 Jonas Maebe
|
1bcc276dcf
* remove implicit typecast to extended automatically inserted for trunc/round
|
10 лет назад |
|
Jonas Maebe
|
9079227c56
* don't give an internalerror when trying to prefetch a regvar or even
|
10 лет назад |
 florian
|
de17e9fa1c
+ cpu capability CPUX86_HAS_CMOV
|
10 лет назад |
|
florian
|
8207e0ef22
+ make use of vfnmsub*/vfmsub*/vfnmadd* instructions if possible
|
11 лет назад |
|
florian
|
33602db802
* try to registers more cleverly during fma code generation
|
11 лет назад |
|
florian
|
18233f2c0c
* fix compilation of i8086 compiler
|
11 лет назад |
|
florian
|
d88d644925
+ support for FMA intrinsic: if there is no hardware support, the compiler throws an error.
|
11 лет назад |
|
Jonas Maebe
|
d452686c39
* moved pbestrealtype from symdef to symcpu
|
11 лет назад |
|
Jonas Maebe
|
6b16c05a98
* fixed compilation for i8086 with 2.6.x (it gave an unreachable code
|
11 лет назад |
 sergei
|
e4819e863b
+ Consider target_info.cprefix in symbol names.
|
11 лет назад |
|
Jonas Maebe
|
4065483a50
* completed thlcgobj.location_force_fpureg(), use it everywhere and removed
|
11 лет назад |
|
florian
|
c00d752813
+ create avx code in abs as well
|
11 лет назад |
|
florian
|
1882ab8d5a
* i8086 has no cpu_capabilites
|
11 лет назад |
|
florian
|
84f7f5f934
* popcnt has a separate cpuid flag so we should separate it from sse
|
11 лет назад |
 nickysn
|
a7b4953795
+ proper (i.e. not broken) implementation of arctan() for the 8087 and 80287
|
11 лет назад |
|
nickysn
|
09a3fc74e3
* use the fwait prefixed FSTCW, followed by FWAIT, instead of FNSTCW in
|
11 лет назад |
|
sergei
|
bca2c464da
* Don't generate x87 instructions for sin and cos on win64, somehow managed to not commit this part in r25995. Mantis #17273.
|
11 лет назад |
|
sergei
|
2b1e5f7014
* Mantis #17273: don't generate x87 instructions on win64 target.
|
11 лет назад |
|
sergei
|
486a7ef81e
* tx86inlinenode.second_popcnt: use emit_ref_reg instead of appending instruction directly, this provides proper PIC handling of the reference. emit_reg_reg part is not strictly necessary, but done for consistency.
|
11 лет назад |
|
florian
|
fc72490166
* handle LOC_*FPUREGISTER correctly when using sse2 sqr, resolve #26408
|
12 лет назад |
|
florian
|
e81d2d1f3b
* basic avx support for floating point operations (use -Cfavx to activate)
|
12 лет назад |
|
Jonas Maebe
|
2dd75e707e
* renamed thlcgobj.tcgsize2orddef to defutil.cgsize_orddef
|
12 лет назад |
|
Jonas Maebe
|
5051453806
+ support for LOC_(C)MMREGISTER in hlcg
|
12 лет назад |
|
nickysn
|
2b583bb265
* do not use fsin and fcos on 286-, since they're 387+
|
12 лет назад |
|
nickysn
|
09249b31b8
* use the generic implementation of second_IncludeExclude in case the cpu is earlier than 386, since then it does not support BTR and BTS
|
12 лет назад |
|
nickysn
|
9f0c23bd4c
* fixed round() on i8086 after r24502
|
12 лет назад |
|
nickysn
|
3665ed4b03
+ enabled returning int64 in ax:bx:cx:dx for the i8086
|
12 лет назад |
|
nickysn
|
3152df6e7f
* fixed bugs in tx86inlinenode, introduced in r24237
|
12 лет назад |
|
nickysn
|
97ca0fa323
- removed ti8086inlinenode.second_round_real, second_trunc_real and load_fpu_location as they mostly repeat the code in tx86inlinenode.load_fpu_location; tx86inlinenode.load_fpu_location adapted for i8086 instead
|
12 лет назад |
|
nickysn
|
44ef72e1a4
* use the generic implementation of abs for the i8086
|
12 лет назад |
