Threading VU1 took a lot of rewrites and new code to make possible (MTGS, microVU, gifUnit...), but we finally got to the point where it was feasible, and now we've done it! (so now everyone can stop complaining that pcsx2 only takes advantages of 2 cores :p).
The speedups in the games that benefit from it are great if you have a cpu with 3+ cores (generally a 10~45% speedup), however games that are GS limited can be a slowdown (especially on dual core cpu's).
The option can be found in the speedhacks section as "MTVU (Multi-Threaded microVU1)". And when enabled it should should show the VU thread-time percentage on the title bar window (Like we currently do for EE/GS/UI threads).
It is listed as a speedhack because in order for threading VU1 to have been a speedup, we need to assume that games will not send gif packets containing Signal/Finish/Label commands from path 1 (vu1's xgkick). The good news is very-few games ever do this, so the compatibility of MTVU is very high (a game that does do this will likely hang).
Note: vs2010 builds and Linux builds need to be updated to include "MTVU.h" and "MTVU.cpp".
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4865 96395faa-99c1-11dd-bbfe-3dabce05a288
Refactored the ProLogicII code into a separate file and changed it to floating point math for easier debugging.
Added a basic ProLogic decoder that doesn't try to do surround separation into L/R.
The setting does not have a GUI option yet, since the PLII decoder is experimental and doesn't give too good results.
To change it, set DplDecodingLevel to one of: 0=none, 1=ProLogic, 2=ProLogicII. The default is 0.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4859 96395faa-99c1-11dd-bbfe-3dabce05a288
Causes internal buffer size allocations to be as small as possible within portaudio and the OS, so it can have bad effects if the number turns out to be too small.
Would be nice if someone with a "real" audio card (Audigy, X-fi or similar) tested WASAPI Exclusive Mode, KS or ASIO with this change.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4851 96395faa-99c1-11dd-bbfe-3dabce05a288
I also enabled WDMKS backend, but I had to change some #defines to get it to compile, and I can't seem to be able to use it on my Win7 machine. If anyone on XP wants to try it, it will be at their own risk. (use WDMKS hostApi name on the .ini)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4850 96395faa-99c1-11dd-bbfe-3dabce05a288
On 4840 I added dynamic tuning (which may have slight negative effect on stability) to adapt to different iterations/sec of different systems, but I didn't like this system too much. On 4840 I also made a small change which, I now understand, factors out the differences between systems in this regard, essentially making dynamic tuning not required anymore. So now dynamic tuning is disabled.
If you enable the SPU2X messages (and the overrun messages), you would see log messages similar to this:
buffers: 44 ms ( 89%), tempo: 1.000000, comp: 1.000, iters: 764, (N-IPS:750 -> avg:50, minokc:50, div:100)
If the iters value (764 at this example) on your system when running at 100% speed (50/60fps) is higher than 1000 or lower than 500, then please comment here and post few of the log lines.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4843 96395faa-99c1-11dd-bbfe-3dabce05a288
I'm not fully happy with it, as it's a bit complicated for my taste, so feedback about how it performs will be appreciated.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4840 96395faa-99c1-11dd-bbfe-3dabce05a288
The problem as I understood is is incomplete COLCLAMP support.
Thanks to sudo for the hack :)
(Unfortunately we couldn't find out more about why these FMV also *all* flicker.)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4837 96395faa-99c1-11dd-bbfe-3dabce05a288
- Fixed a bug with mVU's ESIN implementation. This finally fixes the flower-petal bug in Radiata Stories (and hopefully the clouds-bug in VP2 but havn't tested). Rama and I have wanted to fix this for a long time, but never knew the problem :p
- Improved implementation of some undefined behavior that happens on branches to take into account pipeline stalls.
gif / vif:
- Added some logging code that parses gif packets. It can be enabled by the PRINT_GIF_PACKET macro in Gif.h
- Deleted some old gif code.
- On ReadFIFO_VIF1() clear the out-reg to prevent uninitialized results in case GSreadFIFO() doesn't modify it...
- Create a fallback for GSreadFIFO2() so that its always defined (as long as the plugin implements GSreadFIFO())
pcsx2:
- Organized the virtual folders in vs2008 a bit...
Note: vs2010 and Linux builds need to be modified to add Gif_Logger.cpp and delete GIFpath.cpp
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4835 96395faa-99c1-11dd-bbfe-3dabce05a288
1. Much better behavior at extreme low/high FPS (less skips/lower latency).
2. As a result, I was able to reduce internal latency by 50% (50ms config now equals 100ms before).
Recommended latency config for stretcher: 100ms.
For rhythm games, if your FPS is very stable, I recommend 50ms.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4834 96395faa-99c1-11dd-bbfe-3dabce05a288
Summary of changes...
Fixed menus and HUD for Spyro: A New Beginning and Spyro: the Eternal Night.
Fixed missing dialogs and reduced the VRam usage a bit for Tales of Legendia ("8-bit textures" help a bit more with Vram/speed). Should be playable now in hardware mode. Thanks to rama for figuring it could be hackfixed this way and the initial patch :)
Half-screen fixes for Ty the Tasmanian Tiger 2, X-men Origins: Wolverine, Call of Duty: World at War - Final Fronts.
Modified the Star Wars: Force Unleashed hackfix a bit, shouldn't have the see-thru objects anymore.
Also modified the Metal Gear Solid 3 and Gran Turismo 4 hackfixes as they needed different skip values for each region (PAL/NTSC). Thanx to avih for the region code... and the free C/C++ lessons :P
The hackfix for Sly 3 was still showing stripes for 3-D and two-player modes, fixed for NTSC version now and hopefully works for PAL one too.
Couple GameDB updates after the gif rewrite.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4833 96395faa-99c1-11dd-bbfe-3dabce05a288
(they were crashing with the new Gif rewrite... shouldn't anymore)
Although Jake had added the new "GSgifTransfer" callback to these plugins in the past, he forgot to modify the .def files so pcsx2 can actually use them :p
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4832 96395faa-99c1-11dd-bbfe-3dabce05a288
this should be the only time that an explicit flushing is needed.
doing it this way should be faster than flushing every incomplete gs packet like r4829 was doing...
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4831 96395faa-99c1-11dd-bbfe-3dabce05a288
fixes Bleach bankai problem, and probably other games (wizardry/growlancer 3) but haven't tested...
Also removed a vif FlushA hack that used to be needed for tekken 4/gitaroo man but not anymore...
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4829 96395faa-99c1-11dd-bbfe-3dabce05a288
some bits were kept for reference and should be deleted in the future (along with the entire GIFpath.cpp)
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4823 96395faa-99c1-11dd-bbfe-3dabce05a288
Pcsx2 now has a gifUnit class which acts like the ps2's gif and executes a single gif transfer at a time (and performs path arbitration according to priority and path3 slicing).
This new code is generally a speedup for most games. Particularly VU heavy games like GoW.
This revision breaks old saved state compatibility so don't update if you wish to keep playing with your old saved states.
Leave comments if this revision fixes or breaks anything...
Message to GS Plugin authors:
This new code now uses only 1 gif callback to transfer data to the gs plugin (GSgifTransfer).
pcsx2 also is garaunteed to send full GS primitives to the plugin. So you don't have to account for partial-transfers anymore.
Thanks goes out to shadowlady who tested around 500 games for me :D
Note 1: The old gif code is still in this revision, and can be enabled by the USE_OLD_GIF macro. The old code will be deleted soon.
Note 2: This revision also enables assertion dialogs in devel builds, and changed a lot of assume cases into assertions.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4821 96395faa-99c1-11dd-bbfe-3dabce05a288
Should prevent erratic tempo variations near 100% speeds (compared to the previous implementation), and generally allow much lower latencies while keeping the audio stable (on all audio systems).
E.g. my tests show that at 70ms latency it can properly handle sharp speed variance of 50%-100%. If your game generally doesn't get slower than 70% speed, 50ms latency would probably be fine.
Note that on extreme scenarios (e.g. extreme sharp drops in speed) it doesn't try to keep the audio continues at all costs. If your game exhibits such behavior (audible clicks, etc), increasing the latency a bit at the config would help.
Comments on actual performance/quality on different systems would be greatly appreciated.
git-svn-id: http://pcsx2.googlecode.com/svn/trunk@4820 96395faa-99c1-11dd-bbfe-3dabce05a288
gregory.hainaut