Dibyendu Majumdar
|
76accd1930
|
Update copyright notices
|
4 years ago |
Dibyendu Majumdar
|
08157b102c
|
issue #157 rename bytecodes
|
6 years ago |
Dibyendu Majumdar
|
5e552ab4a7
|
issue #157 more renaming of bytecodes
|
6 years ago |
Dibyendu Majumdar
|
489e118012
|
issue #157 more renaming of bytecodes
|
6 years ago |
Dibyendu Majumdar
|
21028d833b
|
issue #157 more renaming of op codes
|
6 years ago |
Dibyendu Majumdar
|
31f4e178ee
|
issue #157 more opcode renaming
|
6 years ago |
Dibyendu Majumdar
|
fc6929aa58
|
issue #157 more opcode renaming
|
6 years ago |
Dibyendu Majumdar
|
7041827d53
|
issue #157 rename the array get/set op codes
|
6 years ago |
Dibyendu Majumdar
|
eec9a49414
|
issue #157: rename GETTABLE_SK and SETTABLE_SK opcodes
|
6 years ago |
Dibyendu Majumdar
|
4676f220a4
|
issue #114 Prepare for implementing fast C calls; the Lua typecode is extended to 16 bits to allow more type variants to be added. The intention is that for light C functions, we can encode return type / parameter types to enable calling some C functions directly without having to go through the Lua api.
|
6 years ago |
Dibyendu Majumdar
|
a97f974db5
|
issue #144 change the size of TValue.tt to byte
|
6 years ago |
Dibyendu Majumdar
|
03485223ed
|
change GETTABLE_I to work on objects that may not be tables
|
7 years ago |
Dibyendu Majumdar
|
4cae255866
|
use version of arith op with branch weights
|
7 years ago |
Dibyendu Majumdar
|
189be74786
|
fix wrong branch weight
|
7 years ago |
Dibyendu Majumdar
|
c0d2b0e3f8
|
issue #115 experiment with LLVM branch weights
|
7 years ago |
Dibyendu Majumdar
|
9eb8f1d98d
|
issue #115 experiment with table performance - use LuaJIt approach of precomputed hashmask
|
7 years ago |
Dibyendu Majumdar
|
465da9a833
|
issue #118 fix LLVM IT generation bug
|
7 years ago |
Dibyendu Majumdar
|
4f3851e8f3
|
issue #105 #106
|
8 years ago |
Dibyendu Majumdar
|
28b0f77edc
|
doc updates and some refactoring
|
8 years ago |
Dibyendu Majumdar
|
b48a0965e4
|
issue #100 generate calls to specialized functions for string keys when operand of OP_SELF is known to be a short string
|
8 years ago |
Dibyendu Majumdar
|
25cac0cd59
|
add test case and update code gen for TOARRAY opcodes
|
8 years ago |
Dibyendu Majumdar
|
00b61f72fc
|
remove some unused code
|
8 years ago |
Dibyendu Majumdar
|
4af5ecca82
|
issue #100 treat ravi integer and number arrays as subtypes of tables in the lua value similar to how integer and number values are handled
|
8 years ago |
Dibyendu Majumdar
|
d6ed6c150f
|
fix JIT compilation issue with the change in ttistable()
|
8 years ago |
Dibyendu Majumdar
|
a2709d8403
|
issue #100 change the way GETTABLE_SK, SETTABLE_SK, GETTABUP_SK and SETTABUP_SK are compiled
|
8 years ago |
Dibyendu Majumdar
|
7665f6b7c4
|
disable JIT for gettable_sk and gettabup_sk as bug discovered, and refactor the interpreter implementation
|
8 years ago |
Dibyendu Majumdar
|
54bdfe6121
|
issue #100
|
8 years ago |
Dibyendu Majumdar
|
30ed0bd4f6
|
issue #100 add JIT compilation for GETTABLE_SK opcode
|
8 years ago |
Dibyendu Majumdar
|
5a83899351
|
issue #100 #70 #88
|
8 years ago |
Dibyendu Majumdar
|
2fcea1cce2
|
issue #100
|
8 years ago |
dibyendumajumdar
|
704b6faf10
|
more work on issue #73
|
9 years ago |
dibyendumajumdar
|
7531acff65
|
more work on issue #73
|
9 years ago |
dibyendumajumdar
|
f302e34ffc
|
some moe work on fixing issue #73 - more to do
|
9 years ago |
Dibyendu Majumdar
|
852cfee3da
|
issue #71 add specialised bytecode for method calls when type is table and key is short string
|
9 years ago |
Dibyendu Majumdar
|
a11f7bbd27
|
setupval type specialized opcodes should not cause JIT failure
|
9 years ago |
Dibyendu Majumdar
|
c4c8dad5b5
|
issue #61 fix bug in compilation
|
9 years ago |
Dibyendu Majumdar
|
aca5b3ef86
|
issue #60 compile TOTAB and MOVETAB
|
9 years ago |
Dibyendu Majumdar
|
2b7a450baf
|
issue #60 add optimized versions of table access with integer keys
|
9 years ago |
Dibyendu Majumdar
|
62e892e6aa
|
issue #60 add optimized versions of table access with integer keys
|
9 years ago |
Dibyendu Majumdar
|
ff1bffb415
|
issue #60 JIT compile GETTABLE_S
|
9 years ago |
Dibyendu Majumdar
|
f04c5febd6
|
issue #60 JIT compile GETTABLE_S
|
9 years ago |
Dibyendu Majumdar
|
3c9ed08585
|
issue #15 if emit_debug_trace() has been called then we can avoid setting savedpc again
|
9 years ago |
Dibyendu Majumdar
|
b4b54485d4
|
issue #15 updated savedpc
|
9 years ago |
Dibyendu Majumdar
|
ed1ac4b70c
|
code reformat
|
9 years ago |
Dibyendu Majumdar
|
da91cad12a
|
continue investigation of OP_RETURN bug
|
9 years ago |
Dibyendu Majumdar
|
c833336108
|
still debugging the bug related to L->top
|
9 years ago |
Dibyendu Majumdar
|
6b06004c44
|
hard bug - L->top not being set correctly when JIT code calls interpreted code
|
9 years ago |
dibyendumajumdar
|
38967d8b9e
|
add option to disable array range checks
|
9 years ago |
Dibyendu Majumdar
|
e36acf332b
|
in arithmetic operations use constants directly when possible
|
9 years ago |
Dibyendu Majumdar
|
6d0bd5e2a5
|
array index - zext instead of trunc; does it help performance?
|
9 years ago |