sepp256
06e563be0c
Fixed FB tex + 2xSaI tex combination.
2009-10-21 21:57:29 +00:00
sepp256
5ee503b649
Moving headers and changing types in order to compile sepp256/ branch with libogc svn.
2009-10-21 21:56:53 +00:00
sepp256
29173c2ab6
Moved xfb back to MEM1 and restored rendering to full xfb.
...
These features are just commented out for now so we can test them some more.
2009-10-21 21:56:14 +00:00
sepp256
d3644ff615
Fixed 2xSaI conversion bug for GX_TF_A3RGBA5.
...
Fixed 2xSaI scaling for tex rects.
Added 2xSaI texture menu item for glN64 build.
2009-10-21 21:55:32 +00:00
sepp256
901b662e48
Use larger xfb for composite cable also. Revert this later.
2009-10-21 21:54:52 +00:00
sepp256
6a6519f48f
Moved xfb's to MEM2.
...
When using component cable for Wii, GX copies a 640 wide FB into a 720 wide xfb. (revert this later)
Something (libogc?) overwrites the bottom 1/4 of one xfb during the splash screen and again just before ROM load. (This is easier to see with component cables because GX isn't overwriting the whole xfb.)
2009-10-21 21:54:22 +00:00
sepp256
670473d035
Moved xfb's to MEM2 to free up another 1.3MB of MEM1.
...
Emu_kidid: please test the performance. You can switch back to MEM1 by changing main_gc-menu.c line388
Also, HBC enables video and draws garbage over the xfb during reload. Perhaps move xfb to higher in MEM2 to fix this issue.
2009-10-21 21:53:38 +00:00
sepp256
85d06292e2
Changed Vector to guVector so we can compile with libogc svn.
...
Removed external gettick declaration.
2009-10-21 21:53:05 +00:00
tehpola
f8dd17b4db
Rearranged includes to be compatible with newest libogc code
2009-10-21 12:49:24 +00:00
emukidid
0318cb2517
Ported latest /tehpola/r4300 core, but with blocks in MEM2. Default settings in the Makefile.GLN64_dev_wii are expansion pak enabled and 8MB recompCache.
2009-10-21 12:47:35 +00:00
emukidid
0a1719d18d
MEM2.h updated with tehpola's memory size macro in preparation for the core to come in from /tehpola/r4300
2009-10-21 12:45:51 +00:00
tehpola
e5253b1958
Fixed dumb things from previous two revisions.
2009-10-20 22:35:30 +00:00
tehpola
86e567cf56
Use fadds, fsubs, fmuls, and fdivs for single precision floating-point rather than always doing calculations in double precision. This should increase performance and compatibility.
2009-10-20 22:34:34 +00:00
tehpola
2c92bdf46d
Fixed sqrt by just calling sqrt instead of trying to do it cleverly. Slower, but generates less instructions.
2009-10-20 22:33:58 +00:00
tehpola
85c264b7e3
Fixed bug where I would orphan a subtree in a remove. Fixes only stability issue I've seen with func trees.
2009-10-20 22:33:21 +00:00
emukidid
1a24fef146
SQRT interpreted and FuncTree.c added (how was this not in there already?)
2009-10-20 21:26:47 +00:00
emukidid
3e62e1ed75
Fixed save states when expansion pak is used.
2009-10-20 21:25:47 +00:00
emukidid
03b8337341
Clearing of flashram in init_flashram.
...
init_font() only called once now, shouldn't see fonts disappear anymore.
2009-10-20 21:24:46 +00:00
emukidid
fa7ad254c2
Adding /tehpola audio >= fix.
2009-10-20 21:23:28 +00:00
tehpola
315820f760
Separate RecompCache for metadata in MEM2. Little to no affect on games currently.
2009-10-20 20:29:12 +00:00
tehpola
219a3a2544
This audio crackling fix brought to you by WiiSX ;)
2009-10-20 20:28:19 +00:00
sepp256
3a0fcb438d
2xSAI fixes. Some textures are working properly now, but more work needs to be done.
2009-10-20 20:00:11 +00:00
sepp256
a43a13afdf
interpolates are no longer pure virtual
2009-10-20 19:59:51 +00:00
sepp256
5cf0031f61
2xSAI for GX implemented but not working yet.
2009-10-20 19:59:28 +00:00
emukidid
642ccf8c20
Compile fixes for /tehpola branch
2009-10-20 12:48:33 +00:00
emukidid
fa87c50c67
Small fixup in Initialise() which was left unfixed from the ogc_video_reset() days
2009-10-20 12:46:21 +00:00
tehpola
7b4d7fb54e
recompile_block returns a PowerPC_func* so there's no need to try to find it later you recompile the func.
2009-10-20 12:31:11 +00:00
tehpola
12f76e70e1
Fixed pure interpreter invalidate_func to use funcs tree.
2009-10-20 12:30:45 +00:00
tehpola
45f4cce617
PowerPC_func_node is now a BST instead of a LL. Unstable.
2009-10-20 12:30:19 +00:00
tehpola
07adb20716
Don't needlessly research through the funcs linked-list
2009-10-20 12:29:02 +00:00
emukidid
eb97402732
emukidid dir is now gone
2009-10-20 12:27:33 +00:00
emukidid
26c37a14d4
fixup of mem2 rom cache in tehpola, just some missing externs.
2009-10-20 12:23:44 +00:00
emukidid
fbbaedf15f
latest /tehpola -> /googlecode changes of new holes copying and also removal of redundant code from MIPS-to-PPC.c
2009-10-20 12:22:14 +00:00
tehpola
4736be7cf6
When recompiling a function with a hole, copy the hole's holes into the new function
2009-10-20 12:16:21 +00:00
tehpola
6ff511322f
I was conditionally returning to the trampoline in genCheckFP although I had previous established that indeed, I would be taking an exception
2009-10-20 12:15:01 +00:00
emukidid
851bbd16f2
porting latest dynarec changes from /tehpola to /googlecode. ERET and gencheckinterrupt updates
2009-10-20 12:10:35 +00:00
tehpola
a4733da43d
Removed unnecessary comparisons of next_interupt to Count in generated code
2009-10-19 21:40:10 +00:00
tehpola
414eca3b89
Implemented ERET
2009-10-19 21:39:12 +00:00
emukidid
d9581fe3ba
MFC0 recompiled, ported from /tehpola to googlecode.
2009-10-19 21:37:26 +00:00
tehpola
ddd657c779
Implemented MFC0.
2009-10-19 21:36:09 +00:00
emukidid
b6df58bb7d
gencheckfp change to use reg_cop0, rescheduling. both ported from /tehpola to googlecode.
2009-10-19 21:35:01 +00:00
tehpola
14c03b76fc
Check FP Available bit in recompiled code rather than having the called function check.
...
Also including rearranged instructions for improved pipelined performance that I meant to commit earlier.
2009-10-19 21:32:48 +00:00
emukidid
7754970f50
SQRT_FP and recompiler pipeline improvements ported to /googlecode from /tehpola
2009-10-19 21:31:37 +00:00
tehpola
51db3f4674
Use Newton-Raphson refinement to increase accuracy of sqrt instruction (should fix issues with OoT)
2009-10-19 21:29:48 +00:00
emukidid
8e23156b7e
Fixed ROM-Cache-MEM2.c in /tehpola branch.
2009-10-19 21:28:08 +00:00
emukidid
31c3eeffe5
forgot a LB,LH,etc update from tehpola branch, also ported the relative branching stuff.
2009-10-19 21:26:57 +00:00
tehpola
285884de2c
Calls to C code via function pointers held in registers replaced with relative branches.
2009-10-19 21:24:42 +00:00
emukidid
8fc57b0a6a
Porting over of tehpola's latest inlined count update stuff from tehpola to googlecode.
2009-10-19 21:23:23 +00:00
tehpola
d14b1e9487
Inline update_count in generated code.
2009-10-19 14:27:53 +00:00
emukidid
00fea27052
blocks is now located in MEM2, yay 4MB more free space in MEM1. Removed libfat USB stuff cause it's only on the libOGC svn for now. Default RecompCache is now 6MB.
2009-10-19 13:21:57 +00:00
