* Fixed a bug in the JCC instruction.
* Changed the cycle count from u32 to u16.
* Added cycle counting to the block linker.
* Optimised the branch exit slightly.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6681 8ced0084-cf51-0410-be5f-012b33b47a6e
assembler for JIT. Replace JIT ToMask() with a different variant. Remove
superfluous zeroWriteBackLog calls(added by me).
Core/Common: Don't bother creating a string and calling into a Logs trigger()
when there is noone listening. Change AtomicLoadAcquire for gcc to just
make the compiler not reorder memory accesses around it instead of doing
a full memory barrier, per the comment in the win32 variant.
Core/AudioCommon: Fix a use of uninitialized variable inside libalsa.
Microbenchmarking results for ToMask variants:(1 000 000 000 iterations):
cpu\variant| shifts | bit scan
intel mobile C2D@2.5GHz | 5.5s | 4.0s
amd athlon64x2@3GHz | 6.1s | 6.4s
(including some constant overhead identical to both variants)
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6667 8ced0084-cf51-0410-be5f-012b33b47a6e
* Completed the JIT versions of the DSP Multiplier instructions (5 instructions added).
* Bug fixed the dec and lsr16 instructions.
* Minor code clean-up.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6657 8ced0084-cf51-0410-be5f-012b33b47a6e
* Completed the JIT versions of the DSP arithmetic instructions (28 instructions added).
* Added JIT versions of maddx and msubx (thanks to LM1234).
x64 only
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6652 8ced0084-cf51-0410-be5f-012b33b47a6e
* Added 21 Arithmetic instructions to the JIT
* Used the DSPAnalyser to identify arithmetic/multiplier instructions that precede a conditional branch. This allows the JIT to skip updating the SR when nothing would read from it. Marked all arithmetic/multiplier instructions in the DSPTable for this purpose.
* Converted CheckExternalInterrupt into ASM
* Fixed a couple instructions in DSP Load/Store
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6633 8ced0084-cf51-0410-be5f-012b33b47a6e
* Added jit versions of 5 DSP LoadStore instructions: srs, lrs, lr, sr, si
* Renamed MainOpFallback to Default for consistency in naming with JIT64
* Made ext_dmem_read and ext_dmem_write more generic for wider use
* Optimised dmem_read and dmem_write slightly
* Added dmem_read_imm and dmem_write_imm optimised versions
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6596 8ced0084-cf51-0410-be5f-012b33b47a6e
* Optimised the updating of g_dsp.pc in the compile loop (code by kiesel-stein)
* Added JIT version of LRI (code by kiesel-stein)
* Added JIT versions of the branch instructions (code by Jack Frost)
* DSP_SendAIBuffer fix (code by Mylek)
* Marked instructions that update g_dsp.pc in the DSP table and updated PC based on the table (speed up)
* Fixed the signed bits not being set properly in the addr instruction
* Created a MainOpFallback function to use interpreted versions of the instructions if necessary (code by kiesel-stein)
* Disabled the jit versions of subarn and addarn as they are slowing down NSMBW
The above work in both x86 and x64 modes.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6582 8ced0084-cf51-0410-be5f-012b33b47a6e
Also add the mask for ADPCM predictor, like r6480 does.
Tested 64 bit linux/windows. At least compiles for 32 bit windows, but
with these binaries most of my guest software fails with
DSPLLE(interpreted and jit) when it gets to do actual work.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6484 8ced0084-cf51-0410-be5f-012b33b47a6e
* Make writeToBackLog private to DSPIntExtOps.cpp
(JIT variants of 'l and 'ln are disabled and broken as is)
* Make zeroing of the backlog conditional on doing an interpreter fallback and
do it at a few more places
* Fix selection of cleanup for extended opcodes.
* Fix the DSP unit tests to correctly emit the function prolog/epilog
(else EBX wouldn't be saved)
* Add a few more DSP unit tests
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6325 8ced0084-cf51-0410-be5f-012b33b47a6e
The block_size seems to be not reliable, even after trying to more closely match
the interpreter by looking at the analysis for (addr+opcode->size-1) for detecting
the end of the block.
Since we need to "calculate" this number shortly before returning to
RunForCycles, it seemed logical to use the (up to now) unused return value
of the blocks.
Improves SMG2 here.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@6266 8ced0084-cf51-0410-be5f-012b33b47a6e
Add some useful checking to J_CC/SetJumpTarget.
Refactor increment/decrement and reuse code with increase/decrease.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@5816 8ced0084-cf51-0410-be5f-012b33b47a6e
* Restore Audio Throttle to function properly, broken by Ayuanx many hundreds of revisions back.
* Simplify DSPLLE JIT dispatcher in preparation for an asm rewrite
* Remove hack that made DSPLLE JIT seem faster than it was by running fewer cycles, but resulting in bad sound. This shows off how mysteriously slow it is - I don't understand why it's not faster. Use the DSPLLE interpreter for now if you want to use DSPLLE.
* Made "DSPLLE on Thread" work properly with correct-ish timing - although the speed benefit is really small now.
If it seems like this change slows anything non-LLE down, try turning off Audio Throttle and use the frame limiter in options instead.
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@5541 8ced0084-cf51-0410-be5f-012b33b47a6e
void pushExtValueFromReg(u16 dreg, u16 sreg);
void popExtValueToReg();
instead of the backlog.
Someone might want to add it to the Unit test
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@5389 8ced0084-cf51-0410-be5f-012b33b47a6e
enabled nr in jit after getting skid_au's help in writing the inscrease_addr_reg.
ector can you please take a look and see if the loop code makes sense? it seems
no one it really sure how loops suppose to work in jit
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@5301 8ced0084-cf51-0410-be5f-012b33b47a6e
