| .. |
|
aopt386.pas
|
ba5297be37
* support disabling the i386 peephole optimizer with -Oonopeephole
|
11 жил өмнө |
|
cgcpu.pas
|
31edfdc05f
* i386: push references with size OS_F64 using less instructions.
|
11 жил өмнө |
|
cpubase.inc
|
bfbb0c5b9d
* optimize mov/lea
|
12 жил өмнө |
|
cpuelf.pas
|
46f8e78d1f
+ Support GOT/gp-relative constants in GAS and internal assemblers, MIPS and i386.
|
11 жил өмнө |
|
cpuinfo.pas
|
d88d644925
+ support for FMA intrinsic: if there is no hardware support, the compiler throws an error.
|
11 жил өмнө |
|
cpunode.pas
|
4a79481c51
* isolated segment-related functionality of tabsolutevarsym into i386/i8086-
|
11 жил өмнө |
|
cpupara.pas
|
2c02e8a726
- i386, x86_64, SPARC: removed paramanager.getintparaloc overrides, it is handled by generic implementation since r24716.
|
11 жил өмнө |
|
cpupi.pas
|
b1dc518ac4
* removed systems_need_16_byte_stack_alignment and use target_info.stackalign instead
|
13 жил өмнө |
|
cputarg.pas
|
d26f0552a0
* Sync with trunk r23404.
|
12 жил өмнө |
|
csopt386.pas
|
4d5119bf1c
* fixes several issues which cause warnings by the dfa code when using it to detect uninitialized variables
|
11 жил өмнө |
|
daopt386.pas
|
4d5119bf1c
* fixes several issues which cause warnings by the dfa code when using it to detect uninitialized variables
|
11 жил өмнө |
|
hlcgcpu.pas
|
26b53607f8
+ added method reference_reset_base with support for different pointer types to
|
11 жил өмнө |
|
i386att.inc
|
dc628b8969
* x86: Completely skip instructions that do not exist for target CPU bit width. The existing behavior of writing mnemonics and properties but no encoding allows an invalid instruction to be recognized by assembler reader or even generated by compiler, but it but won't assemble anyway.
|
11 жил өмнө |
|
i386atts.inc
|
dc628b8969
* x86: Completely skip instructions that do not exist for target CPU bit width. The existing behavior of writing mnemonics and properties but no encoding allows an invalid instruction to be recognized by assembler reader or even generated by compiler, but it but won't assemble anyway.
|
11 жил өмнө |
|
i386int.inc
|
dc628b8969
* x86: Completely skip instructions that do not exist for target CPU bit width. The existing behavior of writing mnemonics and properties but no encoding allows an invalid instruction to be recognized by assembler reader or even generated by compiler, but it but won't assemble anyway.
|
11 жил өмнө |
|
i386nop.inc
|
842e027a9f
+ prove of concept how FMA4 could be supported in inline assembler
|
11 жил өмнө |
|
i386op.inc
|
dc628b8969
* x86: Completely skip instructions that do not exist for target CPU bit width. The existing behavior of writing mnemonics and properties but no encoding allows an invalid instruction to be recognized by assembler reader or even generated by compiler, but it but won't assemble anyway.
|
11 жил өмнө |
|
i386prop.inc
|
dc628b8969
* x86: Completely skip instructions that do not exist for target CPU bit width. The existing behavior of writing mnemonics and properties but no encoding allows an invalid instruction to be recognized by assembler reader or even generated by compiler, but it but won't assemble anyway.
|
11 жил өмнө |
|
i386tab.inc
|
842e027a9f
+ prove of concept how FMA4 could be supported in inline assembler
|
11 жил өмнө |
|
n386add.pas
|
3a3197ae9c
+ i386: inline 64-bit multiplications if overflow checking is disabled and not optimizing for size. Rough testing on Core 2 Duo shows speed improvement by factor of 5, even despite inlined code does not check for zero upper dwords.
|
11 жил өмнө |
|
n386cal.pas
|
1cec09f979
Add -Caoldwin32gnu ABI option to cope with 'ret $4' issues in old GNU mingw32 C compiler (version below 4.7)
|
11 жил өмнө |
|
n386flw.pas
|
1df293679d
* fixed setting of pi_uses_exception flags and unified usage, they should be set in pass_1
|
11 жил өмнө |
|
n386inl.pas
|
66e82f1655
+ i386: generate optimized code for 64-bit arithmetic shifts by constant amount. Shifts by 63 and by less than 32 take just two instructions, shifts by 32..62 bits are done with 3 instructions.
|
12 жил өмнө |
|
n386ld.pas
|
4a79481c51
* isolated segment-related functionality of tabsolutevarsym into i386/i8086-
|
11 жил өмнө |
|
n386mat.pas
|
5356f17fa5
* i386: switch the div/mod node to shared code, leaving in place the specific optimization for division by power of 2.
|
11 жил өмнө |
|
n386mem.pas
|
338c064beb
* moved x86-specific tpointerdef functionality to architecture-specific
|
11 жил өмнө |
|
n386set.pas
|
d0db391d7c
* cleanup of unused units
|
12 жил өмнө |
|
popt386.pas
|
2ee0c8de45
* i386: For integer comparisons with zero, emit "test $-1,%reg" instead of "test %reg,%reg". It is more spilling-friendly, because it transforms into "test $-1,spilltemp" and does not require a register.
|
11 жил өмнө |
|
r386ari.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386att.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386con.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386dwrf.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386int.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386iri.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386nasm.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386nor.inc
|
283ff05127
* merged avx support in inline assembler developed by Torsten Grundke
|
13 жил өмнө |
|
r386nri.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386num.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386ot.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386rni.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386sri.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
r386stab.inc
|
283ff05127
* merged avx support in inline assembler developed by Torsten Grundke
|
13 жил өмнө |
|
r386std.inc
|
5af873ee5b
* x86 targets: Profiling shows that quite a bit of time is spent in findreg_by_number(), despite it uses binary search. Worse, it is repeated for every piece of register information. Trying to get rid of some of these calls: rearranged registers so that their "opcode" matches 3 LSBs of superregister number (with a few exceptions described at the beginning of x86reg.dat). This allows to lookup opcodes in regval() with O(1) complexity, and removes need in rXXXop.inc files.
|
12 жил өмнө |
|
ra386att.pas
|
757ed4e8d3
* standard assembler reader for i386
|
20 жил өмнө |
|
ra386int.pas
|
6c6bf452ca
* Fixed level 2 comment warnings.
|
17 жил өмнө |
|
rgcpu.pas
|
b7fe6797bf
Merged revisions 2921-2922,2925 via svnmerge from
|
19 жил өмнө |
|
rropt386.pas
|
4d5119bf1c
* fixes several issues which cause warnings by the dfa code when using it to detect uninitialized variables
|
11 жил өмнө |
|
symcpu.pas
|
d404d15c1e
* forgot to commit the changes to i386/symcpu.pas in r27549
|
11 жил өмнө |
sergei
5356f17fa5
* i386: switch the div/mod node to shared code, leaving in place the specific optimization for division by power of 2.