Used usleep(1) instead of _mm_pause on Linux and Mac OSX. This reduces the CPU usage in the video thread’s idle loop. Added YieldCPU to the DSP LLE loop.
vs2010: add debugfast configs for wxw (now debugfast links faster!). clean up wxw projects a lot.
Experimental commit and one fix for my last commit.
I think that isFifoBusy bring better sync with VI (video interface) because the CPU emulated threads are waiting for DrawDone in BP Register. So, I do some modifications.
1) Rename «IsFifoBusy» by «isPossibleWaitingSetDrawDone»
2) Only activate isPossibleWaitingSetDrawDone when bFF_GPLinkEnable is true in fifo loop «Inmediate mode» that is because in theory this drawsync function is using in this mode.
3) Deactivate isPossibleWaitingSetDrawDone also in SetFinish in PixelEngine, beside when 32 block is finish.
Please regression in yours games thats can bring some FPS more above all with VPS frame limiter ON (Auto, 60, 50, etc).
— Fix waiting in AbortFrame(), please test mp1/mp2 is fixed again.
vs2010: add debugfast configs for all projects except wxw…will do those soon.
Change to the 1.3 part of the Xrandr api for fullscreen mode switching on linux to support multiple monitor setups. There is probably some more work to be done, but at least it won’t shut off one of your monitors anymore.
Fix wiimotes :P
Remove any wx-specific code from Core.
Move NetPlay code from DolphinWX to Core.
Even NetPlay with nowx should be possible now if anyone felt like implementing a CLI for it ;)
Externals build framework for wxWidgets 2.9.2+.
revert some some of the soren changes to make dolphin compile in vs2008 again until sw plugin is completely fixed.
some correction to dlist caching
Another OSX cmake fix (make openal compile)
Check for wxWidgets 2.9.2 (trunk really) which the OS X build now requires.
CPU… (yes)..ID -.-‘
— again CPUID :°D
— CPUID asm changes for x64 systems.
restored inline assembly code
— CPUID 2nd Attempt
* CPUID changes:
— Added full ability to recognize true Intel HTT CPU capability (support to discriminate Intel Core 2 Duo…)
— Attempt to fix a register trashing (issue 4022 generally, linux and mac systems)
Small fix for cmake on osx, more work left to be done
D3D9: Fix wireframe rendering.
OGL: Fix wireframe rendering.
Fixes issue 4062.
Oops, remember to save plugin settings in both close cases.
Modal dialogs must be closed with EndModal() or all kinds of weird
corruption can happen with wx 2.9.
Fixed the new texture hashing functions to take samples into consideration. Fixes issue 4058
Automatically disconnect and reconnect the Wiimotes after a save state has been loaded. This makes the save states a bit more stable for Wii games.
Deal with some clang-inspired buglets.
Respect the initial fullscreen setting in nowx.
Split shader code generation for lighting into a separate file.
D3D9: Fix 2x anisotropic filtering…
* Keeping resolution aligned to back-buffer size if the latter one is changed (can give a huge performance boost… even until to 100% in some cases)
fix silly string in CPUDetect
LTCG (link time code gen) was off in some projects, so turn it off in all of them so that the linker doesn’t have to restart halfway through.
Rework XF register loading a bit and change how registers are arranged in memory. This removes the assumption that all data for a viewport or projection matrix will be available when index 0 is loaded. Fixes issue 3688 and probably breaks old save states (sorry).
The compiler need not obey the static keyword, so to avoid linker
problems, whole functions in .h files need to also be static in
case they are included in several .cpp files.
Also a few other minor LTO fixes.
Whoops, forgot to remove that.
Changes to hi-res textures. Textures now load correctly when loading/saving a savestate, and can be toggled on and off in game.
Changed non-hi-res textures to use MurmurHash3, which has better performance that the previous hash.
vs2010: get rid of some more wxw-related linker warnings.
update Cg on windows to 3.0 November 2010…I tested it and didn’t see the problems mentioned in issue 4009.
actually do what i said for the r7071 log, sorry about that…
update zlib to 1.2.5
«Version 1.2.5 fixes bugs in gzseek() and gzeof() that were present in version 1.2.4 (March 2010). All users are encouraged to upgrade immediately.»
Update miniLZO to 2.0.4. Apparently fixes a bug when used with gcc 4.5
D3D9: Offset vertices used to copy textures by 0.5.
I __think__ this is the actually correct way to do it, but not sure. Someone please have a look at this…
Not sure if this changes anything anyway, since we’re using scissor rects to clip this stuff anyway… Maybe the screen edges weren’t cleaned properly before though.
D3D9: Some fixes to EFB format reinterpretation. Should be more correct now.
Might fix issue 3780.
D3D11: Implement EFB format reinterpretation.
Now someone port this to OpenGL please, kthxbye.
Fix some warnings, fix a wrong string, …
Changed the hash algorithm to CRC32 utilising the SSE4.2 instruction. The algorithm will automatically be used for the Accurate Texture Cache, EFB to RAM and texture id’s when a SSE4.2 capable CPU is detected. It will fallback to the old algorithm if SSE4.2 is not detected. Using CRC32 speeds up the hash algorithm by around 2X.