Rooba
/
ReC98
mirror of https://github.com/nmlgc/ReC98.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity

[C decompilation] [th01/fuuin] Slow 2x VRAM region scaling 

This function raises one of those essential questions about the eventual ports
we'd like to do. I'll explain everything more thoroughly here, since people
who might complain about the ports not being faithful enough need to
understand this.

----

The original plan was aim for "100% frame-perfect" ports and advertise them as
such. However, the PC-98 is not a console with fixed specs. As the name
implies, it's a computer architecture, and a plethora of different, more and
more powerful PC-98 models were released during its lifespan. Even if we only
consider the subset of products that fulfills the minimum requirements to run
the PC-98 Touhou games, that's still a sizable number of systems.

Therefore, the only true definition of a *frame* can be "everything that is
drawn between two Vsync wait calls". Such a *frame* may contain certain
expensive function calls, and certain systems may run these functions slower
than the developer expected, thus effectively leading to more *frames* than
the developer explicitly specified.

This is one of those functions.

Here, we have a scaling function that appears to be written deliberately to
run very slow, which ends up creating the rolling effect you see in the route
selection and the high score and continue screens of TH01. However, that
doesn't change the fact that the function is still CPU-bound, and neither
waits for Vsync nor is iteratively called by something that does. The faster
your CPU, the faster the rolling effect gets… until ultimately, it's faster
than one frame and therefore vanishes altogether. Mind you, this is true on
both emulators and real hardware. The final PC-98 model, the Ra43, had a CPU
clocked at 433 Mhz, and it may have even been instant there.
If you use more optimized algorithm, it also runs faster on the same CPU (I
tried this, and it worked beautifully)… you get the idea.

Still, it may very well be that this algorithm was not a deliberate choice and
simply resulted from a lack of experience, especially since this was ZUN's
first game.

That leaves us with two approaches to porting functions like these:

1) Look at the recommended system requirements ZUN specified, configure the
   PC-98 emulator accordingly, measure how much of the work is done in each
   frame, then rewrite the function to be bound to that specific frame rate…
2) …or just continue using a CPU-bound algorithm, which will pretty much
   complete instantly on any modern system.

I'd argue that 2) is actually the more "faithful" approach. It will run faster
than the typical clock speeds people emulate the games at, and maybe draw a
bit of criticism because of that, but it seems a lot more rational than the
approximation provided by 1). Not to mention that it's undeniably easier to
implement, and hey, a faster game feels a lot better than a slower one, right?

… Oh well, maybe we'll still encounter some kind of CPU-bound animation that
is so essential to the experience that we do want to lock it to a certain
frame rate…
This commit is contained in:
nmlgc 2015-03-09 17:58:30 +01:00
parent 160d4eb69f
commit 44ad3eb4bc
7 changed files with 120 additions and 296 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Makefile.mak
View File

@ -38,7 +38,7 @@ bin\th01\reiiden.exe: bin\th01\reiiden.obj th01\main_12.c th01\main_16.c
$**
|
bin\th01\fuuin.exe: bin\th01\fuuin.obj th01\fuuin_11.c th01\fuuin_13.c
bin\th01\fuuin.exe: bin\th01\fuuin.obj th01\fuuin_11.c th01\fuuin_12.c th01\fuuin_13.c
$(CC) $(CFLAGS) -ml -3 -nbin\th01\ -eFUUIN.EXE @&&|
$**
|

4
ReC98.h
View File

@ -54,8 +54,8 @@ extern char *VRAM_PLANE_G;
extern char *VRAM_PLANE_R;
extern char *VRAM_PLANE_E;
// Parenthesizing it like this avoids a signed 16-bit overflow.
#define PLANE_SIZE (RES_X * (RES_Y / 8))
#define ROW_SIZE (RES_X / 8)
#define PLANE_SIZE (ROW_SIZE * RES_Y)
#define PLANE_DWORD_BLIT(dst, src) \
for(p = 0; p < PLANE_SIZE; p += 4) { \

3
th01/fuuin_11.c
View File

@ -5,8 +5,5 @@
#include "th01\th01.h"
#undef graph_accesspage
#define graph_accesspage graph_accesspage_func
#include "th01\hardware\vplanset.c"
#include "th01\hardware\egccopyr.c"

10
th01/fuuin_12.c Normal file
View File

@ -0,0 +1,10 @@
/* ReC98
* -----
* Code segment #12 of TH01's FUUIN.EXE
*/
#include "th01/th01.h"
#pragma option -Z
#include "th01/grps2xsc.c"

74
th01/grps2xsc.c Normal file
View File

@ -0,0 +1,74 @@
void scale_2x(unsigned long *dst32, int src16)
{
unsigned long mask = 1;
unsigned long srcex = 0;
unsigned long dst_local;
int i;
srcex = src16;
dst_local = 0;
*dst32 = 0;
for(i = 0; i < 16; i++) {
dst_local |= _lrotl(srcex & mask, ((i * 2) + 0) - i);
dst_local |= _lrotl(srcex & mask, ((i * 2) + 1) - i);
mask = _lrotl(mask, 1);
}
mask = 0x00FF00FF; *dst32 |= _lrotl(dst_local & mask, 8);
mask = 0xFF00FF00; *dst32 |= _lrotr(dst_local & mask, 8);
}
typedef struct {
int B, R, G, E;
} vram_planar_16_pixels_t;
void graph_slow_2xscale_region_1_to_0(
int x0, int y0, int x1, int y1, int w1, int h1
)
{
int row_p1 = (x1 / 8) + (y1 * ROW_SIZE);
int row_p0 = (x0 / 8) + (y0 * ROW_SIZE);
int col16;
int row;
vram_planar_16_pixels_t px16;
int px16_nonzero;
for(row = 0; row < h1; row++) {
int p0 = row_p0;
int p1 = row_p1;
for(col16 = 0; col16 < w1 / 16; col16++) {
int scale_p;
graph_accesspage(1);
px16.B = *(int*)(VRAM_PLANE_B + p1);
px16.R = *(int*)(VRAM_PLANE_R + p1);
px16.G = *(int*)(VRAM_PLANE_G + p1);
px16.E = *(int*)(VRAM_PLANE_E + p1);
px16_nonzero = px16.B | px16.R | px16.G | px16.E;
for(scale_p = 0; scale_p < ROW_SIZE * 2; scale_p += ROW_SIZE) {
unsigned long dst32;
unsigned long px32_nonzero;
graph_accesspage(0);
scale_2x(&px32_nonzero, px16_nonzero);
grcg_setcolor_rmw(0);
*(long*)(VRAM_PLANE_B + p0 + scale_p) = px32_nonzero;
grcg_off();
scale_2x(&dst32, px16.B);
*(long*)(VRAM_PLANE_B + p0 + scale_p) |= dst32;
scale_2x(&dst32, px16.R);
*(long*)(VRAM_PLANE_R + p0 + scale_p) |= dst32;
scale_2x(&dst32, px16.G);
*(long*)(VRAM_PLANE_G + p0 + scale_p) |= dst32;
scale_2x(&dst32, px16.E);
*(long*)(VRAM_PLANE_E + p0 + scale_p) |= dst32;
}
p1 += 2;
p0 += 4;
}
row_p0 += ROW_SIZE * 2;
row_p1 += ROW_SIZE;
}
}

8
th01/th01.h
View File

@ -7,6 +7,14 @@
// Hardware
void graph_accesspage_func(int page);
void grcg_setcolor_rmw(int color);
void grcg_off_func(void);
#undef graph_accesspage
#define graph_accesspage graph_accesspage_func
#undef grcg_off
#define grcg_off grcg_off_func
// Game
#define STAGES_PER_SCENE 4

315
th01_fuuin.asm
View File

@ -34,8 +34,6 @@ include libs/BorlandC/regs.inc
extern LXMUL@:proc
extern SCOPY@:proc
extern __fgetc:proc
extern __lrotl:proc
extern __lrotr:proc
extern __mbcjmstojis:proc
extern __mbctype:byte
extern __setargv__:proc ; main() needs both to be set
@ -3177,7 +3175,7 @@ arg_6 = dword ptr 0Ch
push 0
push 0
push 30h ; '0'
call sub_EE15
call _graph_slow_2xscale_region_1_to_0
add sp, 0Ch
push 1
push 0
@ -3199,7 +3197,7 @@ loc_B666:
push 0
push 0
push 30h ; '0'
call sub_EE15
call _graph_slow_2xscale_region_1_to_0
add sp, 0Ch
loc_B67B:
@ -6923,8 +6921,8 @@ sub_D119 endp
; =============== S U B R O U T I N E =======================================
; Attributes: bp-based frame
sub_D13A proc far
public _grcg_setcolor_rmw
_grcg_setcolor_rmw proc far
arg_0 = word ptr 6
@ -6984,7 +6982,7 @@ loc_D17E:
out dx, al
pop bp
retf
sub_D13A endp
_grcg_setcolor_rmw endp
; =============== S U B R O U T I N E =======================================
@ -7057,8 +7055,8 @@ sub_D184 endp
; =============== S U B R O U T I N E =======================================
; Attributes: bp-based frame
sub_D1CE proc far
public _grcg_off_func
_grcg_off_func proc far
push bp
mov bp, sp
mov dx, 7Ch
@ -7066,7 +7064,7 @@ sub_D1CE proc far
out dx, al
pop bp
retf
sub_D1CE endp
_grcg_off_func endp
; =============== S U B R O U T I N E =======================================
@ -7260,12 +7258,12 @@ _s = dword ptr -4
enter 4, 0
mov [bp+_s], 0A8000000h
push 0
call sub_D13A
call _grcg_setcolor_rmw
push 7D0000FFh ; c
pushd [bp+_s] ; s
call _memset
add sp, 0Ah
call sub_D1CE
call _grcg_off_func
leave
retf
sub_D2D6 endp
@ -7304,12 +7302,12 @@ arg_0 = byte ptr 6
mov al, [bp+arg_0]
cbw
push ax
call sub_D13A
call _grcg_setcolor_rmw
push 7D0000FFh ; c
pushd [bp+_s] ; s
call _memset
add sp, 0Ah
call sub_D1CE
call _grcg_off_func
leave
retf
sub_D31C endp
@ -7937,7 +7935,7 @@ arg_4 = word ptr 0Ah
push bp
mov bp, sp
push [bp+arg_4]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
mov bx, [bp+arg_2]
imul bx, 50h
@ -7952,7 +7950,7 @@ arg_4 = word ptr 0Ah
mov es, ax
assume es:nothing
mov es:[bx], dl
call sub_D1CE
call _grcg_off_func
pop bp
retf
sub_D74A endp
@ -8113,7 +8111,7 @@ loc_D834:
cmp byte_134EA, 0
jnz short loc_D898
push [bp+arg_6]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
loc_D898:
@ -8167,7 +8165,7 @@ loc_D8ED:
loc_D8F0:
cmp byte_134EA, 0
jnz short loc_D8FB
call sub_D1CE
call _grcg_off_func
loc_D8FB:
pop di
@ -8257,7 +8255,7 @@ loc_D95A:
shl ax, cl
or [bp+var_6], ax
push [bp+arg_6]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
mov [bp+var_2], si
jmp short loc_D9B1
@ -8274,7 +8272,7 @@ loc_D99D:
loc_D9B1:
cmp [bp+var_2], di
jle short loc_D99D
call sub_D1CE
call _grcg_off_func
loc_D9BA:
pop di
@ -8578,7 +8576,7 @@ loc_DBFB:
cmp byte_134EA, 0
jnz short loc_DC35
push [bp+arg_8]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
loc_DC35:
@ -8861,7 +8859,7 @@ loc_DE93:
loc_DF2B:
cmp byte_134EA, 0
jnz short loc_DF36
call sub_D1CE
call _grcg_off_func
loc_DF36:
pop di
@ -8941,7 +8939,7 @@ loc_DF9D:
loc_DFAE:
push [bp+arg_8]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
mov ax, si
imul ax, 50h
@ -9011,7 +9009,7 @@ loc_E03E:
loc_E048:
cmp [bp+var_4], di
jle short loc_DFD0
call sub_D1CE
call _grcg_off_func
loc_E051:
pop di
@ -9261,7 +9259,7 @@ loc_E1C2:
loc_E1DA:
push [bp+arg_4]
call sub_D13A
call _grcg_setcolor_rmw
pop cx
mov dx, 68h ; 'h'
mov al, 0Bh
@ -9538,7 +9536,7 @@ loc_E41D:
mov dx, 68h ; 'h'
mov al, 0Ah
out dx, al
call sub_D1CE
call _grcg_off_func
pop di
pop si
leave
@ -10703,270 +10701,7 @@ fuuin_11_TEXT ends
; Segment type: Pure code
fuuin_12_TEXT segment byte public 'CODE' use16
assume cs:fuuin_12_TEXT
;org 2
assume es:nothing, ss:nothing, ds:_DATA, fs:nothing, gs:nothing
; =============== S U B R O U T I N E =======================================
; Attributes: bp-based frame
sub_ED32 proc far
var_C = dword ptr -0Ch
var_8 = dword ptr -8
val = dword ptr -4
arg_0 = dword ptr 6
arg_4 = word ptr 0Ah
enter 0Ch, 0
push si
mov [bp+val], 1
mov [bp+var_8], 0
movsx eax, [bp+arg_4]
mov [bp+var_8], eax
mov [bp+var_C], 0
les bx, [bp+arg_0]
mov dword ptr es:[bx], 0
xor si, si
jmp short loc_EDBF
; ---------------------------------------------------------------------------
loc_ED67:
mov ax, si
add ax, ax
sub ax, si
push ax ; count
mov eax, [bp+var_8]
and eax, [bp+val]
push eax ; val
call __lrotl
push dx
push ax
pop eax
add sp, 6
or [bp+var_C], eax
mov ax, si
add ax, ax
inc ax
sub ax, si
push ax ; count
mov eax, [bp+var_8]
and eax, [bp+val]
push eax ; val
call __lrotl
push dx
push ax
pop eax
add sp, 6
or [bp+var_C], eax
push 1 ; count
pushd [bp+val] ; val
call __lrotl
add sp, 6
mov word ptr [bp+val+2], dx
mov word ptr [bp+val], ax
inc si
loc_EDBF:
cmp si, 10h
jl short loc_ED67
mov [bp+val], 0FF00FFh
push 8 ; count
mov eax, [bp+var_C]
and eax, [bp+val]
push eax ; val
call __lrotl
push dx
push ax
pop eax
add sp, 6
les bx, [bp+arg_0]
or es:[bx], eax
mov [bp+val], 0FF00FF00h
push 8 ; count
mov eax, [bp+var_C]
and eax, [bp+val]
push eax ; val
call __lrotr
push dx
push ax
pop eax
add sp, 6
les bx, [bp+arg_0]
or es:[bx], eax
pop si
leave
retf
sub_ED32 endp
; =============== S U B R O U T I N E =======================================
; Attributes: bp-based frame
sub_EE15 proc far
var_1C = dword ptr -1Ch
var_18 = dword ptr -18h
var_14 = word ptr -14h
var_12 = word ptr -12h
var_10 = word ptr -10h
var_E = word ptr -0Eh
var_C = word ptr -0Ch
var_A = word ptr -0Ah
var_8 = word ptr -8
var_6 = word ptr -6
var_4 = word ptr -4
var_2 = word ptr -2
arg_0 = word ptr 6
arg_2 = word ptr 8
arg_4 = word ptr 0Ah
arg_6 = word ptr 0Ch
arg_8 = word ptr 0Eh
arg_A = word ptr 10h
enter 1Ch, 0
push si
push di
mov ax, [bp+arg_4]
mov bx, 8
cwd
idiv bx
mov dx, [bp+arg_6]
imul dx, 50h
add ax, dx
mov [bp+var_2], ax
mov ax, [bp+arg_0]
cwd
idiv bx
mov dx, [bp+arg_2]
imul dx, 50h
add ax, dx
mov [bp+var_4], ax
mov [bp+var_8], 0
jmp loc_EF72
; ---------------------------------------------------------------------------
loc_EE48:
mov di, [bp+var_4]
mov ax, [bp+var_2]
mov [bp+var_14], ax
mov [bp+var_6], 0
jmp loc_EF56
; ---------------------------------------------------------------------------
loc_EE59:
push 1
call _graph_accesspage_func
pop cx
les bx, _VRAM_PLANE_B
add bx, [bp+var_14]
mov ax, es:[bx]
mov [bp+var_10], ax
les bx, _VRAM_PLANE_R
add bx, [bp+var_14]
mov ax, es:[bx]
mov [bp+var_E], ax
les bx, _VRAM_PLANE_G
add bx, [bp+var_14]
mov ax, es:[bx]
mov [bp+var_C], ax
les bx, _VRAM_PLANE_E
add bx, [bp+var_14]
mov ax, es:[bx]
mov [bp+var_A], ax
mov ax, [bp+var_10]
or ax, [bp+var_E]
or ax, [bp+var_C]
or ax, [bp+var_A]
mov [bp+var_12], ax
xor si, si
jmp loc_EF44
; ---------------------------------------------------------------------------
loc_EEA9:
push 0
call _graph_accesspage_func
push [bp+var_12]
push ss
lea ax, [bp+var_1C]
push ax
call sub_ED32
push 0
call sub_D13A
les bx, _VRAM_PLANE_B
add bx, di
mov eax, [bp+var_1C]
mov es:[bx+si], eax
call sub_D1CE
push [bp+var_10]
push ss
lea ax, [bp+var_18]
push ax
call sub_ED32
les bx, _VRAM_PLANE_B
add bx, di
mov eax, [bp+var_18]
or es:[bx+si], eax
push [bp+var_E]
push ss
lea ax, [bp+var_18]
push ax
call sub_ED32
les bx, _VRAM_PLANE_R
add bx, di
mov eax, [bp+var_18]
or es:[bx+si], eax
push [bp+var_C]
push ss
lea ax, [bp+var_18]
push ax
call sub_ED32
les bx, _VRAM_PLANE_G
add bx, di
mov eax, [bp+var_18]
or es:[bx+si], eax
push [bp+var_A]
push ss
lea ax, [bp+var_18]
push ax
call sub_ED32
add sp, 22h
les bx, _VRAM_PLANE_E
add bx, di
mov eax, [bp+var_18]
or es:[bx+si], eax
add si, 50h ; 'P'
loc_EF44:
cmp si, 0A0h
jl loc_EEA9
add [bp+var_14], 2
add di, 4
inc [bp+var_6]
loc_EF56:
mov ax, [bp+arg_8]
mov bx, 10h
cwd
idiv bx
cmp ax, [bp+var_6]
jg loc_EE59
add [bp+var_4], 0A0h
add [bp+var_2], 50h ; 'P'
inc [bp+var_8]
loc_EF72:
mov ax, [bp+var_8]
cmp ax, [bp+arg_A]
jl loc_EE48
pop di
pop si
leave
retf
sub_EE15 endp
extern _graph_slow_2xscale_region_1_to_0:proc
fuuin_12_TEXT ends
; ===========================================================================