use16 align 1 include "..\include\mem.h" include "..\include\graphic.h" include "..\include\3d.h" org 0h header exe 1,exports,imports,0,0 exporting declare draw3d_point declare draw3d_line declare draw3d_hidden declare draw3d_hidden_fill declare load3ds declare translate declare translatex declare translatey declare translatez declare scale declare rescale declare copy declare fill declare identity declare rotationx declare rotationy declare rotationz declare rotation declare project declare transform declare multiply ende importing use GRAPHIC.LIB,line use GRAPHIC.LIB,polyfill use VIDEO,showpixel endi ;affiche liste vertex %0 proc draw3d_point uses cx si, vertex3d:word,vertex2d:word,camera:word,color:word mov si,[vertex2d] stdcall project,si,[vertex3d],[camera] mov cx,[si] inc si inc si .draw: virtual at si .vertex2d vertex2d end virtual invoke showpixel,[.vertex2d.px],[.vertex2d.py],[color] add si,4 dec cx jnz .draw ret endp ;affiche liste vertex %0 proc draw3d_line uses ax bx cx dx si di, type:word,faces:word,vertex3d:word,vertex2d:word,camera:word,color:word mov di,[faces] mov si,[vertex2d] stdcall project,si,[vertex3d],[camera] mov cx,[di] inc si inc si inc di inc di .draw: mov ax,[type] dec al mov dx,di .drawset: mov bx,[di] shl bx,2 push [color] virtual at bx+si .vertex2d vertex2d end virtual push [.vertex2d.py] push [.vertex2d.px] add di,2 mov bx,[di] shl bx,2 push [.vertex2d.py] push [.vertex2d.px] invoke line dec al jnz .drawset push di mov di,dx mov bx,[di] pop di shl bx,2 push [color] push [.vertex2d.py] push [.vertex2d.px] mov bx,[di] shl bx,2 push word [bx+si+2] push word [bx+si] invoke line add di,2 dec cx jnz .draw ret endp ;affiche liste vertex %0 proc draw3d_hidden uses ax bx cx dx si di, type:word,faces:word,vertex3d:word,vertex2d:word,camera:word,color:word local a1:WORD,a2:WORD,b1:WORD mov di,[faces] mov si,[vertex2d] stdcall project,si,[vertex3d],[camera] mov cx,[di] inc si inc si inc di inc di .draw: push cx mov si,[vertex2d] inc si inc si mov bx,[di+2] shl bx,2 virtual at bx+si .vertex2d vertex2d end virtual mov cx,[.vertex2d.px] mov dx,[.vertex2d.py] mov bx,[di] shl bx,2 mov ax,[.vertex2d.px] sub ax,cx mov [a1],ax mov ax,[.vertex2d.py] sub ax,dx mov [b1],ax mov bx,[di+4] shl bx,2 mov ax,[.vertex2d.px] sub ax,cx mov [a2],ax mov ax,[.vertex2d.py] sub ax,dx xor edx,edx imul [a1] and eax,0FFFFh rol eax,16 add eax,edx rol eax,16 mov ecx,eax mov ax,[b1] imul [a2] and eax,0FFFFh rol eax,16 add eax,edx rol eax,16 sub ecx,eax pop cx jge .nohidden mov ax,[type] shl ax,1 add di,ax dec cx jnz .draw jmp .endof .nohidden: mov ax,[type] dec al mov dx,di .drawset: mov bx,[di] shl bx,2 push [color] push [.vertex2d.py] push [.vertex2d.px] add di,2 mov bx,[di] shl bx,2 push [.vertex2d.py] push [.vertex2d.px] invoke line dec al jnz .drawset push di mov di,dx mov bx,[di] pop di shl bx,2 push [color] push [.vertex2d.py] push [.vertex2d.px] mov bx,[di] shl bx,2 push word [bx+si+2] push word [bx+si] invoke line add di,2 dec cx jnz .draw .endof: ret endp ;creer table pour face caché %0 proc draw3d_hidden_fill uses eax bx ecx edx si di, type:word,faces:word,vertex3d:word,vertex2d:word,camera:word,color:word local a1:WORD,a2:WORD,b1:WORD mov di,[faces] stdcall project,[vertex2d],[vertex3d],[camera] mov cx,[di] inc di inc di .calculvect: push cx mov si,[vertex2d] inc si inc si mov bx,[di+2] shl bx,2 virtual at bx+si .vertex2d vertex2d end virtual mov cx,[.vertex2d.px] mov dx,[.vertex2d.py] mov bx,[di] shl bx,2 mov ax,[.vertex2d.px] sub ax,cx mov [a1],ax mov ax,[.vertex2d.py] sub ax,dx mov [b1],ax mov bx,[di+4] shl bx,2 mov ax,[.vertex2d.px] sub ax,cx mov [a2],ax mov ax,[.vertex2d.py] sub ax,dx xor edx,edx imul [a1] rol eax,16 add eax,edx rol eax,16 mov ecx,eax mov ax,[b1] imul [a2] rol eax,16 add eax,edx rol eax,16 sub ecx,eax pop cx jl .hidden mov ax,[type] mov si,[vertex3d] inc si inc si fld [cs:zero] .calcz: mov bx,[di] mov dx,bx shl dx,2 shl bx,3 add bx,dx virtual at bx+si .vertex3d vertex3d end virtual fadd [.vertex3d.tz] add di,2 dec al jnz .calcz .hidden: mov ax,[type] shl ax,1 add di,ax .enofvalue: dec cx jnz .calculvect ret endp ;charge un fichier 3ds logé en %0 renvoie error ;sauvegarde en : ;- %1 le nom de l'objet ;- %2 les vertex 3D de l'objet ;- %3 la matrice de transformation de l'objet ;- %4 les faces de l'objet ;- %5 le type de face de l'objet ;1 non 3ds ;2 non 3 et > proc load3ds uses eax bx cx si di ds es fs, seg:word,addt:word,objectname:word,vertex:word,matrix:word,face:word push ds pop fs mov si,[addt] mov ds,[seg] cmp word [si],main jne .error1 cmp word [si+28],3 jb .error2 .reading: mov ax,[si] mov bx,[si+2] cmp ax,main je .enter cmp ax,edit je .enter cmp ax,mesh je .enter cmp ax,object je .readobject cmp ax,vertex je .readvertex cmp ax,locale je .readmatrix cmp ax,face je .readfaces .next: add si,bx jmp .reading .enter: add si,6 jmp .reading .readobject: add si,6 mov di,si mov al,0 mov es,[seg] mov cx,12 cld repne scasb mov cx,di sub cx,si mov di,[objectname] push fs pop es cld rep movsb jmp .reading .readvertex: add si,6 mov ax,[si] mov di,[vertex] mov [fs:di],ax inc si inc si mov cx,ax shl ax,1 add cx,ax add di,2 cld push fs pop es cld rep movsd jmp .reading .readmatrix: add si,6 mov di,[matrix] mov eax,[si] virtual at di .mat mat end virtual mov [fs:.mat.p1],eax mov eax,[si+4*1] mov [fs:.mat.p5],eax mov eax,[si+4*2] mov [fs:.mat.p9],eax mov eax,[si+4*3] mov [fs:.mat.p2],eax mov eax,[si+4*4] mov [fs:.mat.p6],eax mov eax,[si+4*5] mov [fs:.mat.p10],eax mov eax,[si+4*6] mov [fs:.mat.p3],eax mov eax,[si+4*7] mov [fs:.mat.p7],eax mov eax,[si+4*8] mov [fs:.mat.p11],eax mov eax,[si+4*9] mov [fs:.mat.p4],eax mov eax,[si+4*10] mov [fs:.mat.p8],eax mov eax,[si+4*11] mov [fs:.mat.p12],eax mov eax,[cs:un] mov [fs:.mat.p16],eax mov eax,0 mov [fs:.mat.p13],eax mov [fs:.mat.p14],eax mov [fs:.mat.p15],eax add si,12*4 jmp .reading .readfaces: add si,6 mov ax,[si] mov di,[face] mov [fs:di],ax inc si inc si add di,2 push fs pop es cld .readall: mov cx,3 rep movsw inc si inc si dec ax jnz .readall ;;jmp .reading .error1: .error2: ret endp un dd 1.0 zero dd 0.0 ;initialise une matrice de translation pour une translation TX,TY,TZ (%1,%2,%3) dans MATRICE %0 ; mat[ 0] = 1.0 ; mat[ 5] = 1.0 ; mat[10] = 1.0 ; mat[15] = 1.0 ; mat[ 3] = vecteur->x ; mat[ 7] = vecteur->y ; mat[11] = vecteur->z proc translate uses eax ecx di es, mat:word,x:dword,y:dword,z:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[cs:un] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov [di+15*4],eax mov eax,[x] mov [di+3*4],eax mov eax,[y] mov [di+7*4],eax mov eax,[z] mov [di+11*4],eax ret endp ;initialise une matrice de translation pour une translation TX %1 dans MATRICE %0 ; mat[ 0] = 1.0 ; mat[ 5] = 1.0 ; mat[10] = 1.0 ; mat[15] = 1.0 ; mat[ 3] = value ; mat[ 7] = 1 ; mat[11] = 1 proc translatex uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[cs:un] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov [di+15*4],eax mov [di+7*4],eax mov [di+11*4],eax mov eax,[value] mov [di+3*4],eax ret endp ;initialise une matrice de translation pour une translation TY %1 dans MATRICE %0 ; mat[ 0] = 1.0 ; mat[ 5] = 1.0 ; mat[10] = 1.0 ; mat[15] = 1.0 ; mat[ 3] = 1 ; mat[ 7] = value ; mat[11] = 1 proc translatey uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[cs:un] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov [di+15*4],eax mov [di+3*4],eax mov [di+11*4],eax mov eax,[value] mov [di+7*4],eax ret endp ;initialise une matrice de translation pour une translation TZ %1 dans MATRICE %0 ; mat[ 0] = 1.0 ; mat[ 5] = 1.0 ; mat[10] = 1.0 ; mat[15] = 1.0 ; mat[ 3] = 1 ; mat[ 7] = 1 ; mat[11] = value proc translatez uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[cs:un] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov [di+15*4],eax mov [di+3*4],eax mov [di+7*4],eax mov eax,[value] mov [di+11*4],eax ret endp ;initialise une matrice d'echelle %0 de facteur (x,y,z) %1-%3 ; mat[ 0] = factorx ; mat[ 5] = factory ; mat[10] = factorz ; mat[15] = 1.0 ; reste a 0 proc scale uses eax ecx di es, mat:word,x:dword,y:dword,z:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[x] mov [di+0*4],eax mov eax,[y] mov [di+5*4],eax mov eax,[z] mov [di+10*4],eax mov eax,[cs:un] mov [di+15*4],eax ret endp ;initialise une matrice d'echelle %0 de facteur value %1 ; mat[ 0] = factor ; mat[ 5] = factor ; mat[10] = factor ; mat[15] = 1.0 ; reste a 0 proc rescale uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[value] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov eax,[cs:un] mov [di+15*4],eax ret endp ;copy une matrice %0 en %1 ;mat2=mat1 proc copy uses ecx si di es, mat1:word,mat2:word mov si,[mat1] mov di,[mat2] mov ecx,16 push ds pop es cld rep stosd ret endp ;initialise une matrice %0 avec la valeur %1 ;mat[i]=value proc fill uses eax ecx di es, mat:word,value:dword mov di,[mat] mov eax,[value] mov ecx,16 push ds pop es cld rep stosd ret endp ;initialise une matrice d'identité %0 ; mat[0] = 1.0 ; mat[5] = 1.0 ; mat[10] = 1.0 ; mat[15] = 1.0 ; reste a 0 proc identity uses eax ecx di es, mat:word mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] mov eax,[cs:un] mov [di+0*4],eax mov [di+5*4],eax mov [di+10*4],eax mov [di+15*4],eax ret endp ;initialise une matrice de rotation %0 autour de X de %1 degrees ; mat[ 5] = cos(angle) ; mat[ 6] = -sin(angle) ; mat[ 9] = sin(angle) ; mat[10] = cos(angle) ; mat[ 0] = 1.0 ; mat[15] = 1.0 ; reste a 0 proc rotationx uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] fld [value] fsincos ;mat[ 5] = cos(angle); fst dword [di+5*4] ;mat[ 10] = cos(angle); fstp dword [di+10*4] ;mat[ 9] = sin(angle); fst dword [di+9*4] ;mat[ 6] = -sin(angle); fchs fstp dword [di+6*4] ;mat[ 0] = 1.0 ;mat[15] = 1.0 mov eax,[cs:un] mov [di+0*4],eax mov [di+15*4],eax ret endp ;initialise une matrice de rotation BX autour de Y de %0 degrees ; mat[ 0] = cos(angle) ; mat[ 8] = -sin(angle) ; mat[ 2] = sin(angle) ; mat[10] = cos(angle) ; mat[ 5] = 1.0 ; mat[15] = 1.0 ; reste a 0 proc rotationy uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] fld [value] fsincos ;mat[ 0] = cos(angle) fst dword [di+0*4] ;mat[ 10] = cos(angle) fstp dword [di+10*4] ;mat[ 2] = sin(angle) fst dword [di+2*4] fchs ;mat[ 8] = -sin(angle) fstp dword [ di+8*4] ;mat[ 5] = 1.0 ;mat[15] = 1.0 mov eax,[cs:un] mov [di+5*4],eax mov [di+15*4],eax ret endp ;initialise une matrice de rotation %0 autour de Z de %1 degrees ; mat[ 0] = cos(angle) ; mat[ 1] = -sin(angle) ; mat[ 4] = sin(angle) ; mat[ 5] = cos(angle) ; mat[10] = 1.0 ; mat[15] = 1.0 ; reste a 0 proc rotationz uses eax ecx di es, mat:word,value:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] fld [value] fsincos ;mat[ 0] = cos(angle) fst dword [ di+0*4] ;mat[ 5] = cos(angle) fstp dword [ di+5*4] ;mat[ 4] = sin(angle) fst dword [ di+4*4] fchs ;mat[ 1] = -sin(angle) fstp dword [ di+1*4] ;mat[10] = 1.0 ;mat[15] = 1.0 mov eax,[cs:un] mov [di+10*4],eax mov [di+15*4],eax ret endp ;initialise une matrice de rotation %0 autour de X,Y,Z de %0-%3 degrees ; mat[ 0] = cos(angleY)*cos(angleZ) ; mat[ 1] = cos(angleY)*sin(angleZ) ; mat[ 2] = -sin(angleY) ; mat[ 4] = sin(angleX)*sin(angleY)*cos(angleZ)+cos(angleX)*-sin(angleZ) ; mat[ 5] = sin(angleX)*sin(angleY)*sin(angleZ)+cos(angleX)*cos(angleZ) ; mat[ 6] = sin(angleX)*cos(angleY) ; mat[ 8] = cos(angleX)*sin(angleY)*cos(angleZ)+sin(angleX)*sin(angleZ) ; mat[ 9] = cos(angleX)*sin(angleY)*sin(angleZ)-sin(angleX)*cos(angleZ) ; mat[ 10] = cos(angleX)*cos(angleY) ; mat[3] = 0.0 ; mat[7] = 0.0 ; mat[11] = 0.0 ; mat[12] = 0.0 ; mat[13] = 0.0 ; mat[14] = 0.0 ; mat[15] = 1.0 ; reste a 0 proc rotation uses eax ecx di es, mat:word,anglex:dword,angley:dword,anglez:dword mov di,[mat] mov ecx,16 mov eax,0 push ds pop es cld rep stosd mov di,[mat] ;st1 sin(angleZ) ;st2 cos(angleZ) ;st3 sin(angleY) ;st4 cos(angleY) ;st5 sin(angleX) ;st6 cos(angleX) fld [anglex] fsincos fld [angley] fsincos fld [anglez] fsincos ;Cos(angleY)*Cos(angleZ) fld st3 fmul st0,st2 ;mat[0] fstp dword [ di+0*4] ;Cos(angleY)*Sin(angleZ) fld st3 fmul st0,st1 ;mat[1] fstp dword [ di+1*4] ;-Sin(angley) fld st2 fchs ;mat[2] fstp dword [ di+2*4] mov eax,[cs:un] mov [di+15*4],eax ret endp factor dd 128.0 ;Transforme la liste de vertex 3D pointé par %0 en vertex 2D dans %1 en utilisant %2 comme origine pour %3 valeurs ; vertex2d[i].px=int((vertex3d[i].tx*factor)/(vertex3d[i].tz+origin.Z)+origin.X) ; vertex2d[i].py=int((vertex3d[i].ty*factor)/(vertex3d[i].tz+origin.Z)+origin.Y) proc project uses bx cx si di, vertex2d:word,vertex3d:word,origin:word mov si,[vertex3d] mov bx,[vertex2d] mov di,[origin] mov cx,[si] mov [bx],cx inc si inc si inc bx inc bx .boucle: virtual at 0 .vertex3d vertex3d end virtual virtual at 0 .vertex2d vertex2d end virtual virtual at si .vertex3dsrc vertex3d end virtual virtual at di .vertex3ddst vertex3d end virtual virtual at bx .vertex2dsrc vertex2d end virtual ;(vertex3d[i].z+origZ) fld [.vertex3dsrc.tz] fadd [.vertex3ddst.tz] ;(vertex3d[i].tx*factor) fld [.vertex3dsrc.tx] fmul [cs:factor] ;(vertex3d[i].tx*factor)/(vertex3d[i].tz+origZ) fdiv st,st1 ;(vertex3d[i].tx*factor)/(vertex3d[i].tz+origZ)+origX fadd [.vertex3ddst.tx] ;vertex2d[i].tx=int((vertex3d[i].tx*factor)/(vertex3d[i].tz+origZ)+origX) fistp [.vertex2dsrc.px] ;(vertex3d[i].z+origZ) ;(vertex3d[i].ty*factor) fld [.vertex3dsrc.ty] fmul [cs:factor] ;(vertex3d[i].ty*factor)/(vertex3d[i].tz+origZ) ;fdivrp st1 ;(vertex3d[i].ty*factor)/(vertex3d[i].tz+origZ)+origY fadd [.vertex3ddst.ty] ;vertex2d[i].ty=int((vertex3d[i].ty*factor)/(vertex3d[i].tz+origZ)+origY) fistp [.vertex2dsrc.py] add si,.vertex3d.sizeof add bx,.vertex2d.sizeof dec cx jnz .boucle ret endp ;transforme les points %0 avec la matrice %1 ; vertex3d_2 = mat * vertex3d ; w = mat[12]*vertex3d[i].tx + mat[13]*vertex3d[i].ty + mat[14]*vertex3d[i].tz + mat[15] ; vertex3d_2[i].tx = mat[ 0]*vertex3d[i].tx + mat[ 1]*vertex3d[i].ty + mat[ 2]*vertex3d[i].tz + mat[ 3] ; vertex3d_2[i].ty = mat[ 4]*vertex3d[i].tx + mat[ 5]*vertex3d[i].ty + mat[ 6]*vertex3d[i].tz + mat[ 7] ; vertex3d_2[i].tz = mat[ 8]*vertex3d[i].tx + mat[ 9]*vertex3d[i].ty + mat[10]*vertex3d[i].tz + mat[11] proc transform uses cx si di, vertex3d:word,mat:word mov si,[vertex3d] mov cx,[si] inc si inc si mov di,[mat] .boucle: ;Calcul du facteur echelle ;mat[12]*vertex3d[i].tx virtual at si .vertex3d vertex3d end virtual fld [.vertex3d.tx] fmul dword [ di+12*4] ;mat[13]*vertex3d[i].ty fld [.vertex3d.ty] fmul dword [ di+13*4] ;mat[12]*vertex3d[i].tx + mat[13]*vertex3d[i].ty ;faddp st1 ;mat[14]*vertex3d.tz fld [.vertex3d.tz] fmul dword [ di+14*4] ;mat[12]*vertex3d[i].tx + mat[13]*vertex3d[i].ty + mat[14]*vertex3d[i].tz ;faddp st1 ;mat[12]*vertex3d[i].tx + mat[13]*vertex3d[i].ty + mat[14]*vertex3d[i].tz + mat[15] fadd dword [ di+15*4] ;w=0.0 ?? ftst ;-> AX fstsw ax ;vertex3d.tx vertex3d.ty et vertex3d.tz fld [.vertex3d.tx] fmul dword [ di+0*4] fld [.vertex3d.ty] fmul dword [ di+1*4] ;faddp st1 fld [.vertex3d.tz] fmul dword [ di+2*4] ;faddp st1 fadd dword [ di+3*4] fld [.vertex3d.tx] fmul dword [ di+4*4] fld [.vertex3d.ty] fmul dword [ di+5*4] ;faddp st1 fld [.vertex3d.tz] fmul dword [ di+6*4] ;faddp st1 fadd dword [ di+7*4] fld [.vertex3d.tx] fmul dword [ di+8*4] fld [.vertex3d.ty] fmul dword [ di+9*4] ;faddp st1 fld [.vertex3d.tz] fmul dword [ di+10*4] ;faddp st1 fadd dword [ di+11*4] ;w=0.0 sahf jz .divby0 ;vertex3d.tx=vertex3d[i].tz/w fdiv st,st3 fstp [.vertex3d.tz] ;vertex3d.ty=vertex3d[i].ty/w fdiv st,st2 fstp [.vertex3d.ty] ;vertex3d.tz=vertex3d[i].tx/w fdivrp st1,st fstp [.vertex3d.tx] ;No div by 0 jmp .nodiv .divby0: fstp [.vertex3d.tz] fstp [.vertex3d.ty] fstp [.vertex3d.tx] ;ffree .nodiv: virtual at 0 .vertex3dori vertex3d end virtual add si,.vertex3dori.sizeof dec cx jnz .boucle ret endp ;Multiplie la matrice de transformation %0 par celle en %1 et met les resultat en %2 ; mat = p1 * p2 ; mat[ 0] = (p1[ 0]*p2[ 0])+(p1[ 1]*p2[ 4])+(p1[ 2]*p2[ 8])+(p1[ 3]*p2[12]); ; mat[ 4] = (p1[ 4]*p2[ 0])+(p1[ 5]*p2[ 4])+(p1[ 6]*p2[ 8])+(p1[ 7]*p2[12]); ; mat[ 8] = (p1[ 8]*p2[ 0])+(p1[ 9]*p2[ 4])+(p1[10]*p2[ 8])+(p1[11]*p2[12]); ; mat[12] = (p1[12]*p2[ 0])+(p1[13]*p2[ 4])+(p1[14]*p2[ 8])+(p1[15]*p2[12]); ; mat[ 1] = (p1[ 0]*p2[ 1])+(p1[ 1]*p2[ 5])+(p1[ 2]*p2[ 9])+(p1[ 3]*p2[13]); ; mat[ 5] = (p1[ 4]*p2[ 1])+(p1[ 5]*p2[ 5])+(p1[ 6]*p2[ 9])+(p1[ 7]*p2[13]); ; mat[ 9] = (p1[ 8]*p2[ 1])+(p1[ 9]*p2[ 5])+(p1[10]*p2[ 9])+(p1[11]*p2[13]); ; mat[13] = (p1[12]*p2[ 1])+(p1[13]*p2[ 5])+(p1[14]*p2[ 9])+(p1[15]*p2[13]); ; mat[ 2] = (p1[ 0]*p2[ 2])+(p1[ 1]*p2[ 6])+(p1[ 2]*p2[10])+(p1[ 3]*p2[14]); ; mat[ 6] = (p1[ 4]*p2[ 2])+(p1[ 5]*p2[ 6])+(p1[ 6]*p2[10])+(p1[ 7]*p2[14]); ; mat[10] = (p1[ 8]*p2[ 2])+(p1[ 9]*p2[ 6])+(p1[10]*p2[10])+(p1[11]*p2[14]); ; mat[14] = (p1[12]*p2[ 2])+(p1[13]*p2[ 6])+(p1[14]*p2[10])+(p1[15]*p2[14]); ; mat[ 3] = (p1[ 0]*p2[ 3])+(p1[ 1]*p2[ 7])+(p1[ 2]*p2[11])+(p1[ 3]*p2[15]); ; mat[ 7] = (p1[ 4]*p2[ 3])+(p1[ 5]*p2[ 7])+(p1[ 6]*p2[11])+(p1[ 7]*p2[15]); ; mat[11] = (p1[ 8]*p2[ 3])+(p1[ 9]*p2[ 7])+(p1[10]*p2[11])+(p1[11]*p2[15]); ; mat[15] = (p1[12]*p2[ 3])+(p1[13]*p2[ 7])+(p1[14]*p2[11])+(p1[15]*p2[15]); proc multiply uses bx si di, p1:word,p2:word,mat:word mov si,[p1] mov di,[p2] mov bx,[mat] ;0,0 ;p1[ 0]*p2[ 0] fld dword [ si+0*4] fmul dword [ di+0*4] ;p1[ 1]*p2[ 4] fld dword [ si+1*4] fmul dword [ di+4*4] ;(p1[ 0]*p2[ 0])+(p1[ 1]*p2[ 4]) ;;faddp st1 ;p1[ 2]*p2[ 8] fld dword [ si+2*4] fmul dword [ di+8*4] ;(p1[ 0]*p2[ 0])+(p1[ 1]*p2[ 4])+(p1[ 2]*p2[ 8]) ;;faddp st1 ;p1[ 3]*p2[12] fld dword [ si+3*4] fmul dword [ di+12*4] ;(p1[ 0]*p2[ 0])+(p1[ 1]*p2[ 4])+(p1[ 2]*p2[ 8])+(p1[ 3]*p2[12]) ;;faddp st1 fstp dword [ bx+0*4] ;mat[ 0] ;1,0 fld dword [ si+0*4] fmul dword [ di+1*4] fld dword [ si+1*4] fmul dword [ di+5*4] ;faddp st1 fld dword [ si+2*4] fmul dword [ di+9*4] ;faddp st1 fld dword [ si+3*4] fmul dword [ di+13*4] ;faddp st1 fstp dword [ bx+1*4] ;2,0 fld dword [ si+0*4] fmul dword [ di+2*4] fld dword [ si+1*4] fmul dword [ di+6*4] ;faddp st1 fld dword [ si+2*4] fmul dword [ di+10*4] ;faddp st1 fld dword [ si+3*4] fmul dword [ di+14*4] ;faddp st1 fstp dword [ bx+2*4] ;3,0 fld dword [ si+0*4] fmul dword [ di+3*4] fld dword [ si+1*4] fmul dword [ di+7*4] ;faddp st1 fld dword [ si+2*4] fmul dword [ di+11*4] ;faddp st1 fld dword [ si+3*4] fmul dword [ di+15*4] ;faddp st1 fstp dword [ bx+3*4] ;0,1 fld dword [ si+4*4] fmul dword [ di+0*4] fld dword [ si+5*4] fmul dword [ di+4*4] ;faddp st1 fld dword [ si+6*4] fmul dword [ di+8*4] ;faddp st1 fld dword [ si+7*4] fmul dword [ di+12*4] ;faddp st1 fstp dword [ bx+4*4] ;1,1 fld dword [ si+4*4] fmul dword [ di+1*4] fld dword [ si+5*4] fmul dword [ di+5*4] ;faddp st1 fld dword [ si+6*4] fmul dword [ di+9*4] ;faddp st1 fld dword [ si+7*4] fmul dword [ di+13*4] ;faddp st1 fstp dword [ bx+5*4] ;2,1 fld dword [ si+4*4] fmul dword [ di+2*4] fld dword [ si+5*4] fmul dword [ di+6*4] ;faddp st1 fld dword [ si+6*4] fmul dword [ di+10*4] ;faddp st1 fld dword [ si+7*4] fmul dword [ di+14*4] ;faddp st1 fstp dword [ bx+6*4] ;3,1 fld dword [ si+4*4] fmul dword [ di+3*4] fld dword [ si+5*4] fmul dword [ di+7*4] ;faddp st1 fld dword [ si+6*4] fmul dword [ di+11*4] ;faddp st1 fld dword [ si+7*4] fmul dword [ di+15*4] ;faddp st1 fstp dword [ bx+7*4] ;0,2 fld dword [ si+8*4] fmul dword [ di+0*4] fld dword [ si+9*4] fmul dword [ di+4*4] ;faddp st1 fld dword [ si+10*4] fmul dword [ di+8*4] ;faddp st1 fld dword [ si+11*4] fmul dword [ di+12*4] ;faddp st1 fstp dword [ bx+8*4] ;1,2 fld dword [ si+8*4] fmul dword [ di+1*4] fld dword [ si+9*4] fmul dword [ di+5*4] ;faddp st1 fld dword [ si+10*4] fmul dword [ di+9*4] ;faddp st1 fld dword [ si+11*4] fmul dword [ di+13*4] ;faddp st1 fstp dword [ bx+9*4] ;2,2 fld dword [ si+8*4] fmul dword [ di+2*4] fld dword [ si+9*4] fmul dword [ di+6*4] ;faddp st1 fld dword [ si+10*4] fmul dword [ di+10*4] ;faddp st1 fld dword [ si+11*4] fmul dword [ di+14*4] ;faddp st1 fstp dword [ bx+10*4] ;3,2 fld dword [ si+8*4] fmul dword [ di+3*4] fld dword [ si+9*4] fmul dword [ di+7*4] ;faddp st1 fld dword [ si+10*4] fmul dword [ di+11*4] ;faddp st1 fld dword [ si+11*4] fmul dword [ di+15*4] ;faddp st1 fstp dword [ bx+11*4] ;0,3 fld dword [ si+12*4] fmul dword [ di+0*4] fld dword [ si+13*4] fmul dword [ di+4*4] ;faddp st1 fld dword [ si+14*4] fmul dword [ di+8*4] ;faddp st1 fld dword [ si+15*4] fmul dword [ di+12*4] ;faddp st1 fstp dword [ bx+12*4] ;1,3 fld dword [ si+12*4] fmul dword [ di+1*4] fld dword [ si+13*4] fmul dword [ di+5*4] ;faddp st1 fld dword [ si+14*4] fmul dword [ di+9*4] ;faddp st1 fld dword [ si+15*4] fmul dword [ di+13*4] ;faddp st1 fstp dword [ bx+13*4] ;2,3 fld dword [ si+12*4] fmul dword [ di+2*4] fld dword [ si+13*4] fmul dword [ di+6*4] ;faddp st1 fld dword [ si+14*4] fmul dword [ di+10*4] ;faddp st1 fld dword [ si+15*4] fmul dword [ di+14*4] ;faddp st1 fstp dword [ bx+14*4] ;3,3 fld dword [ si+12*4] fmul dword [ di+3*4] fld dword [ si+13*4] fmul dword [ di+7*4] ;faddp st1 fld dword [ si+14*4] fmul dword [ di+11*4] ;faddp st1 fld dword [ si+15*4] fmul dword [ di+15*4] ;faddp st1 fstp dword [ bx+15*4] ret endp