[!] Yoshi_Memrect no longer adds any overhead to RDP cmds, instead we'll just retrieve from LoadDetailsMap
[~] Small clean ups..
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@439 42e9bfbe-799a-4a2d-bad1-236e862a387a
[!] Moved Yoshi_MemRect to S2DEX, also it'll be used from S2DEX_RDPHalf_0 to avoid adding overhead to TexRect
[!] Fixed debug build
[!] Removed redundancy in makefile
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@432 42e9bfbe-799a-4a2d-bad1-236e862a387a
[!] Patch ucode name table to create custom name tables (Dlist dumps for custom ucodes are accurate now)
[+] Added GBI1_S2DEX ucode name table
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@426 42e9bfbe-799a-4a2d-bad1-236e862a387a
Note : The benefit of modifying our ucode tables directly
1 - We don't need to build any array which can be really expensive if it happens
2 - No need for a another InstructionLookup
Only issue is that I need to find a clean way to undo any patching to avoid creating conflicts
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@423 42e9bfbe-799a-4a2d-bad1-236e862a387a
[!] Removed all custom ucode tables since they aren't needed anymore (saves alot of space to binary)
[!] Set ucode, patch ucodes, set vtx multiplier in one go (ToDo : set ucode names in one go too)
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@412 42e9bfbe-799a-4a2d-bad1-236e862a387a
[!] Make DLParser_PopDL static to hint compiler to inline (saves ~20 ops)
[!] Reverted changes to DLParser_GBI2_GeometryMode, only made code less readable and slower (saves ~10 ops)
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@410 42e9bfbe-799a-4a2d-bad1-236e862a387a
[!] Some clean up, mostly in S2DEX
[!] Reverted changes to DLParser_InitGeometryMode, it only made the code less readable and slower (+10 ops)
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@409 42e9bfbe-799a-4a2d-bad1-236e862a387a
[+] Old TMEM emulation is a compile option in buildconfig.h
[-] removed redundant texture info in struct
[-] TMEM is not using VRAM (was no speed benefit)
git-svn-id: https://subversion.assembla.com/svn/Daedalusx64/trunk@408 42e9bfbe-799a-4a2d-bad1-236e862a387a