ShuriZma
/
bsnes
mirror of https://github.com/bsnes-emu/bsnes.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
bsnes/higan/processor/spc700/spc700.cpp

279 lines
7.8 KiB
C++
Raw Normal View History

Update to v088r07 release. byuu says: (r05 and r06 were save points between large core modifications) I would really appreciate extensive regression testing (especially around SuperFX, Cx4, ST018, DSP-n, ST-01n, NES, GB) at this point. The most critical core changes should be completed now. And it was an unbelievable amount of restructuring. Changelog: - SuperFX core moved to Processor::GSU - SNES::CPU core moved to Processor::R65816 - SNES::SMP core moved to Processor::SPC700 - NES::CPU core renamed to Processor::R6502 - use filestream to load RAM files from interface - save states store SHA256 instead of CRC32 (CRC32 usage removed entirely from bsnes) - nes/ -> fc/ and NES -> FC - snes/ -> sfc/ and SNES -> SFC - SuperFamicom::MappedRAM::copy uses stream instead of data+size - Linux port uses gcc-4.7 (still using only gcc-4.6 subset, so you can make a gcc-4.6 symlink for now if you like.) - all profiles and all targets compile and work properly All eight instruction set cores have been moved to processor/ now. Consistency's a wonderful thing. The last remnants of NES/SNES are now limited to target-ui code; and the nall/(system) folder names. I'm building with gcc-4.7 on my Linux box now because the resultant binaries are up to 20% faster (seriously) than gcc-4.6.
2012-04-29 06:16:44 +00:00
#include <processor/processor.hpp>
#include "spc700.hpp"
namespace Processor {
#include "algorithms.cpp"
#include "instructions.cpp"
#include "disassembler.cpp"
#include "serialization.cpp"
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
#define op(id, name, ...) case id: return op_##name(__VA_ARGS__);
#define fp(name) &SPC700::op_##name
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
auto SPC700::instruction() -> void {
switch(opcode = readPC()) {
op(0x00, nop)
op(0x01, jst)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x02, set_bit)
op(0x03, branch_bit)
op(0x04, read_dp, fp(or), regs.a)
op(0x05, read_addr, fp(or), regs.a)
op(0x06, read_ix, fp(or))
op(0x07, read_idpx, fp(or))
op(0x08, read_const, fp(or), regs.a)
op(0x09, write_dp_dp, fp(or))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x0a, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x0b, adjust_dp, fp(asl))
op(0x0c, adjust_addr, fp(asl))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x0d, push, regs.p)
op(0x0e, test_addr, 1)
op(0x0f, brk)
op(0x10, branch, regs.p.n == 0)
op(0x11, jst)
op(0x12, set_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x13, branch_bit)
op(0x14, read_dpi, fp(or), regs.a, regs.x)
op(0x15, read_addri, fp(or), regs.x)
op(0x16, read_addri, fp(or), regs.y)
op(0x17, read_idpy, fp(or))
op(0x18, write_dp_const, fp(or))
op(0x19, write_ix_iy, fp(or))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x1a, adjust_dpw, -1)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x1b, adjust_dpx, fp(asl))
op(0x1c, adjust, fp(asl), regs.a)
op(0x1d, adjust, fp(dec), regs.x)
op(0x1e, read_addr, fp(cmp), regs.x)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x1f, jmp_iaddrx)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0x20, set_flag, regs.p.p.bit, 0)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x21, jst)
op(0x22, set_bit)
op(0x23, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x24, read_dp, fp(and), regs.a)
op(0x25, read_addr, fp(and), regs.a)
op(0x26, read_ix, fp(and))
op(0x27, read_idpx, fp(and))
op(0x28, read_const, fp(and), regs.a)
op(0x29, write_dp_dp, fp(and))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x2a, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x2b, adjust_dp, fp(rol))
op(0x2c, adjust_addr, fp(rol))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x2d, push, regs.a)
op(0x2e, bne_dp)
op(0x2f, branch, true)
op(0x30, branch, regs.p.n == 1)
op(0x31, jst)
op(0x32, set_bit)
op(0x33, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x34, read_dpi, fp(and), regs.a, regs.x)
op(0x35, read_addri, fp(and), regs.x)
op(0x36, read_addri, fp(and), regs.y)
op(0x37, read_idpy, fp(and))
op(0x38, write_dp_const, fp(and))
op(0x39, write_ix_iy, fp(and))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x3a, adjust_dpw, +1)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x3b, adjust_dpx, fp(rol))
op(0x3c, adjust, fp(rol), regs.a)
op(0x3d, adjust, fp(inc), regs.x)
op(0x3e, read_dp, fp(cmp), regs.x)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x3f, jsr_addr)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0x40, set_flag, regs.p.p.bit, 1)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x41, jst)
op(0x42, set_bit)
op(0x43, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x44, read_dp, fp(eor), regs.a)
op(0x45, read_addr, fp(eor), regs.a)
op(0x46, read_ix, fp(eor))
op(0x47, read_idpx, fp(eor))
op(0x48, read_const, fp(eor), regs.a)
op(0x49, write_dp_dp, fp(eor))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x4a, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x4b, adjust_dp, fp(lsr))
op(0x4c, adjust_addr, fp(lsr))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x4d, push, regs.x)
op(0x4e, test_addr, 0)
op(0x4f, jsp_dp)
op(0x50, branch, regs.p.v == 0)
op(0x51, jst)
op(0x52, set_bit)
op(0x53, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x54, read_dpi, fp(eor), regs.a, regs.x)
op(0x55, read_addri, fp(eor), regs.x)
op(0x56, read_addri, fp(eor), regs.y)
op(0x57, read_idpy, fp(eor))
op(0x58, write_dp_const, fp(eor))
op(0x59, write_ix_iy, fp(eor))
op(0x5a, read_dpw, fp(cpw))
op(0x5b, adjust_dpx, fp(lsr))
op(0x5c, adjust, fp(lsr), regs.a)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x5d, transfer, regs.a, regs.x)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x5e, read_addr, fp(cmp), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x5f, jmp_addr)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0x60, set_flag, regs.p.c.bit, 0)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x61, jst)
op(0x62, set_bit)
op(0x63, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x64, read_dp, fp(cmp), regs.a)
op(0x65, read_addr, fp(cmp), regs.a)
op(0x66, read_ix, fp(cmp))
op(0x67, read_idpx, fp(cmp))
op(0x68, read_const, fp(cmp), regs.a)
op(0x69, write_dp_dp, fp(cmp))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x6a, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x6b, adjust_dp, fp(ror))
op(0x6c, adjust_addr, fp(ror))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x6d, push, regs.y)
op(0x6e, bne_dpdec)
op(0x6f, rts)
op(0x70, branch, regs.p.v == 1)
op(0x71, jst)
op(0x72, set_bit)
op(0x73, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x74, read_dpi, fp(cmp), regs.a, regs.x)
op(0x75, read_addri, fp(cmp), regs.x)
op(0x76, read_addri, fp(cmp), regs.y)
op(0x77, read_idpy, fp(cmp))
op(0x78, write_dp_const, fp(cmp))
op(0x79, write_ix_iy, fp(cmp))
op(0x7a, read_dpw, fp(adw))
op(0x7b, adjust_dpx, fp(ror))
op(0x7c, adjust, fp(ror), regs.a)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x7d, transfer, regs.x, regs.a)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x7e, read_dp, fp(cmp), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x7f, rti)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0x80, set_flag, regs.p.c.bit, 1)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x81, jst)
op(0x82, set_bit)
op(0x83, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x84, read_dp, fp(adc), regs.a)
op(0x85, read_addr, fp(adc), regs.a)
op(0x86, read_ix, fp(adc))
op(0x87, read_idpx, fp(adc))
op(0x88, read_const, fp(adc), regs.a)
op(0x89, write_dp_dp, fp(adc))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x8a, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x8b, adjust_dp, fp(dec))
op(0x8c, adjust_addr, fp(dec))
op(0x8d, read_const, fp(ld), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x8e, plp)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x8f, write_dp_const, fp(st))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x90, branch, regs.p.c == 0)
op(0x91, jst)
op(0x92, set_bit)
op(0x93, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0x94, read_dpi, fp(adc), regs.a, regs.x)
op(0x95, read_addri, fp(adc), regs.x)
op(0x96, read_addri, fp(adc), regs.y)
op(0x97, read_idpy, fp(adc))
op(0x98, write_dp_const, fp(adc))
op(0x99, write_ix_iy, fp(adc))
op(0x9a, read_dpw, fp(sbw))
op(0x9b, adjust_dpx, fp(dec))
op(0x9c, adjust, fp(dec), regs.a)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0x9d, transfer, regs.s, regs.x)
op(0x9e, div_ya_x)
op(0x9f, xcn)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0xa0, set_flag, regs.p.i.bit, 1)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xa1, jst)
op(0xa2, set_bit)
op(0xa3, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xa4, read_dp, fp(sbc), regs.a)
op(0xa5, read_addr, fp(sbc), regs.a)
op(0xa6, read_ix, fp(sbc))
op(0xa7, read_idpx, fp(sbc))
op(0xa8, read_const, fp(sbc), regs.a)
op(0xa9, write_dp_dp, fp(sbc))
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xaa, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xab, adjust_dp, fp(inc))
op(0xac, adjust_addr, fp(inc))
op(0xad, read_const, fp(cmp), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xae, pull, regs.a)
op(0xaf, sta_ixinc)
op(0xb0, branch, regs.p.c == 1)
op(0xb1, jst)
op(0xb2, set_bit)
op(0xb3, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xb4, read_dpi, fp(sbc), regs.a, regs.x)
op(0xb5, read_addri, fp(sbc), regs.x)
op(0xb6, read_addri, fp(sbc), regs.y)
op(0xb7, read_idpy, fp(sbc))
op(0xb8, write_dp_const, fp(sbc))
op(0xb9, write_ix_iy, fp(sbc))
op(0xba, read_dpw, fp(ldw))
op(0xbb, adjust_dpx, fp(inc))
op(0xbc, adjust, fp(inc), regs.a)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xbd, transfer, regs.x, regs.s)
op(0xbe, das)
op(0xbf, lda_ixinc)
Update to v098r19 release. byuu says: Changelog: - added nall/bit-field.hpp - updated all CPU cores (sans LR35902 due to some complexities) to use BitFields instead of bools - updated as many CPU cores as I could to use BitFields instead of union { struct { uint8_t ... }; }; pairs The speed changes are mostly a wash for this. In some instances, I noticed a ~2-3% speedup (eg SNES emulation), and in others a 2-3% slowdown (eg Famicom emulation.) It's within the margin of error, so it's safe to say it has no impact. This does give us a lot of new useful things, however: - no more manual reconstruction of flag values from lots of left shifts and ORs - no more manual deconstruction of flag values from lots of ANDs - ability to get completely free aliases to flag groups (eg GSU can provide alt2, alt1 and also alt (which is alt2,alt1 combined) - removes the need for the nasty order_lsbN macro hack (eventually will make higan 100% endian independent) - saves us from insane compilers that try and do nasty things with alignment on union-structs - saves us from insane compilers that try to store bit-field bits in reverse order - will allow some really novel new use cases (I'm planning an instant-decode ARM opcode function, for instance.) - reduces code size (we can serialize flag registers in one line instead of one for each flag) However, I probably won't use it for super critical code that's constantly reading out register values (eg PPU MMIO registers.) I think there we would end up with a performance penalty.
2016-06-08 22:26:35 +00:00
op(0xc0, set_flag, regs.p.i.bit, 0)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xc1, jst)
op(0xc2, set_bit)
op(0xc3, branch_bit)
op(0xc4, write_dp, regs.a)
op(0xc5, write_addr, regs.a)
op(0xc6, sta_ix)
op(0xc7, sta_idpx)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xc8, read_const, fp(cmp), regs.x)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xc9, write_addr, regs.x)
op(0xca, set_addr_bit)
op(0xcb, write_dp, regs.y)
op(0xcc, write_addr, regs.y)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xcd, read_const, fp(ld), regs.x)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xce, pull, regs.x)
op(0xcf, mul_ya)
op(0xd0, branch, regs.p.z == 0)
op(0xd1, jst)
op(0xd2, set_bit)
op(0xd3, branch_bit)
op(0xd4, write_dpi, regs.a, regs.x)
op(0xd5, write_addri, regs.x)
op(0xd6, write_addri, regs.y)
op(0xd7, sta_idpy)
op(0xd8, write_dp, regs.x)
op(0xd9, write_dpi, regs.x, regs.y)
op(0xda, stw_dp)
op(0xdb, write_dpi, regs.y, regs.x)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xdc, adjust, fp(dec), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xdd, transfer, regs.y, regs.a)
op(0xde, bne_dpx)
op(0xdf, daa)
op(0xe0, clv)
op(0xe1, jst)
op(0xe2, set_bit)
op(0xe3, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xe4, read_dp, fp(ld), regs.a)
op(0xe5, read_addr, fp(ld), regs.a)
op(0xe6, read_ix, fp(ld))
op(0xe7, read_idpx, fp(ld))
op(0xe8, read_const, fp(ld), regs.a)
op(0xe9, read_addr, fp(ld), regs.x)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xea, set_addr_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xeb, read_dp, fp(ld), regs.y)
op(0xec, read_addr, fp(ld), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xed, cmc)
op(0xee, pull, regs.y)
op(0xef, wait)
op(0xf0, branch, regs.p.z == 1)
op(0xf1, jst)
op(0xf2, set_bit)
op(0xf3, branch_bit)
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
op(0xf4, read_dpi, fp(ld), regs.a, regs.x)
op(0xf5, read_addri, fp(ld), regs.x)
op(0xf6, read_addri, fp(ld), regs.y)
op(0xf7, read_idpy, fp(ld))
op(0xf8, read_dp, fp(ld), regs.x)
op(0xf9, read_dpi, fp(ld), regs.x, regs.y)
op(0xfa, write_dp_dp, fp(st))
op(0xfb, read_dpi, fp(ld), regs.y, regs.x)
op(0xfc, adjust, fp(inc), regs.y)
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
op(0xfd, transfer, regs.a, regs.y)
op(0xfe, bne_ydec)
op(0xff, wait)
Update to v088r07 release. byuu says: (r05 and r06 were save points between large core modifications) I would really appreciate extensive regression testing (especially around SuperFX, Cx4, ST018, DSP-n, ST-01n, NES, GB) at this point. The most critical core changes should be completed now. And it was an unbelievable amount of restructuring. Changelog: - SuperFX core moved to Processor::GSU - SNES::CPU core moved to Processor::R65816 - SNES::SMP core moved to Processor::SPC700 - NES::CPU core renamed to Processor::R6502 - use filestream to load RAM files from interface - save states store SHA256 instead of CRC32 (CRC32 usage removed entirely from bsnes) - nes/ -> fc/ and NES -> FC - snes/ -> sfc/ and SNES -> SFC - SuperFamicom::MappedRAM::copy uses stream instead of data+size - Linux port uses gcc-4.7 (still using only gcc-4.6 subset, so you can make a gcc-4.6 symlink for now if you like.) - all profiles and all targets compile and work properly All eight instruction set cores have been moved to processor/ now. Consistency's a wonderful thing. The last remnants of NES/SNES are now limited to target-ui code; and the nall/(system) folder names. I'm building with gcc-4.7 on my Linux box now because the resultant binaries are up to 20% faster (seriously) than gcc-4.6.
2012-04-29 06:16:44 +00:00
}
}
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
#undef op
Update to v098r16 release. byuu says: Changelog: - GNUmakefile: reverted $(call unique,) to $(strip) - processor/r6502: removed templates; reduces object size from 146.5kb to 107.6kb - processor/lr35902: removed templates; reduces object size from 386.2kb to 197.4kb - processor/spc700: merged op macros for switch table declarations - sfc/coprocessor/sa1: partial cleanups; flattened directory structure - sfc/coprocessor/superfx: partial cleanups; flattened directory structure - sfc/coprocessor/icd2: flattened directory structure - gb/ppu: changed behavior of STAT IRQs Major caveat! The GB/GBC STAT IRQ changes has a major bug in it somewhere that's seriously breaking most games. I'm pushing the WIP anyway, because I believe the changes to be mostly correct. I'd like to get more people looking at these changes, and also try more heavy-handed hacking and diff comparison logging between the previous WIP and this one.
2016-06-05 05:03:21 +00:00
#undef fp
Update to v098r15 release. byuu says: Changelog: - removed template usage from processor/spc700; cleaned up many function names and the switch table - object size: 176.8kb => 127.3kb - source code size: 43.5kb => 37.0kb - fixed processor/r65816 BRK/COP vector regression [hex_usr] - corrected HuC3 unmapped RAM read value; fixes Robopon [endrift] - cosmetic: simplified the butterworth constant calculation [Wolfram|Alpha] The SPC700 core changes took forever, about three hours of work. Only the LR35902 and R6502 still need their template functions removed. The point of this is that it doesn't cause any speed penalty to do so, and it results in smaller binary sizes and faster compilation times.
2016-06-05 04:52:43 +00:00
Update to v088r07 release. byuu says: (r05 and r06 were save points between large core modifications) I would really appreciate extensive regression testing (especially around SuperFX, Cx4, ST018, DSP-n, ST-01n, NES, GB) at this point. The most critical core changes should be completed now. And it was an unbelievable amount of restructuring. Changelog: - SuperFX core moved to Processor::GSU - SNES::CPU core moved to Processor::R65816 - SNES::SMP core moved to Processor::SPC700 - NES::CPU core renamed to Processor::R6502 - use filestream to load RAM files from interface - save states store SHA256 instead of CRC32 (CRC32 usage removed entirely from bsnes) - nes/ -> fc/ and NES -> FC - snes/ -> sfc/ and SNES -> SFC - SuperFamicom::MappedRAM::copy uses stream instead of data+size - Linux port uses gcc-4.7 (still using only gcc-4.6 subset, so you can make a gcc-4.6 symlink for now if you like.) - all profiles and all targets compile and work properly All eight instruction set cores have been moved to processor/ now. Consistency's a wonderful thing. The last remnants of NES/SNES are now limited to target-ui code; and the nall/(system) folder names. I'm building with gcc-4.7 on my Linux box now because the resultant binaries are up to 20% faster (seriously) than gcc-4.6.
2012-04-29 06:16:44 +00:00
}