INFINITY, -3)));
EXPECT_STREQ("0", fmtf(powf(-INFINITY, -3.1)));
EXPECT_STREQ("-inf", fmtf(powf(-INFINITY, 1)));
EXPECT_STREQ("inf", fmtf(powf(-INFINITY, 1.1)));
EXPECT_STREQ("inf", fmtf(powf(-INFINITY, 2)));
EXPECT_STREQ("inf", fmtf(powf(-INFINITY, 2.1)));
EXPECT_STREQ("-inf", fmtf(powf(-INFINITY, 3)));
EXPECT_STREQ("inf", fmtf(powf(-INFINITY, 3.1)));
EXPECT_STREQ("0", fmtf(powf(INFINITY, -1)));
EXPECT_STREQ("0", fmtf(powf(INFINITY, -.1)));
EXPECT_STREQ("inf", fmtf(powf(INFINITY, 1)));
EXPECT_STREQ("inf", fmtf(powf(INFINITY, .1)));
EXPECT_STREQ("1", fmtf(powf(INFINITY, 0)));
EXPECT_STREQ("1", fmtf(powf(INFINITY, -0.)));
EXPECT_STREQ("1", fmtf(powf(0, 0)));
EXPECT_STREQ("1", fmtf(powf(0, -0.)));
EXPECT_STREQ("inf", fmtf(powf(0, -(MAX(rando, 1) | 1))));
EXPECT_STREQ("-inf", fmtf(powf(-0., -(MAX(rando, 1) | 1))));
EXPECT_STREQ("inf", fmtf(powf(0, -(rando & -2))));
EXPECT_STREQ("inf", fmtf(powf(-0., -(rando & -2))));
EXPECT_STREQ("-0.333333", fmtf(powf(-3, -1)));
EXPECT_STREQ("0.111111", fmtf(powf(-3, -2)));
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("-0", _gc(xdtoaf(powf(-0., MAX(1, rando) | 1))));
EXPECT_STREQ("0", _gc(xdtoaf(powf(-0., MAX(1, rando) & ~1))));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
}
Get libc/tinymath/ compiling on aarch64
2023-05-03 01:35:25 +00:00
#if 0
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
TEST(powl, errors) {
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(0., 0.)));
EXPECT_STREQ("1", fmtd(POW(0., -0.)));
EXPECT_STREQ("0", fmtd(POW(0., .5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(0., -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(0., 1.)));
EXPECT_STREQ("inf", fmtd(POW(0., -1.)));
EXPECT_STREQ("0", fmtd(POW(0., 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(0., -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(0., 2.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(0., -2.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(0., NAN)));
EXPECT_TRUE(isnan(POW(0., -NAN)));
EXPECT_STREQ("0", fmtd(POW(0., INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(0., -INFINITY)));
EXPECT_STREQ("0", fmtd(POW(0., __DBL_MIN__)));
EXPECT_STREQ("0", fmtd(POW(0., __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(-0., 0.)));
EXPECT_STREQ("1", fmtd(POW(-0., -0.)));
EXPECT_STREQ("0", fmtd(POW(-0., .5)));
errno = 0;
EXPECT_STREQ("inf", fmtd(POW(-0., -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("-0", fmtd(POW(-0., 1.)));
EXPECT_STREQ("-inf", fmtd(POW(-0., -1.)));
EXPECT_STREQ("0", fmtd(POW(-0., 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(-0., -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(-0., 2.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(-0., -2.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(ERANGE, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-0., NAN)));
EXPECT_TRUE(isnan(POW(-0., -NAN)));
EXPECT_STREQ("0", fmtd(POW(-0., INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(-0., -INFINITY)));
EXPECT_STREQ("0", fmtd(POW(-0., __DBL_MIN__)));
EXPECT_STREQ("0", fmtd(POW(-0., __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(.5, 0.)));
EXPECT_STREQ("1", fmtd(POW(.5, -0.)));
EXPECT_STREQ("0.707106781186548", fmtd(POW(.5, .5)));
EXPECT_STREQ("1.4142135623731", fmtd(POW(.5, -.5)));
EXPECT_STREQ("0.5", fmtd(POW(.5, 1.)));
EXPECT_STREQ("2", fmtd(POW(.5, -1.)));
EXPECT_STREQ("0.353553390593274", fmtd(POW(.5, 1.5)));
EXPECT_STREQ("2.82842712474619", fmtd(POW(.5, -1.5)));
EXPECT_STREQ("0.25", fmtd(POW(.5, 2.)));
EXPECT_STREQ("4", fmtd(POW(.5, -2.)));
EXPECT_TRUE(isnan(POW(.5, NAN)));
EXPECT_TRUE(isnan(POW(.5, -NAN)));
EXPECT_STREQ("0", fmtd(POW(.5, INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(.5, -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(.5, __DBL_MIN__)));
errno = 0;
EXPECT_STREQ("0", fmtd(POW(.5, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(-.5, 0.)));
EXPECT_STREQ("1", fmtd(POW(-.5, -0.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-.5, .5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-.5, -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("-0.5", fmtd(POW(-.5, 1.)));
EXPECT_STREQ("-2", fmtd(POW(-.5, -1.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-.5, 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-.5, -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0.25", fmtd(POW(-.5, 2.)));
EXPECT_STREQ("4", fmtd(POW(-.5, -2.)));
EXPECT_TRUE(isnan(POW(-.5, NAN)));
EXPECT_TRUE(isnan(POW(-.5, -NAN)));
EXPECT_STREQ("0", fmtd(POW(-.5, INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(-.5, -INFINITY)));
errno = 0;
EXPECT_TRUE(isnan(POW(-.5, __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(-.5, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(1., 0.)));
EXPECT_STREQ("1", fmtd(POW(1., -0.)));
EXPECT_STREQ("1", fmtd(POW(1., .5)));
EXPECT_STREQ("1", fmtd(POW(1., -.5)));
EXPECT_STREQ("1", fmtd(POW(1., 1.)));
EXPECT_STREQ("1", fmtd(POW(1., -1.)));
EXPECT_STREQ("1", fmtd(POW(1., 1.5)));
EXPECT_STREQ("1", fmtd(POW(1., -1.5)));
EXPECT_STREQ("1", fmtd(POW(1., 2.)));
EXPECT_STREQ("1", fmtd(POW(1., -2.)));
EXPECT_STREQ("1", fmtd(POW(1., NAN)));
EXPECT_STREQ("1", fmtd(POW(1., -NAN)));
EXPECT_STREQ("1", fmtd(POW(1., INFINITY)));
EXPECT_STREQ("1", fmtd(POW(1., -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(1., __DBL_MIN__)));
EXPECT_STREQ("1", fmtd(POW(1., __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(-1., 0.)));
EXPECT_STREQ("1", fmtd(POW(-1., -0.)));
errno = 0;
EXPECT_TRUE(isnan(POW(-1., .5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1., -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("-1", fmtd(POW(-1., 1.)));
EXPECT_STREQ("-1", fmtd(POW(-1., -1.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1., 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1., -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(-1., 2.0)));
EXPECT_STREQ("1", fmtd(POW(-1., -2.0)));
EXPECT_TRUE(isnan(POW(-1., NAN)));
EXPECT_TRUE(isnan(POW(-1., -NAN)));
EXPECT_STREQ("1", fmtd(POW(-1., INFINITY)));
EXPECT_STREQ("1", fmtd(POW(-1., -INFINITY)));
errno = 0;
EXPECT_TRUE(isnan(POW(-1., __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(-1., __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(1.5, 0.)));
EXPECT_STREQ("1", fmtd(POW(1.5, -0.)));
EXPECT_STREQ("1.22474487139159", fmtd(POW(1.5, .5)));
EXPECT_STREQ("0.816496580927726", fmtd(POW(1.5, -.5)));
EXPECT_STREQ("1.5", fmtd(POW(1.5, 1.)));
EXPECT_STREQ("0.666666666666667", fmtd(POW(1.5, -1.)));
EXPECT_STREQ("1.83711730708738", fmtd(POW(1.5, 1.5)));
EXPECT_STREQ("0.544331053951817", fmtd(POW(1.5, -1.5)));
EXPECT_STREQ("2.25", fmtd(POW(1.5, 2.0)));
EXPECT_STREQ("0.444444444444444", fmtd(POW(1.5, -2.0)));
EXPECT_TRUE(isnan(POW(1.5, NAN)));
EXPECT_TRUE(isnan(POW(1.5, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(1.5, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(1.5, -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(1.5, __DBL_MIN__)));
errno = 0;
EXPECT_STREQ("inf", fmtd(POW(1.5, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(-1.5, 0.)));
EXPECT_STREQ("1", fmtd(POW(-1.5, -0.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1.5, .5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1.5, -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("-1.5", fmtd(POW(-1.5, 1.)));
EXPECT_STREQ("-0.666666666666667", fmtd(POW(-1.5, -1.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1.5, 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-1.5, -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("2.25", fmtd(POW(-1.5, 2.0)));
EXPECT_STREQ("0.444444444444444", fmtd(POW(-1.5, -2.0)));
EXPECT_TRUE(isnan(POW(-1.5, NAN)));
EXPECT_TRUE(isnan(POW(-1.5, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(-1.5, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(-1.5, -INFINITY)));
errno = 0;
EXPECT_TRUE(isnan(POW(-1.5, __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(-1.5, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(+2.0, 0.)));
EXPECT_STREQ("1", fmtd(POW(+2.0, -0.)));
EXPECT_STREQ("1.4142135623731", fmtd(POW(+2.0, .5)));
EXPECT_STREQ("0.707106781186548", fmtd(POW(+2.0, -.5)));
EXPECT_STREQ("2", fmtd(POW(+2.0, 1.)));
EXPECT_STREQ("0.5", fmtd(POW(+2.0, -1.)));
EXPECT_STREQ("2.82842712474619", fmtd(POW(+2.0, 1.5)));
EXPECT_STREQ("0.353553390593274", fmtd(POW(+2.0, -1.5)));
EXPECT_STREQ("4", fmtd(POW(+2.0, 2.0)));
EXPECT_STREQ("0.25", fmtd(POW(+2.0, -2.0)));
EXPECT_TRUE(isnan(POW(+2.0, NAN)));
EXPECT_TRUE(isnan(POW(+2.0, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(+2.0, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(+2.0, -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(+2.0, __DBL_MIN__)));
errno = 0;
EXPECT_STREQ("inf", fmtd(POW(+2.0, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(-2.0, 0.)));
EXPECT_STREQ("1", fmtd(POW(-2.0, -0.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-2.0, .5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-2.0, -.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("-2", fmtd(POW(-2.0, 1.)));
EXPECT_STREQ("-0.5", fmtd(POW(-2.0, -1.)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-2.0, 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(-2.0, -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("4", fmtd(POW(-2.0, 2.0)));
EXPECT_STREQ("0.25", fmtd(POW(-2.0, -2.0)));
EXPECT_TRUE(isnan(POW(-2.0, NAN)));
EXPECT_TRUE(isnan(POW(-2.0, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(-2.0, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(-2.0, -INFINITY)));
errno = 0;
EXPECT_TRUE(isnan(POW(-2.0, __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
EXPECT_EQ(EDOM, errno);
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(-2.0, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(NAN, 0.)));
EXPECT_STREQ("1", fmtd(POW(NAN, -0.)));
EXPECT_TRUE(isnan(POW(NAN, .5)));
EXPECT_TRUE(isnan(POW(NAN, -.5)));
EXPECT_TRUE(isnan(POW(NAN, 1.)));
EXPECT_TRUE(isnan(POW(NAN, -1.)));
EXPECT_TRUE(isnan(POW(NAN, 1.5)));
EXPECT_TRUE(isnan(POW(NAN, -1.5)));
EXPECT_TRUE(isnan(POW(NAN, 2.0)));
EXPECT_TRUE(isnan(POW(NAN, -2.0)));
EXPECT_TRUE(isnan(POW(NAN, NAN)));
EXPECT_TRUE(isnan(POW(NAN, -NAN)));
EXPECT_TRUE(isnan(POW(NAN, INFINITY)));
EXPECT_TRUE(isnan(POW(NAN, -INFINITY)));
EXPECT_TRUE(isnan(POW(NAN, __DBL_MIN__)));
EXPECT_TRUE(isnan(POW(NAN, __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(-NAN, 0.)));
EXPECT_STREQ("1", fmtd(POW(-NAN, -0.)));
EXPECT_TRUE(isnan(POW(-NAN, .5)));
EXPECT_TRUE(isnan(POW(-NAN, -.5)));
EXPECT_TRUE(isnan(POW(-NAN, 1.)));
EXPECT_TRUE(isnan(POW(-NAN, -1.)));
EXPECT_TRUE(isnan(POW(-NAN, 1.5)));
EXPECT_TRUE(isnan(POW(-NAN, -1.5)));
EXPECT_TRUE(isnan(POW(-NAN, 2.0)));
EXPECT_TRUE(isnan(POW(-NAN, -2.0)));
EXPECT_TRUE(isnan(POW(-NAN, NAN)));
EXPECT_TRUE(isnan(POW(-NAN, -NAN)));
EXPECT_TRUE(isnan(POW(-NAN, INFINITY)));
EXPECT_TRUE(isnan(POW(-NAN, -INFINITY)));
EXPECT_TRUE(isnan(POW(-NAN, __DBL_MIN__)));
EXPECT_TRUE(isnan(POW(-NAN, __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(INFINITY, 0.)));
EXPECT_STREQ("1", fmtd(POW(INFINITY, -0.)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, .5)));
EXPECT_STREQ("0", fmtd(POW(INFINITY, -.5)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, 1.)));
EXPECT_STREQ("0", fmtd(POW(INFINITY, -1.)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, 1.5)));
EXPECT_STREQ("0", fmtd(POW(INFINITY, -1.5)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, 2.0)));
EXPECT_STREQ("0", fmtd(POW(INFINITY, -2.0)));
EXPECT_TRUE(isnan(POW(INFINITY, NAN)));
EXPECT_TRUE(isnan(POW(INFINITY, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(INFINITY, -INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, __DBL_MIN__)));
EXPECT_STREQ("inf", fmtd(POW(INFINITY, __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(-INFINITY, 0.)));
EXPECT_STREQ("1", fmtd(POW(-INFINITY, -0.)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, .5)));
EXPECT_STREQ("0", fmtd(POW(-INFINITY, -.5)));
EXPECT_STREQ("-inf", fmtd(POW(-INFINITY, 1.)));
EXPECT_STREQ("-0", fmtd(POW(-INFINITY, -1.)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, 1.5)));
EXPECT_STREQ("0", fmtd(POW(-INFINITY, -1.5)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, 2.0)));
EXPECT_STREQ("0", fmtd(POW(-INFINITY, -2.0)));
EXPECT_TRUE(isnan(POW(-INFINITY, NAN)));
EXPECT_TRUE(isnan(POW(-INFINITY, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(-INFINITY, -INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, __DBL_MIN__)));
EXPECT_STREQ("inf", fmtd(POW(-INFINITY, __DBL_MAX__)));
EXPECT_STREQ("1", fmtd(POW(__DBL_MIN__, 0.)));
EXPECT_STREQ("1", fmtd(POW(__DBL_MIN__, -0.)));
EXPECT_STREQ("1.49166814624004e-154", fmtd(POW(__DBL_MIN__, .5)));
EXPECT_STREQ("6.7039039649713e+153", fmtd(POW(__DBL_MIN__, -.5)));
EXPECT_STREQ("2.2250738585072e-308", fmtd(POW(__DBL_MIN__, 1.)));
EXPECT_STREQ("4.49423283715579e+307", fmtd(POW(__DBL_MIN__, -1.)));
errno = 0;
EXPECT_STREQ("0", fmtd(POW(__DBL_MIN__, 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(__DBL_MIN__, -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(__DBL_MIN__, 2.0)));
EXPECT_STREQ("inf", fmtd(POW(__DBL_MIN__, -2.0)));
EXPECT_TRUE(isnan(POW(__DBL_MIN__, NAN)));
EXPECT_TRUE(isnan(POW(__DBL_MIN__, -NAN)));
EXPECT_STREQ("0", fmtd(POW(__DBL_MIN__, INFINITY)));
EXPECT_STREQ("inf", fmtd(POW(__DBL_MIN__, -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(__DBL_MIN__, __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* errno = 0; */ /* wut */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
/* EXPECT_STREQ("0", fmtd(POW(__DBL_MIN__, __DBL_MAX__))); */
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("1", fmtd(POW(__DBL_MAX__, 0.)));
EXPECT_STREQ("1", fmtd(POW(__DBL_MAX__, -0.)));
EXPECT_STREQ("1.34078079299426e+154", fmtd(POW(__DBL_MAX__, .5)));
EXPECT_STREQ("7.45834073120021e-155", fmtd(POW(__DBL_MAX__, -.5)));
EXPECT_STREQ("1.79769313486232e+308", fmtd(POW(__DBL_MAX__, 1.)));
EXPECT_STREQ("5.562684646268e-309", fmtd(POW(__DBL_MAX__, -1.)));
errno = 0;
EXPECT_STREQ("inf", fmtd(POW(__DBL_MAX__, 1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(__DBL_MAX__, -1.5)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(__DBL_MAX__, 2.0)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("0", fmtd(POW(__DBL_MAX__, -2.0)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_TRUE(isnan(POW(__DBL_MAX__, NAN)));
EXPECT_TRUE(isnan(POW(__DBL_MAX__, -NAN)));
EXPECT_STREQ("inf", fmtd(POW(__DBL_MAX__, INFINITY)));
EXPECT_STREQ("0", fmtd(POW(__DBL_MAX__, -INFINITY)));
EXPECT_STREQ("1", fmtd(POW(__DBL_MAX__, __DBL_MIN__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
errno = 0;
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EXPECT_STREQ("inf", fmtd(POW(__DBL_MAX__, __DBL_MAX__)));
Make gdtoa less tunable (#104) The -fno-math-errno flag shouldn't impact libraries since it's mostly intended for permitting the compiler to generate sqrt() instructions.
2021-03-08 00:34:16 +00:00
/* EXPECT_EQ(ERANGE, errno); */
Pay off more technical debt This makes breaking changes to add underscores to many non-standard function names provided by the c library. MODE=tiny is now tinier and we now use smaller locks that are better for tiny apps in this mode. Some headers have been renamed to be in the same folder as the build package, so it'll be easier to know which build dependency is needed. Certain old misguided interfaces have been removed. Intel intrinsics headers are now listed in libc/isystem (but not in the amalgamation) to help further improve open source compatibility. Header complexity has also been reduced. Lastly, more shell scripts are now available.
2022-09-13 06:10:38 +00:00
EXPECT_STREQ("1", _gc(xasprintf("%.15g", POW(0., 0))));
EXPECT_STREQ("1", _gc(xasprintf("%.15g", POW(-0., 0))));
EXPECT_STREQ("-0", _gc(xasprintf("%.15g", POW(-0., 1))));
EXPECT_STREQ("-0", _gc(xasprintf("%.15g", POW(-0., 11))));
EXPECT_STREQ("-0", _gc(xasprintf("%.15g", POW(-0., 111))));
EXPECT_STREQ("0", _gc(xasprintf("%.15g", POW(-0., 2))));
EXPECT_STREQ("0", _gc(xasprintf("%.15g", POW(-0., 22))));
EXPECT_STREQ("0", _gc(xasprintf("%.15g", POW(-0., 222))));
EXPECT_STREQ("0", _gc(xasprintf("%.15g", POW(-0., 2.5))));
Initial import
2020-06-15 14:18:57 +00:00
}
Get libc/tinymath/ compiling on aarch64
2023-05-03 01:35:25 +00:00
#endif
Initial import
2020-06-15 14:18:57 +00:00
BENCH(powl, bench) {
Improve signal handling and math - Polyfill ucontext_t on FreeBSD/OpenBSD/NetBSD - Add tests confirming signals can edit CPU state - Work towards supporting ZIP filesystem on bare metal - Add more tinymath unit tests for POSIX conformance - Add X87 and SSE status flags to crash report - Fix some bugs in blinkenlights - Fix llvm build breakage
2021-02-26 02:30:17 +00:00
double _pow(double, double) asm("pow");
float _powf(float, float) asm("powf");
long double _powl(long double, long double) asm("powl");
Use ARM's faster math functions on non-tiny builds
2022-07-12 01:34:10 +00:00
EZBENCH2("pow", donothing, _pow(.7, .2)); /* ~18ns */
Import more Musl math
2022-07-12 22:49:11 +00:00
EZBENCH2("powf", donothing, _powf(.7, .2)); /* ~16ns */
Add more math fixes
2021-03-06 19:41:01 +00:00
EZBENCH2("powl", donothing, _powl(.7, .2)); /* ~56ns */
Initial import
2020-06-15 14:18:57 +00:00
}
Reference in a new issue Copy permalink