 pierre
|
61d16e1063
Allow compilation of softfpu unit with 128-bit upport on big endian code (code generated might be wrong)
|
7 gadi atpakaļ |
 Károly Balogh
|
a57d0d51b4
softfpu: make it possible to optionally inline some of the simple functions. also made it possible to replace the MUL32TO64 function with compiler generated code. If the new defines are enabled, they results in 15-30% speedup on m68k (68020+) softfpu code
|
8 gadi atpakaļ |
 florian
|
8c5fdc9dc2
* softfpu patch by Max Nazhalov: Small fixes to softfpu.pp complementing r27215
|
9 gadi atpakaļ |
|
Károly Balogh
|
8d98f31afc
two cases of use casts instead of anding with when extracting 16bit values
|
11 gadi atpakaļ |
 sergei
|
5279693919
* softfpu.pp: fixed two more porting bugs, and also added two typecasts to avoid unnecessary 64-bit operations.
|
11 gadi atpakaļ |
|
sergei
|
ee8fcc8bdb
* softfpu.pp: reverted r28318 and replaced it with a improved solution:
|
11 gadi atpakaļ |
|
Károly Balogh
|
3c8b24eead
a proposed fix to the qword_to_float64 problem
|
11 gadi atpakaļ |
|
Károly Balogh
|
baac5114a7
added a missing comment from the SoftFPU C source. probably explains why qword_to_float64 fails with values which have the uppermost bit set
|
11 gadi atpakaļ |
|
sergei
|
0dea20d8e7
* softfloat: Fixed int64->double and qword->double conversions to take rounding mode into account (actually, brought them up to date with Softfloat-2b C source). Now test/units/math/troundm.pp is consistent for all targets.
|
11 gadi atpakaļ |
|
sergei
|
bf51683e78
* Softfloat: added some typecasts to prevent promotion to 64 bits where it is not intended.
|
11 gadi atpakaļ |
|
sergei
|
96abd7f370
* Softfloat: marked ExtractFloat32*, ExtractFloat64* and PackFloat32 helpers as inline.
|
11 gadi atpakaļ |
|
sergei
|
cf2b4a4dbd
* softfpu cleanup: use Pascal boolean types where appropriate, use 64-bit overload of shift64RightJamming() to simplify things.
|
11 gadi atpakaļ |
|
sergei
|
22e099d000
* sqrt(real): for targets with emulated floating point, invoke float64_sqrt or float32_sqrt from softfpu code. Testing on ARM CPU shows that float64_sqrt executes about twice faster and offers better accuracy than fpc_sqrt_real from genmath.inc.
|
11 gadi atpakaļ |
|
sergei
|
eb984a5adb
* Replaced last remaining C tokens.
|
11 gadi atpakaļ |
|
sergei
|
20cc82dfea
* Two more "&& vs &" porting bugs fixed.
|
11 gadi atpakaļ |
|
sergei
|
5c1f404631
* softfpu.pp: clean up.
|
11 gadi atpakaļ |
|
sergei
|
94a045aa3d
* Moved declarations of TFPURoundingMode,TFPUExceptionMask and TFPUPrecisionMode to System unit. Declarations in Math unit changed to aliases.
|
11 gadi atpakaļ |
|
sergei
|
825900671b
* Typo in r27184.
|
11 gadi atpakaļ |
|
sergei
|
fcfed19c84
* Fixed one of the C porting bugs (wonder how many are still left...)
|
11 gadi atpakaļ |
|
sergei
|
13da05a778
* Factored direct access to softfloat_exception_flags into a helper procedure.
|
11 gadi atpakaļ |
 masta
|
dad4079e0c
Fix shift128Right in rtl/inc/softfpu.pp
|
12 gadi atpakaļ |
|
florian
|
162cb30528
* another patch by Max Nazhalov to fix shifting and rounding of float128
|
12 gadi atpakaļ |
|
florian
|
a0292b8962
* patch by Max Nazhalov to fix more conversion issues of the C code translation, resolves #24691
|
12 gadi atpakaļ |
|
florian
|
caef53e63b
* another patch by Max Nazhalov to fix missing exits in the softfpu code
|
12 gadi atpakaļ |
|
florian
|
7f6aa98553
* restored changes made in r24864 which I accidently overwrote with r24871
|
12 gadi atpakaļ |
|
florian
|
980563d576
* several small errors fixed, float128<->float64 conversion works now
|
12 gadi atpakaļ |
|
florian
|
92477a14b8
o patch by Max Nazholov:
|
12 gadi atpakaļ |
|
pierre
|
aae318403d
Fix a int64<->single conversion failure exposed in test/units/softfpu/sfttst.pp
|
12 gadi atpakaļ |
|
florian
|
f0487f8fde
* remove packed used for nan records in the softfpu code, it served no purpose
|
12 gadi atpakaļ |
|
florian
|
b01e6e1671
* remove duplicate types after r20603, resolves #21548
|
13 gadi atpakaļ |
