 sergei
|
5f985602fb
* Replaced fpc_ln_real with modern port from fdlibm/uclibc, it has even better accuracy than damath library.
|
11 years ago |
|
sergei
|
aadb407e31
* Convert int64 to double in more straightforward way.
|
11 years ago |
|
sergei
|
6dd845a183
* Removed TTabCoef type and unused zero members in coefficient arrays (the actual number of coefficients is passed to polevl/p1evl anyways).
|
11 years ago |
 Jonas Maebe
|
bf59a848b4
* fixed compilation of genmath.inc for the JVM target. In the future, please
|
11 years ago |
 nickysn
|
779b92459e
* fixed ldexp() on 16/8-bit CPUs. Among other things, this fixes sin() and cos()
|
11 years ago |
|
sergei
|
472b82c09a
* Removed 'absolute'-style aliasing records to floating-point variables, because it causes internal compiler error on some targets.
|
11 years ago |
|
sergei
|
b48ac658fb
+ Payne/Hanek argument reduction routine, improves accuracy of generic sin/cos functions over larger argument range. Tested on x86_64-win64 and mipsel-linux.
|
11 years ago |
|
sergei
|
949b994a68
- Got rid of "goto".
|
11 years ago |
|
sergei
|
4f0b3f61ec
* fpc_trunc_real: explicitly cast float64.high and float32 to longint, fixes test suite regressions on softfloat targets. These targets have different definitions of float64/float32 (with unsigned fields), causing comparisons like "float64.high<0" to be optimized out.
|
11 years ago |
|
sergei
|
897c8b8f7b
* Cleanup fpc_trunc_real implementation.
|
11 years ago |
|
sergei
|
203efdcd0e
* Let order of fields in float64 record depend on FPC_DOUBLE_HILO_SWAPPED, so messing with this order on every usage of float64 is no longer necessary.
|
11 years ago |
|
sergei
|
d83fbd7602
* Replaced fpc_arctan_real() implementation with one providing a better precision.
|
11 years ago |
|
sergei
|
441e37b8b4
- Eliminated separate extractFloat32* routines, they are used only once and manually inlining them does not harm to readability.
|
11 years ago |
|
sergei
|
797653dd14
* Reverted overflow/underflow handling in fpc_exp_real to original state of C port. In particular, changed constant values to typed constants. Resolves #24839.
|
11 years ago |
|
sergei
|
690995fa12
* RTL: fixed some warnings when compiling for x86_64:
|
12 years ago |
 florian
|
92477a14b8
o patch by Max Nazholov:
|
12 years ago |
|
Jonas Maebe
|
aee5380ae0
* merged trunk up to r20882
|
13 years ago |
|
Jonas Maebe
|
e8dae4f30a
* set result of exp() to 0 in case of underflow with masked exceptions
|
13 years ago |
|
sergei
|
2642403d5b
* float_raise: Reduce amount of threadvar accesses, gains a bit more compact code. Functionality is not changed.
|
13 years ago |
|
Jonas Maebe
|
8a95a04e16
* extracted dynarray helpers from system unit into jdynarr.inc (were
|
14 years ago |
|
Jonas Maebe
|
26cc281390
* fixed real48 to double conversion for 0.0 (patch by "jeng", mantis #16863)
|
15 years ago |
|
Jonas Maebe
|
310ec39757
* fixed + optimized generic round_real
|
17 years ago |
|
Jonas Maebe
|
30a51c2dee
+ support for the different rounding modes in the generic rounding
|
17 years ago |
|
florian
|
81fd66f6f6
* typo in define fixed
|
17 years ago |
 micha
|
83b7f81f81
+ add no-fpu support to compiler/rtl for powerpc-linux (-Cfnone)
|
17 years ago |
 yury
|
5e3462b6fc
* Fixed warnings and notes.
|
17 years ago |
|
yury
|
04b640880c
* fixed exp() for hardfloat on arm.
|
18 years ago |
|
florian
|
5596cd5391
* ensure that qword_to_double is calculated using doubles
|
18 years ago |
 tom_at_work
|
e501fab034
* replaced all HandleError() calls to appropriate float_raise() calls
|
18 years ago |
|
yury
|
fa5865845a
* use float_raise instead of HandleError in fpc_exp_real.
|
18 years ago |
