NJ updates his CPS2 emulator again! (10/09/06)

The modification point is to be rather many, but the effect which it corresponds to hardship completely not seeing.With the unreasonable [tsu] lever you abandoned the improvement of speed above this already.The after all place, as for about 7 tenths of processing with the feeling which depends on the processing of video emulation the shank. When it is CPS2, when it is the usual game picture, unless the 1000~1500 about sprite must be drawn in one picture, because, it is rather hard, don't you think? is. If equal to number the combining [e] which is about 2 times that NEOGEO. By the way, doing the sound mixing with ME, speed does not rise and after (on the other hand speed goes down with synchronous processing, sound unstably). If it tries turning sound to OFF, it is understood, but almost there is no load, it is, don't you think? assuming, that (perhaps speed rose, it probably is 1fps).
When after, it compiles with kernel mode, it seems that the person who is thought that speed rises, is, but it does not rise. With only kernel mode such as network the processing and active memory which cannot be used just increases about 300~400KB. With being the case that it is said, temporarily, CPS2 leaving for a while with this, starts CPS1. romcnv.exe was renewed. Correction of trouble of the below-mentioned drawing as for the game which does.　Please do again to draw up cash file. Correcting the trouble of drawing of the game below. Street fighter ZERO 2: Left and right of the victory mark of ranking indication is strange. Part priority of opening demonstration is strange. 　Score of the top and bottom of ranking indication of the [maitei] pan is not processed the mask. 　[ezahoden] pre-world war 2 demonstration is not indicated in normally with D&D S.O.M. Part sprite of [pazurupu] 2 is not indicated. (Bug of ROM converter. Correction being completed). In every game texture size modification to optimum value. 　But, because it was allotted at the value to which problem does not happen originally as for speed the change. It tried not to read the PCM data when sound is made invalid. It corresponds to the paddle input of [pazurupu] 2. (The same method as forgottn of CPS1PSP). Modifying the calculation method of the pallet. Whether fading out and the like somewhat became good. C68K substantially rewriting. (The M68000 core was drawn up anew, but bug correction and the like Because it is not in time, utilizing continuation C68K. Because speed does not change almost, This way whether it does,) Z80 modification to CZ80 which customizes the core. CZ80 is to be lacking a little in accuracy, but　Because it is not the strange sound driver different from NEOGEO, especially problem seems that does not occur. 　Because there is a bug, it corrected. There is no compatibility of the state data attendant upon above-mentioned and modification. it is the schedule which even in formal edition and is modified. Is not bug, is, but the case where with the grapple game Raster Effects is turned to OFF the gap of the background part　Cancellation. However, because drawing of the picture increases, (about average 1~2fps) speed falls somewhat.
　* Really, as for speed it is to improve delicately but you deducted with this correction and when such as zero whether. Other things, detailed modification point large number.

-For the time being, the fixes for the M68000 bug are finished, and now if I can root out the CZ80 bugs it looks like I'll be able to release the fix version. Upon reverting to test 5's Z80, the speed dropped quite a bit, so I'd like to stick with CZ80 as much as possible.
-The CZ80 bugs are also fixed, I've checked all of the opcodes and found problems with some cycle counts and flag math and have fixed them. Maybe I'll be able to release this tonight?