http://re4rainbow.4shared.com/ mirror site: http://www.mediafire.com/re4rainbow Legend: '//blah-blah-blah//' or '/blah-blah-blah/' is re4rainbow's comment. Just FYI. Quoted from: http://code.google.com/p/pcsx2-playground/source/list rev 447: Fixed VS 2008 Express compilation errors by changing afxres.h to afxresmw.h in pcsx2.rc. Added missing cycle ratio mods to some load/store instructions, and set up the cycle ratio mod to be a little more tweak-able. rev 446: Added somewhat more "accurate" cycle timings, by weighting mul/div and load/store instructions to more closely match typical/avg cycle counts for those instructions on the PS2. In addition to helping fix some obscure timing glitches, it also allows the emulator to run a little more efficiently. True accuracy is impossible since we can't realistically emulate the MIPs superscalar 8-stage pipeline, branch predictions, or data cache hits/misses. But at least the "average" cycle timings should be a little closer to the real thing now. Improved stability of both EE/X2 and EE/X3 sync hacks. In particular, some FMVs that broke under the X2 hack should work a bit better now, and the X3 hack should behave a lot better now too (but will still break pretty much any FMV/audio sync). X2 sync hack performance also improved, by maybe 6% or so. rev 445: Moved the default .sln file to a more convenient location. Maybe not more appropriate, but certainly more convenient. ;) rev 444: Added comments for my previous VU0 sync fix, along with some other code cleanups. rev 443: Enabled the slow dvd gamefix again. It makes Tales of the Abyss crash way less (no crashes encountered with it yet). We plan to fix this properly soon, once the real issue is found :) rev 442: Resolve compilation errors in Linux. Leave pages of linker errors. rev 441: Improved EE/VU0 sync -- fixes 'new stage' crash in Naruto Ultimate Ninja 2 (part of Issue 58). Also rolled back some troubleshooting code that made it's way into the previous update. rev 440: Fixed a problem with EE's hsync-mode counters. They weren't detecting overflows and targets accurately. This most notably fixes several issues in Tales of the Abyss. Fixed debug mode build, it no longer causes an assertion immediately on starting a game. Removed the final warning from the new C++ build, and cleaned up some code along the way. Deleted the old SSE2EMU functions, which are no longer used since removing SSE1 support from the recompiler a while back. rev 439: fixed a compiler warning, and removed the 2 'flag' speedhacks. no game seemed to need them; and i think i remember someone said it breaks a game. by 'remove' it means that they act like they're always ON. i.e. Overflow flags aren't set for calculations, since i've only seen this hurt games. (most likely due to the ps2's floats being able to contain twice as much values before overflowing; it also could be a problem with other opcodes not 'reseting' the overflow flags when they're supposed to.) rev 438: Created a BSCPropagate class and cleaned out the #define mess in ir5900tables.c. Fixed TLB build compilation errors. Cleaned up some of the headers and #includes use. rev 437: fixed a warning in ir5900tables.c not 100% sure its correct, but it looks like what was wanted :p rev 436: Team decided to axe 2k5 project files :p rev 435: Converted codebase to use the C++ compiler. The C++ option is force-enabled through the project files manually, instead of renaming files to *.cpp (files may be renamed at a later date). Note: VC2005 project is currently broken, as is the Linux build. Both will be updated soon. rev 434: Fix for the better FPU clamp, thanks to Zeydlitz. rev 433: Let's not use -fomit-frame-pointer for the moment. rev 432: Prevent t2reg from having a chance of being assigned to the same register as t1reg. Turn FPU_ADD_SUB_HACK off for 64 bit builds. rev 431: FPU Extra Overflow Clamping should be more accurate now. it should be the best mode for all games, but of course theres always those picky games... rev 430: fix for last revision. rev 429: added Nneeve's (q.w.e.r.t.y) fixed add/sub code. its on by default, theres a macro to turn it off. just set FPU_ADD_SUB_HACK to 0 to use the old way. and i think i know a way to improve fpu clamping, i'm going to try and code it for my next update. rev 428: Clean up some Vif debugging code. rev 427: Stop pcsx2 from crashing in Linux if a plugin is missing. --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 59.126.238.100 ※ 编辑: choan 来自: 59.126.238.100 (12/18 12:51)