AstralTransRocketries/ClassiCube
1
0
Fork 0
mirror of https://github.com/ClassiCube/ClassiCube.git synced 2025-09-09 07:18:34 -04:00
Code Issues Packages Projects Releases Wiki Activity

Dreamcast: assembly optimised version nearly works

Browse Source
This commit is contained in:
UnknownShadow200 2024-07-07 21:42:53 +10:00
parent c891f09b7e
commit 333f0e5c9b
2 changed files with 145 additions and 172 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
misc/dreamcast/VertexClip.S
View File

@ -63,11 +63,11 @@ _ClipLine:
mov IN1, TM1 ! MT, tmp = &v1
fldi0 fr4 ! LS, fr4 = 0
add #12, TM1 ! EX, tmp = &v1->z
fmov.s @r1, fr2 ! LS, fr2 = v1->z
fmov.s @TM1, fr2 ! LS, fr2 = v1->z
mov IN2, TM1 ! MT, tmp = &v2
fldi0 fr5 ! LS, fr5 = 0
add #12, TM1 ! EX, tmp = &v2->z
fmov.s @r1,fr11 ! LS, fr11 = v2->z
fmov.s @TM1,fr11 ! LS, fr11 = v2->z
fsub fr2,fr11 ! FE, fr11 = v2->z - v1->z
fldi0 fr8 ! LS, fr8 = 0
fmul fr11,fr11 ! FE, fr11 = (v2->z - v1->z) * (v2->z * v1->z)

313
misc/dreamcast/VertexDraw.S
View File

@ -6,35 +6,47 @@
! r13 = cur vertex
! r14 = next vertex (prefetch)
!fr8 = VIEWPORT_HWIDTH
!fr9 = VIEWPORT_HHEIGHT
!fr10 = VIEWPORT_X_PLUS_HWIDTH
!fr11 = VIEWPORT_Y_PLUS_HHEIGHT
!fr12 = VIEWPORT_HWIDTH
!fr13 = VIEWPORT_HHEIGHT
!fr14 = VIEWPORT_X_PLUS_HWIDTH
!fr15 = VIEWPORT_Y_PLUS_HHEIGHT
#define F_HW fr12
#define F_HH fr13
#define F_XP fr14
#define F_YP fr15
#define R_VTX r10
#define R_EOL r11
#define REG_CMD_VTX r10
#define REG_CMD_EOL r11
#define REG_CLIPFUNC r12
.align 4
! Pushes a vertex to the store queue
! CLOBBERS: r3
! CLOBBERS: r2
! INPUTS: R (vertex), r8 (SQ global)
! OUTPUTS: r8 altered
.macro PushVertex R
! memcpy(r8, \R, 32)
mov.l @(0,\R), r3
mov.l r3, @(0,r8)
mov.l @(4,\R), r3
mov.l r3, @(4,r8)
mov.l @(8,\R), r3
mov.l r3, @(8,r8)
mov.l @(12,\R),r3
mov.l r3,@(12,r8)
mov.l @(16,\R),r3
mov.l r3,@(16,r8)
mov.l @(20,\R),r3
mov.l r3,@(20,r8)
mov.l @(24,\R),r3
mov.l r3,@(24,r8)
mov.l @(28,\R),r3
mov.l r3,@(28,r8)
mov.l @(0,\R), r2
mov.l r2, @(0,r8)
mov.l @(4,\R), r2
mov.l r2, @(4,r8)
mov.l @(8,\R), r2
mov.l r2, @(8,r8)
mov.l @(12,\R),r2
mov.l r2,@(12,r8)
mov.l @(16,\R),r2
mov.l r2,@(16,r8)
mov.l @(20,\R),r2
mov.l r2,@(20,r8)
mov.l @(24,\R),r2
mov.l r2,@(24,r8)
mov.l @(28,\R),r2
mov.l r2,@(28,r8)
pref @r8 ! LS, Trigger SQ
add #32,r8 ! EX, SQ += 32
.endm
@ -55,16 +67,16 @@
! TRANSFORM X
fmov.s @\R,fr4 ! LS, fr4 = vertex->x
fmov fr10,fr5 ! LS, fr5 = VIEWPORT_X_PLUS_HWIDTH
fmul fr8,fr4 ! FE, fr4 = VIEWPORT_HWIDTH * vertex->x
fmov F_XP,fr5 ! LS, fr5 = VIEWPORT_X_PLUS_HWIDTH
fmul F_HW,fr4 ! FE, fr4 = VIEWPORT_HWIDTH * vertex->x
fmac fr0,fr4,fr5 ! FE, fr5 = fr0 * fr4 + fr5 -- (X * F * hwidth) + x_plus_hwidth
fmov.s fr5,@\R ! LS, vertex->x = fr5
add #4, \R ! EX, \R = &vertex->y
! TRANSFORM Y
fmov.s @\R,fr4 ! LS, fr4 = vertex->y
fmov fr11,fr5 ! LS, fr5 = VIEWPORT_Y_PLUS_HHEIGHT
fmul fr9,fr4 ! FE, fr4 = VIEWPORT_HHEIGHT * vertex->y
fmov F_YP,fr5 ! LS, fr5 = VIEWPORT_Y_PLUS_HHEIGHT
fmul F_HH,fr4 ! FE, fr4 = VIEWPORT_HHEIGHT * vertex->y
fmac fr0,fr4,fr5 ! FE, fr5 = fr0 * fr4 + fr5 -- (Y * F * hheight) + y_plus_hheight
fmov.s fr5,@\R ! LS, vertex->y = fr5
add #4, \R ! EX, \R = &vertex->z
@ -75,54 +87,52 @@
.endm
! Transforms then pushes a vertex to the store queue
! CLOBBERS: r3, fr0, fr4, fr5
! CLOBBERS: r2, fr0, fr4, fr5
! INPUTS: R (vertex), r8 (SQ global)
! OUTPUTS: r8 altered
! OUTPUTS: R, r8 altered
.macro TransformVertex R
! INVERSE W CALCULATION
add #28, \R ! EX, \R = &vertex->w
fmov.s @\R,fr0 ! LS, fr0 = vertex->w
add #28, \R ! EX, SRC += 28
fmov.s @\R,fr0 ! LS, fr0 = v->w
fmul fr0,fr0 ! FE, fr0 = fr0 * fr0
add #-24, \R ! EX, \R = &vertex->x
add #-28, \R ! EX, SRC -= 28
mov.l @\R+, r2 ! LS, tmp = SRC->flags, SRC += 4
mov.l r2,@r8 ! LS, DST->flags = tmp
fsrra fr0 ! FE, fr0 = 1 / sqrt(fr0) -> 1 / vertex->w
add #4, r8 ! EX, DST += 4
! COPY U,V
mov.l @(12,\R),r2 ! LS, tmp = SRC->u
mov.l r2,@(12,r8) ! LS, DST->u = tmp
mov.l @(16,\R),r2 ! LS, tmp = SRC->v
mov.l r2,@(16,r8) ! LS, DST->v = tmp
! TRANSFORM X
fmov.s @\R,fr4 ! LS, fr4 = vertex->x
fmov fr10,fr5 ! LS, fr5 = VIEWPORT_X_PLUS_HWIDTH
fmul fr8,fr4 ! FE, fr4 = VIEWPORT_HWIDTH * vertex->x
fmov.s @\R,fr4 ! LS, fr4 = SRC->x
fmov F_XP,fr5 ! LS, fr5 = VIEWPORT_X_PLUS_HWIDTH
fmul F_HW,fr4 ! FE, fr4 = VIEWPORT_HWIDTH * SRC->x
mov.l @(20,\R),r2 ! LS, tmp = SRC->bgra
mov.l r2,@(20,r8) ! LS, SRC->bgra = tmp
fmac fr0,fr4,fr5 ! FE, fr5 = fr0 * fr4 + fr5 -- (X * F * hwidth) + x_plus_hwidth
fmov.s fr5,@\R ! LS, vertex->x = fr5
add #4, \R ! EX, \R = &vertex->y
add #4, \R ! EX, SRC += 4
fmov.s fr5,@r8 ! LS, DST->x = fr5
! TRANSFORM Y
fmov.s @\R,fr4 ! LS, fr4 = vertex->y
fmov fr11,fr5 ! LS, fr5 = VIEWPORT_Y_PLUS_HHEIGHT
fmul fr9,fr4 ! FE, fr4 = VIEWPORT_HHEIGHT * vertex->y
fmov F_YP,fr5 ! LS, fr5 = VIEWPORT_Y_PLUS_HHEIGHT
fmul F_HH,fr4 ! FE, fr4 = VIEWPORT_HHEIGHT * vertex->y
fmac fr0,fr4,fr5 ! FE, fr5 = fr0 * fr4 + fr5 -- (Y * F * hheight) + y_plus_hheight
fmov.s fr5,@\R ! LS, vertex->y = fr5
add #4, \R ! EX, \R = &vertex->z
add #4, r8 ! EX, DST += 4
add #-8, \R ! EX, src -= 8 (back to start of vertex)
fmov.s fr5,@r8 ! LS, DST->y = fr5
! ASSIGN Z
fmov.s fr0,@\R ! LS, vertex->z = fr0
add #-12, \R ! EX, \R -= 12 (back to start of vertex)
add #4, r8 ! EX, DST += 4
fmov.s fr0,@r8 ! LS, DST->z = fr0
add #-12,r8 ! EX, DST -= 12 (back to start of vertex)
! memcpy(r8, \R, 28)
mov.l @(0,\R), r3
mov.l r3, @(0,r8)
mov.l @(4,\R), r3
mov.l r3, @(4,r8)
mov.l @(8,\R), r3
mov.l r3, @(8,r8)
mov.l @(12,\R),r3
mov.l r3,@(12,r8)
mov.l @(16,\R),r3
mov.l r3,@(16,r8)
mov.l @(20,\R),r3
mov.l r3,@(20,r8)
mov.l @(24,\R),r3
mov.l r3,@(24,r8)
pref @r8 ! LS, Trigger SQ
add #32,r8 ! EX, SQ += 32
pref @r8 ! LS, Trigger SQ
add #32,r8 ! EX, SQ += 32
.endm
@ -134,10 +144,6 @@
#define REG_V2 r6
#define REG_V3 r7
#define REG_CMD_VTX r10
#define REG_CMD_EOL r11
#define REG_CLIPFUNC r12
! r3 also matches out parameter for ClipLine
#define REG_TMP r3
#define TMP_SET_A \
@ -148,28 +154,25 @@
_Case_0_0_0_1:
rts; nop
! v0
! / |
! / |
! .....A....B...
! / |
! v3--v2---v1
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TMP_SET_B
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TransformVertex REG_V0
TMP_SET_B
@ -178,31 +181,28 @@ _Case_0_0_0_1:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_0_0_1_0:
rts; nop
! v1
! / |
! / |
! ....A.....B...
! / |
! v0--v3---v2
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TMP_SET_A
TransformVertex REG_TMP
@ -211,31 +211,28 @@ _Case_0_0_1_0:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_0_1_0_0:
rts; nop
! v2
! / |
! / |
! ....A.....B...
! / |
! v1--v0---v3
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TMP_SET_A
TransformVertex REG_TMP
@ -244,31 +241,28 @@ _Case_0_1_0_0:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_0_0_0:
rts; nop
! v3
! / |
! / |
! ....A.....B...
! / |
! v2--v1---v0
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
TransformVertex REG_TMP
@ -277,31 +271,28 @@ _Case_1_0_0_0:
TransformVertex REG_V3
rts
lds.l @r15+,pr
lds r13,pr
_Case_0_0_1_1:
rts; nop
! v0-----------v1
! \ |
! ....B..........A...
! \ |
! v3-----v2
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TransformVertex REG_V1
TMP_SET_A
@ -311,30 +302,27 @@ _Case_0_0_1_1:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_0_0_1:
rts; nop
! v3-----------v0
! \ |
! ....B..........A...
! \ |
! v2-----v1
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_A
TransformVertex REG_TMP
@ -344,30 +332,27 @@ _Case_1_0_0_1:
TransformVertex REG_V3
rts
lds.l @r15+,pr
lds r13,pr
_Case_0_1_1_0:
rts; nop
! v1-----------v2
! \ |
! ....B..........A...
! \ |
! v0-----v3
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TMP_SET_B
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TransformVertex REG_V1
TransformVertex REG_V2
@ -377,30 +362,27 @@ _Case_0_1_1_0:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_1_0_0:
rts; nop
! v2-----------v3
! \ |
! ....B..........A...
! \ |
! v1-----v0
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
TransformVertex REG_TMP
@ -410,10 +392,9 @@ _Case_1_1_0_0:
TransformVertex REG_V3
rts
lds.l @r15+,pr
lds r13,pr
_Case_0_1_1_1:
rts; nop
! --v1--
! v0-- --v2
! \ |
@ -421,21 +402,19 @@ _Case_0_1_1_1:
! \ |
! v3
! v1,v2,v0 v2,v0,A v0,A,B
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_EOL, @REG_TMP
TransformVertex REG_V1
TransformVertex REG_V2
@ -446,10 +425,9 @@ _Case_0_1_1_1:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_0_1_1:
rts; nop
! --v0--
! v3-- --v1
! \ |
@ -457,20 +435,19 @@ _Case_1_0_1_1:
! \ |
! v2
! v0,v1,v3 v1,v3,A v3,A,B
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V2, REG_CLIP1
mov REG_V3, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
mov.l REG_CMD_EOL, @REG_TMP
mov.l REG_CMD_VTX, @REG_V3
TransformVertex REG_V0
@ -482,10 +459,9 @@ _Case_1_0_1_1:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_1_0_1:
rts; nop
! --v3--
! v2-- --v0
! \ |
@ -493,20 +469,19 @@ _Case_1_1_0_1:
! \ |
! v1
! v3,v0,v2 v0,v2,A v2,A,B
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V1, REG_CLIP1
mov REG_V2, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
mov.l REG_CMD_EOL, @REG_TMP
mov.l REG_CMD_VTX, @REG_V3
TransformVertex REG_V3
@ -518,10 +493,9 @@ _Case_1_1_0_1:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_1_1_0:
rts; nop
! --v2--
! v1-- --v3
! \ |
@ -529,20 +503,19 @@ _Case_1_1_1_0:
! \ |
! v0
! v2,v3,v1 v3,v1,A v1,A,B
sts.l pr,@-r15
sts pr,r13
TMP_SET_A
mov REG_V3, REG_CLIP1
mov REG_V0, REG_CLIP2
mov.l REG_CMD_VTX, @REG_TMP
jsr @REG_CLIPFUNC
nop
mov.l REG_CMD_VTX, @REG_TMP
TMP_SET_B
mov REG_V0, REG_CLIP1
mov REG_V1, REG_CLIP2
mov.l REG_CMD_EOL, @REG_TMP
jsr @REG_CLIPFUNC
mov.l REG_CMD_EOL, @REG_TMP
mov.l REG_CMD_VTX, @REG_V3
TransformVertex REG_V2
@ -554,7 +527,7 @@ _Case_1_1_1_0:
TransformVertex REG_TMP
rts
lds.l @r15+,pr
lds r13,pr
_Case_1_1_1_1:
! Triangle strip: {1,2,0} {2,0,3}
@ -579,16 +552,16 @@ _ProcessVertexList:
mov.l r14,@-r15
sts.l pr,@-r15
! STORE FPU REGISTERS
fmov.s fr8,@-r15
fmov.s fr9,@-r15
fmov.s fr10,@-r15
fmov.s fr11,@-r15
fmov.s F_HW,@-r15
fmov.s F_HH,@-r15
fmov.s F_XP,@-r15
fmov.s F_YP,@-r15
! VIEWPORT SETUP
mov.l .VP_1,r0 ! LS, r0 = &vp
fmov.s @r0+,fr8 ! LS, fr8 = vp.HWIDTH
fmov.s @r0+,fr9 ! LS, fr9 = vp.HHEIGHT
fmov.s @r0+,fr10 ! LS, fr10 = vp.X_PLUS_HWIDTH
fmov.s @r0+,fr11 ! LS, fr11 = vp.Y_PLUS_HHEIGHT
mov.l .VP_1,r0 ! LS, &vp
fmov.s @r0+,F_HW ! LS, vp.HWIDTH
fmov.s @r0+,F_HH ! LS, vp.HHEIGHT
fmov.s @r0+,F_XP ! LS, vp.X_PLUS_HWIDTH
fmov.s @r0+,F_YP ! LS, vp.Y_PLUS_HHEIGHT
! REGISTER SETUP
mov r4,r14
mov r4,r13
@ -616,17 +589,17 @@ DO_CMD:
add #4,r4
mov.l .VP_1,r2
! Load VIEWPORT registers
fmov.s @r4+, fr8 ! VIEWPORT_HWIDTH = src->x
fmov.s @r4+, fr9 ! VIEWPORT_HHEIGHT = src->y
fmov.s @r4+,fr10 ! VIEWPORT_X_PLUS_HWIDTH = src->z
fmov.s @r4+,F_HW ! VIEWPORT_HWIDTH = src->x
fmov.s @r4+,F_HH ! VIEWPORT_HHEIGHT = src->y
fmov.s @r4+,F_XP ! VIEWPORT_X_PLUS_HWIDTH = src->z
add #16,r2
fmov.s @r4+,fr11 ! VIEWPORT_Y_PLUS_HHEIGHT = src->u
fmov.s @r4+,F_YP ! VIEWPORT_Y_PLUS_HHEIGHT = src->u
! And store to vp global
fmov.s fr11,@-r2
fmov.s fr10,@-r2
fmov.s fr9,@-r2
fmov.s F_YP,@-r2
fmov.s F_XP,@-r2
fmov.s F_HH,@-r2
bra NEXT_ITER
fmov.s fr8,@-r2
fmov.s F_HW,@-r2
SUBMIT_LOOP:
mov.l @r13,r0 ! FLAGS = CUR->flags
@ -665,15 +638,15 @@ NEXT_ITER:
! VIEWPORT SAVE
mov.l .VP_1,r0
add #16,r0
fmov.s fr11,@-r0
fmov.s fr10,@-r0
fmov.s fr9,@-r0
fmov.s fr8,@-r0
fmov.s F_YP,@-r0
fmov.s F_XP,@-r0
fmov.s F_HH,@-r0
fmov.s F_HW,@-r0
! RESTORE FPU REGISTERS
fmov.s @r15+, fr8
fmov.s @r15+, fr9
fmov.s @r15+,fr10
fmov.s @r15+,fr11
fmov.s @r15+,F_YP
fmov.s @r15+,F_XP
fmov.s @r15+,F_HH
fmov.s @r15+,F_HW
! RESTORE CPU REGISTERS
lds.l @r15+,pr
mov.l @r15+,r14