mirror of
https://github.com/thug1src/thug.git
synced 2025-01-22 05:43:47 +00:00
614 lines
20 KiB
Plaintext
614 lines
20 KiB
Plaintext
|
|
||
|
|
.global MPGStart0
|
||
|
|
.global MPGEnd0
|
||
|
|
|
||
|
|
.global TestFunc
|
||
|
|
.global InitialiseOccluders
|
||
|
|
.global TestSphereAgainstOccluders
|
||
|
|
.global RayTriangleCollision
|
||
|
|
.global BatchRayTriangleCollision
|
||
|
|
.global ViewCullTest
|
||
|
|
.global OuterCullTest
|
||
|
|
.global BothCullTests
|
||
|
|
|
||
|
|
; align to a 2^4=16 byte boundary, so it can be the target of a dma::ref
|
||
|
|
.align 4
|
||
|
|
|
||
|
|
MPGStart0:
|
||
|
|
|
||
|
|
MPG 0, *
|
||
|
|
|
||
|
|
|
||
|
|
;-----------------------------------------------------------------------------------------------------------------------------
|
||
|
|
|
||
|
|
|
||
|
|
TestFunc:
|
||
|
|
|
||
|
|
; just a little test func to make sure I understand how to communicate with a vu0 microprogram
|
||
|
|
|
||
|
|
; integer register test
|
||
|
|
NOP IADD VI02,VI01,VI01
|
||
|
|
|
||
|
|
; float register test
|
||
|
|
ADDw VF02,VF01,VF00w NOP
|
||
|
|
|
||
|
|
; memory test
|
||
|
|
NOP LQ VF03,0(VI00)
|
||
|
|
ADDw VF03,VF03,VF00w NOP
|
||
|
|
|
||
|
|
; quit
|
||
|
|
NOP[E] NOP
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
InitialiseOccluders:
|
||
|
|
|
||
|
|
; this function will be optimised later... not a priority now since only called once per frame (per camera)
|
||
|
|
|
||
|
|
; assumes an initial state where the 5 planes of occluder 0 are at addresses 0,1,2,3,4
|
||
|
|
; occluder 1 at 5,6,7,8,9, etc, and the number of occluders is in VI01
|
||
|
|
|
||
|
|
|
||
|
|
; get address of 1st spare occluder
|
||
|
|
NOP IADD VI05,VI01,VI01
|
||
|
|
NOP IADD VI05,VI05,VI05
|
||
|
|
NOP IADD VI05,VI05,VI01
|
||
|
|
|
||
|
|
; set up a sentinel plane that's colossally far away
|
||
|
|
NOP LOI -1.0e+35
|
||
|
|
MULi VF01,VF00,I NOP
|
||
|
|
|
||
|
|
; add 1 sentinel occluder (5 sentinel planes)
|
||
|
|
NOP SQI VF01,(VI05++)
|
||
|
|
NOP SQI VF01,(VI05++)
|
||
|
|
NOP SQI VF01,(VI05++)
|
||
|
|
NOP SQI VF01,(VI05++)
|
||
|
|
NOP SQI VF01,(VI05++)
|
||
|
|
|
||
|
|
; rearrange memory
|
||
|
|
NOP IADDIU VI02,VI00,0
|
||
|
|
NOP MFIR.xyzw VF31,VI00
|
||
|
|
|
||
|
|
MemLoop:
|
||
|
|
; load 4 occluders, 5 planes each
|
||
|
|
|
||
|
|
NOP LQ VF01,0(VI02)
|
||
|
|
NOP LQ VF02,5(VI02)
|
||
|
|
NOP LQ VF03,10(VI02)
|
||
|
|
NOP LQ VF04,15(VI02)
|
||
|
|
NOP LQ VF05,1(VI02)
|
||
|
|
NOP LQ VF06,6(VI02)
|
||
|
|
NOP LQ VF07,11(VI02)
|
||
|
|
NOP LQ VF08,16(VI02)
|
||
|
|
NOP LQ VF09,2(VI02)
|
||
|
|
NOP LQ VF10,7(VI02)
|
||
|
|
NOP LQ VF11,12(VI02)
|
||
|
|
NOP LQ VF12,17(VI02)
|
||
|
|
NOP LQ VF13,3(VI02)
|
||
|
|
NOP LQ VF14,8(VI02)
|
||
|
|
NOP LQ VF15,13(VI02)
|
||
|
|
NOP LQ VF16,18(VI02)
|
||
|
|
NOP LQ VF17,4(VI02)
|
||
|
|
NOP LQ VF18,9(VI02)
|
||
|
|
NOP LQ VF19,14(VI02)
|
||
|
|
NOP LQ VF20,19(VI02)
|
||
|
|
|
||
|
|
|
||
|
|
; transpose VF01,VF02,VF03,VF04
|
||
|
|
|
||
|
|
NOP MOVE VF21,VF01
|
||
|
|
NOP MOVE VF22,VF02
|
||
|
|
NOP MOVE VF23,VF03
|
||
|
|
|
||
|
|
ADDx.y VF01,VF31,VF02x NOP
|
||
|
|
ADDx.z VF01,VF31,VF03x NOP
|
||
|
|
ADDx.w VF01,VF31,VF04x NOP
|
||
|
|
|
||
|
|
ADDy.x VF02,VF31,VF21y NOP
|
||
|
|
ADDy.z VF02,VF31,VF03y NOP
|
||
|
|
ADDy.w VF02,VF31,VF04y NOP
|
||
|
|
|
||
|
|
ADDz.x VF03,VF31,VF21z NOP
|
||
|
|
ADDz.y VF03,VF31,VF22z NOP
|
||
|
|
ADDz.w VF03,VF31,VF04z NOP
|
||
|
|
|
||
|
|
ADDw.x VF04,VF31,VF21w NOP
|
||
|
|
ADDw.y VF04,VF31,VF22w NOP
|
||
|
|
ADDw.z VF04,VF31,VF23w NOP
|
||
|
|
|
||
|
|
|
||
|
|
; transpose VF05,VF06,VF07,VF08
|
||
|
|
|
||
|
|
NOP MOVE VF21,VF05
|
||
|
|
NOP MOVE VF22,VF06
|
||
|
|
NOP MOVE VF23,VF07
|
||
|
|
|
||
|
|
ADDx.y VF05,VF31,VF06x NOP
|
||
|
|
ADDx.z VF05,VF31,VF07x NOP
|
||
|
|
ADDx.w VF05,VF31,VF08x NOP
|
||
|
|
|
||
|
|
ADDy.x VF06,VF31,VF21y NOP
|
||
|
|
ADDy.z VF06,VF31,VF07y NOP
|
||
|
|
ADDy.w VF06,VF31,VF08y NOP
|
||
|
|
|
||
|
|
ADDz.x VF07,VF31,VF21z NOP
|
||
|
|
ADDz.y VF07,VF31,VF22z NOP
|
||
|
|
ADDz.w VF07,VF31,VF08z NOP
|
||
|
|
|
||
|
|
ADDw.x VF08,VF31,VF21w NOP
|
||
|
|
ADDw.y VF08,VF31,VF22w NOP
|
||
|
|
ADDw.z VF08,VF31,VF23w NOP
|
||
|
|
|
||
|
|
|
||
|
|
; transpose VF09,VF10,VF11,VF12
|
||
|
|
|
||
|
|
NOP MOVE VF21,VF09
|
||
|
|
NOP MOVE VF22,VF10
|
||
|
|
NOP MOVE VF23,VF11
|
||
|
|
|
||
|
|
ADDx.y VF09,VF31,VF10x NOP
|
||
|
|
ADDx.z VF09,VF31,VF11x NOP
|
||
|
|
ADDx.w VF09,VF31,VF12x NOP
|
||
|
|
|
||
|
|
ADDy.x VF10,VF31,VF21y NOP
|
||
|
|
ADDy.z VF10,VF31,VF11y NOP
|
||
|
|
ADDy.w VF10,VF31,VF12y NOP
|
||
|
|
|
||
|
|
ADDz.x VF11,VF31,VF21z NOP
|
||
|
|
ADDz.y VF11,VF31,VF22z NOP
|
||
|
|
ADDz.w VF11,VF31,VF12z NOP
|
||
|
|
|
||
|
|
ADDw.x VF12,VF31,VF21w NOP
|
||
|
|
ADDw.y VF12,VF31,VF22w NOP
|
||
|
|
ADDw.z VF12,VF31,VF23w NOP
|
||
|
|
|
||
|
|
|
||
|
|
; transpose VF13,VF14,VF15,VF16
|
||
|
|
|
||
|
|
NOP MOVE VF21,VF13
|
||
|
|
NOP MOVE VF22,VF14
|
||
|
|
NOP MOVE VF23,VF15
|
||
|
|
|
||
|
|
ADDx.y VF13,VF31,VF14x NOP
|
||
|
|
ADDx.z VF13,VF31,VF15x NOP
|
||
|
|
ADDx.w VF13,VF31,VF16x NOP
|
||
|
|
|
||
|
|
ADDy.x VF14,VF31,VF21y NOP
|
||
|
|
ADDy.z VF14,VF31,VF15y NOP
|
||
|
|
ADDy.w VF14,VF31,VF16y NOP
|
||
|
|
|
||
|
|
ADDz.x VF15,VF31,VF21z NOP
|
||
|
|
ADDz.y VF15,VF31,VF22z NOP
|
||
|
|
ADDz.w VF15,VF31,VF16z NOP
|
||
|
|
|
||
|
|
ADDw.x VF16,VF31,VF21w NOP
|
||
|
|
ADDw.y VF16,VF31,VF22w NOP
|
||
|
|
ADDw.z VF16,VF31,VF23w NOP
|
||
|
|
|
||
|
|
|
||
|
|
; transpose VF17,VF18,VF19,VF20
|
||
|
|
|
||
|
|
NOP MOVE VF21,VF17
|
||
|
|
NOP MOVE VF22,VF18
|
||
|
|
NOP MOVE VF23,VF19
|
||
|
|
|
||
|
|
ADDx.y VF17,VF31,VF18x NOP
|
||
|
|
ADDx.z VF17,VF31,VF19x NOP
|
||
|
|
ADDx.w VF17,VF31,VF20x NOP
|
||
|
|
|
||
|
|
ADDy.x VF18,VF31,VF21y NOP
|
||
|
|
ADDy.z VF18,VF31,VF19y NOP
|
||
|
|
ADDy.w VF18,VF31,VF20y NOP
|
||
|
|
|
||
|
|
ADDz.x VF19,VF31,VF21z NOP
|
||
|
|
ADDz.y VF19,VF31,VF22z NOP
|
||
|
|
ADDz.w VF19,VF31,VF20z NOP
|
||
|
|
|
||
|
|
ADDw.x VF20,VF31,VF21w NOP
|
||
|
|
ADDw.y VF20,VF31,VF22w NOP
|
||
|
|
ADDw.z VF20,VF31,VF23w NOP
|
||
|
|
|
||
|
|
|
||
|
|
; store the rearranged occluders
|
||
|
|
|
||
|
|
NOP SQI VF04,(VI02++)
|
||
|
|
NOP SQI VF01,(VI02++)
|
||
|
|
NOP SQI VF02,(VI02++)
|
||
|
|
NOP SQI VF03,(VI02++)
|
||
|
|
NOP SQI VF08,(VI02++)
|
||
|
|
NOP SQI VF05,(VI02++)
|
||
|
|
NOP SQI VF06,(VI02++)
|
||
|
|
NOP SQI VF07,(VI02++)
|
||
|
|
NOP SQI VF12,(VI02++)
|
||
|
|
NOP SQI VF09,(VI02++)
|
||
|
|
NOP SQI VF10,(VI02++)
|
||
|
|
NOP SQI VF11,(VI02++)
|
||
|
|
NOP SQI VF16,(VI02++)
|
||
|
|
NOP SQI VF13,(VI02++)
|
||
|
|
NOP SQI VF14,(VI02++)
|
||
|
|
NOP SQI VF15,(VI02++)
|
||
|
|
NOP SQI VF20,(VI02++)
|
||
|
|
NOP SQI VF17,(VI02++)
|
||
|
|
NOP SQI VF18,(VI02++)
|
||
|
|
NOP SQI VF19,(VI02++)
|
||
|
|
|
||
|
|
|
||
|
|
; loop if we haven't just processed the sentinel planes
|
||
|
|
|
||
|
|
NOP ISUB VI01,VI02,VI05
|
||
|
|
NOP NOP
|
||
|
|
NOP IBLTZ VI01,MemLoop
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
; quit
|
||
|
|
NOP[E] NOP
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
;-----------------------------------------------------------------------------------------------------------------------------
|
||
|
|
|
||
|
|
|
||
|
|
TestSphereAgainstOccluders:
|
||
|
|
|
||
|
|
; sphere is in VF07 as (x,y,z,R)
|
||
|
|
; for each plane p, we must compute px*x + py*w + pz*z + pw + R
|
||
|
|
; uses a sentinel occluder to terminate
|
||
|
|
|
||
|
|
NOP IADDIU VI02,VI00,0 ; initialise pointer
|
||
|
|
|
||
|
|
; loop prologue
|
||
|
|
ADD.y VF05,VF00,VF00 LQI VF04,(VI02++)
|
||
|
|
NOP LQI VF01,(VI02++)
|
||
|
|
NOP LQI VF02,(VI02++)
|
||
|
|
NOP LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y LQI VF02,(VI02++)
|
||
|
|
MADDz VF00,VF03,VF07z IADDIU VI10,VI00,0xF0 ; VI10 = mask for Sxyzw FMAC flags
|
||
|
|
|
||
|
|
; main loop, tests 4 occluders (20 planes) per pass
|
||
|
|
Loop: ADDw.y VF05,VF05,VF00w LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y FMAND VI01,VI10
|
||
|
|
MADDz VF00,VF03,VF07z LQI VF02,(VI02++)
|
||
|
|
ADDw.y VF05,VF05,VF00w LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y FMAND VI01,VI01
|
||
|
|
MADDz VF00,VF03,VF07z LQI VF02,(VI02++)
|
||
|
|
ADDw.y VF05,VF05,VF00w LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y FMAND VI01,VI01
|
||
|
|
MADDz VF00,VF03,VF07z LQI VF02,(VI02++)
|
||
|
|
ADDw.y VF05,VF05,VF00w LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y FMAND VI01,VI01
|
||
|
|
MADDz VF00,VF03,VF07z LQI VF02,(VI02++)
|
||
|
|
FTOI0.y VF06,VF05 LQI VF03,(VI02++)
|
||
|
|
ADDAw ACC,VF04,VF07w LQI VF04,(VI02++)
|
||
|
|
MADDAx ACC,VF01,VF07x LQI VF01,(VI02++)
|
||
|
|
MADDAy ACC,VF02,VF07y FMAND VI01,VI01
|
||
|
|
NOP IBEQ VI01,VI00,Loop
|
||
|
|
MADDz VF00,VF03,VF07z LQI VF02,(VI02++)
|
||
|
|
|
||
|
|
; (macflags & 0x80) => possible occluder in 1st slot of group of 4
|
||
|
|
NOP ISUBIU VI10,VI01,0x80
|
||
|
|
NOP MTIR VI06,VF06y
|
||
|
|
NOP IBGEZ VI10,TestForSentinel
|
||
|
|
NOP ISUBIU VI03,VI02,22
|
||
|
|
|
||
|
|
; (macflags & 0x40) => possible occluder in 2nd slot of group of 4
|
||
|
|
NOP ISUBIU VI10,VI01,0x40
|
||
|
|
NOP IADDIU VI06,VI06,1
|
||
|
|
NOP IBGEZ VI10,TestForSentinel
|
||
|
|
NOP ISUBIU VI03,VI02,17
|
||
|
|
|
||
|
|
; (macflags & 0x20) => possible occluder in 3rd slot of group of 4
|
||
|
|
NOP ISUBIU VI10,VI01,0x20
|
||
|
|
NOP IADDIU VI06,VI06,1
|
||
|
|
NOP IBGEZ VI10,TestForSentinel
|
||
|
|
NOP ISUBIU VI03,VI02,12
|
||
|
|
|
||
|
|
; (macflags & 0x10) => possible occluder in 4th slot of group of 4
|
||
|
|
NOP IADDIU VI06,VI06,1
|
||
|
|
NOP ISUBIU VI03,VI02,7
|
||
|
|
|
||
|
|
TestForSentinel:
|
||
|
|
; sentinel if VI03==VI05, else occluded
|
||
|
|
NOP[E] ISUB VI01,VI03,VI05
|
||
|
|
NOP ISUBIU VI06,VI06,4
|
||
|
|
|
||
|
|
;-----------------------------------------------------------------------------------------------------------------------------
|
||
|
|
|
||
|
|
BatchRayTriangleCollision:
|
||
|
|
NOP ILW.y VI04, 0(VI00) ; total number of collision tests
|
||
|
|
NOP ILW.x VI05, 0(VI00) ; starting index
|
||
|
|
|
||
|
|
NOP IADDIU VI01, VI00, 0 ; init number of found collisions
|
||
|
|
NOP IADDIU VI03, VI00, 1 ; init address
|
||
|
|
NOP IADD VI06, VI00, VI03 ; init output address (writes over input)
|
||
|
|
|
||
|
|
NOP ISUB VI04, VI04, VI05 ; decrement the start index from the array size
|
||
|
|
|
||
|
|
NOP LQ VF06, 2(VI03) ; v0
|
||
|
|
NOP LQ VF08, 4(VI03) ; v2
|
||
|
|
NOP LQ VF07, 3(VI03) ; v1
|
||
|
|
NOP LQ VF04, 0(VI03) ; rayStart
|
||
|
|
NOP LQ VF05, 1(VI03) ; rayDir
|
||
|
|
|
||
|
|
NextCollision:
|
||
|
|
SUB.xyz VF10, VF08, VF06 NOP ; edge2 = v2 - v0
|
||
|
|
SUB.xyz VF09, VF07, VF06 MFIR.w VF17w, VI05 ; edge1 = v1 - v0 ; store index in VF17w (output)
|
||
|
|
SUB.xyz VF13, VF04, VF06 FSSET 0 ; tvec = rayStart - v0
|
||
|
|
ADDW.x VF02, VF00, VF00w LOI 1.00001 ; Load 1.0f to VF02x ; load 1 + EPSILON into I
|
||
|
|
|
||
|
|
OPMULA.xyz ACC, VF05, VF10 NOP ; Cross product of ray normal and edge2 (pvec)
|
||
|
|
OPMSUB.xyz VF12,VF10, VF05 NOP ; Second part of cross product
|
||
|
|
OPMULA.xyz ACC, VF13, VF09 IADDIU VI14, VI00, 0x80 ; qvec = crossProd(tvec, edge1) ; Set the mask X sign flag
|
||
|
|
OPMSUB.xyz VF14,VF09, VF13 IADDIU VI13, VI00, 0x10 ; Set the mask W sign flag
|
||
|
|
SUBi.x VF03,VF00, I NOP ; put -(1 + EPSILON) in VF03x
|
||
|
|
|
||
|
|
MUL.xyz VF11,VF09, VF12 NOP ; det = edge1 * pvec [start]
|
||
|
|
MUL.xyz VF15,VF13, VF12 NOP ; u = tvec * pvec [start]
|
||
|
|
MULAx.x ACC, VF09, VF12x IADDI VI04, VI04, -1 ; decrement number of collision tests
|
||
|
|
MADDAy.x ACC, VF02, VF11y IADDIU VI03, VI03, 5 ; increment address
|
||
|
|
MADDz.x VF11,VF02, VF11z NOP ; det = edge1 * pvec [ready]
|
||
|
|
MULAx.x ACC, VF02, VF15x LOI 0.00001 ; load 1 + EPSILON into I
|
||
|
|
MADDAy.x ACC, VF02, VF15y LQ VF04, 0(VI03) ; next rayStart
|
||
|
|
MUL.xyz VF17,VF10, VF14 NOP ; t = edge2 * qvec [start]
|
||
|
|
SUBi.x VF01,VF11, I DIV Q, VF00w,VF11x ; If det < EPSILON ; Q = 1.0f / det
|
||
|
|
|
||
|
|
MADDz.x VF15,VF02, VF15z LQ VF06, 2(VI03) ; u = tvec * pvec [ready] ; next v0
|
||
|
|
MULAx.x ACC, VF10, VF14x LQ VF08, 4(VI03) ; next v2
|
||
|
|
MADDAy.x ACC, VF02, VF17 LQ VF07, 3(VI03) ; next v1
|
||
|
|
MUL.xyz VF16,VF05, VF14 FMAND VI15,VI14 ; v = rayDir * qvec [start] ; check if det result (MAC register)
|
||
|
|
MADDz.x VF17,VF02, VF17z IBNE VI15,VI00,BatchRayTriDone ; t = edge2 * qvec [ready]
|
||
|
|
MULAx.x ACC, VF05, VF14x LQ VF05, 1(VI03) ; next rayDir
|
||
|
|
MULQ.x VF15,VF15, Q NOP ; u = (tvec * pvec) / det
|
||
|
|
MADDAy.x ACC, VF02, VF16 NOP
|
||
|
|
MADDz.x VF16,VF02, VF16z NOP ; v = rayDir * qvec [ready]
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
ADDx.x VF00,VF15, VF03x FMAND VI15,VI14 ; if u > 1 + EPSILON ; check if u result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,BatchRayTriDone
|
||
|
|
MULQ.x VF16,VF16, Q NOP ; v = (rayDir * qvec) / det;
|
||
|
|
MULQ.x VF17,VF17, Q NOP ; t = (edge2 * qvec) / det
|
||
|
|
NOP FMAND VI15,VI14 ; check if u > 1 + EPSILON result (MAC register)
|
||
|
|
NOP IBEQ VI15,VI00,BatchRayTriDone
|
||
|
|
|
||
|
|
ADDA.x ACC ,VF15, VF16 FMAND VI15,VI14 ; check if v < 0 result (MAC register)
|
||
|
|
MADDw.x VF31,VF03, VF00w IBNE VI15,VI00,BatchRayTriDone ; if (u+v) > 1 + EPSILON [start]
|
||
|
|
|
||
|
|
NOP NOP
|
||
|
|
SUB.x VF00,VF17, VF00 NOP
|
||
|
|
NOP NOP
|
||
|
|
SUBx.w VF00,VF00, VF17 FMAND VI15,VI14 ; if t > 1 [start] ; check if u+v > 1 + EPSILON result (MAC register)
|
||
|
|
NOP IBEQ VI15,VI00,BatchRayTriDone
|
||
|
|
|
||
|
|
NOP FMAND VI15,VI14 ; check if t < 0 result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,BatchRayTriDone
|
||
|
|
|
||
|
|
NOP FMAND VI15,VI13 ; check if t > 1 result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,BatchRayTriDone
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
NOP SQI.xw VF17, (VI06++) ; store distance (x) and index (w)
|
||
|
|
NOP IADDIU VI01, VI01, 1 ; inc number of found collisions
|
||
|
|
|
||
|
|
BatchRayTriDone:
|
||
|
|
NOP IBNE VI04, VI00,NextCollision
|
||
|
|
NOP IADDIU VI05, VI05, 1 ; increment index for output
|
||
|
|
|
||
|
|
NOP[T] NOP
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
RayTriangleCollision:
|
||
|
|
|
||
|
|
SUB.xyz VF10, VF08, VF06 NOP ; edge2 = v2 - v0
|
||
|
|
SUB.xyz VF09, VF07, VF06 FSSET 0 ; edge1 = v1 - v0
|
||
|
|
SUB.xyz VF13, VF04, VF06 NOP ; tvec = rayStart - v0
|
||
|
|
ADDW.x VF02, VF00, VF00w LOI 1.00001 ; Load 1.0f to VF02x ; load 1 + EPSILON into I
|
||
|
|
|
||
|
|
OPMULA.xyz ACC, VF05, VF10 NOP ; Cross product of ray normal and edge2 (pvec)
|
||
|
|
OPMSUB.xyz VF12,VF10, VF05 NOP ; Second part of cross product
|
||
|
|
OPMULA.xyz ACC, VF13, VF09 IADDIU VI14, VI00, 0x80 ; qvec = crossProd(tvec, edge1) ; Set the mask X sign flag
|
||
|
|
OPMSUB.xyz VF14,VF09, VF13 IADDIU VI13, VI00, 0x10 ; Set the mask W sign flag
|
||
|
|
SUBi.x VF03,VF00, I IADDIU VI02, VI00, 0 ; put -(1 + EPSILON) in VF03x ; store 0 for return value
|
||
|
|
|
||
|
|
MUL.xyz VF11,VF09, VF12 NOP ; det = edge1 * pvec [start]
|
||
|
|
MUL.xyz VF15,VF13, VF12 NOP ; u = tvec * pvec [start]
|
||
|
|
MULAx.x ACC, VF09, VF12x NOP
|
||
|
|
MADDAy.x ACC, VF02, VF11y NOP
|
||
|
|
MADDz.x VF11,VF02, VF11z NOP ; det = edge1 * pvec [ready]
|
||
|
|
MULAx.x ACC, VF02, VF15x LOI 0.00001 ; load 1 + EPSILON into I
|
||
|
|
MADDAy.x ACC, VF02, VF15y NOP
|
||
|
|
MUL.xyz VF17,VF10, VF14 NOP ; t = edge2 * qvec [start]
|
||
|
|
SUBi.x VF01,VF11, I DIV Q, VF00w,VF11x ; If det < EPSILON ; Q = 1.0f / det
|
||
|
|
|
||
|
|
MADDz.x VF15,VF02, VF15z NOP ; u = tvec * pvec [ready]
|
||
|
|
MULAx.x ACC, VF10, VF14x NOP
|
||
|
|
MADDAy.x ACC, VF02, VF17 NOP
|
||
|
|
MUL.xyz VF16,VF05, VF14 FMAND VI15,VI14 ; v = rayDir * qvec [start] ; check if det result (MAC register)
|
||
|
|
MADDz.x VF17,VF02, VF17z IBNE VI15,VI00,vu0RayTriDone ; t = edge2 * qvec [ready]
|
||
|
|
MULAx.x ACC, VF05, VF14x NOP
|
||
|
|
MULQ.x VF15,VF15, Q NOP ; u = (tvec * pvec) / det
|
||
|
|
MADDAy.x ACC, VF02, VF16 NOP
|
||
|
|
MADDz.x VF16,VF02, VF16z NOP ; v = rayDir * qvec [ready]
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
ADDx.x VF00,VF15, VF03x FMAND VI15,VI14 ; if u > 1 + EPSILON ; check if u result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,vu0RayTriDone
|
||
|
|
MULQ.x VF16,VF16, Q NOP ; v = (rayDir * qvec) / det;
|
||
|
|
MULQ.x VF17,VF17, Q NOP ; t = (edge2 * qvec) / det
|
||
|
|
NOP FMAND VI15,VI14 ; check if u > 1 + EPSILON result (MAC register)
|
||
|
|
NOP IBEQ VI15,VI00,vu0RayTriDone
|
||
|
|
|
||
|
|
ADDA.x ACC ,VF15, VF16 FMAND VI15,VI14 ; check if v < 0 result (MAC register)
|
||
|
|
MADDw.x VF31,VF03, VF00w IBNE VI15,VI00,vu0RayTriDone ; if (u+v) > 1 + EPSILON [start]
|
||
|
|
|
||
|
|
NOP NOP
|
||
|
|
SUB.x VF00,VF17, VF00 NOP
|
||
|
|
NOP NOP
|
||
|
|
SUBx.w VF00,VF00, VF17 FMAND VI15,VI14 ; if t > 1 [start] ; check if u+v > 1 + EPSILON result (MAC register)
|
||
|
|
NOP IBEQ VI15,VI00,vu0RayTriDone
|
||
|
|
|
||
|
|
NOP FMAND VI15,VI14 ; check if t < 0 result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,vu0RayTriDone
|
||
|
|
|
||
|
|
NOP FMAND VI15,VI13 ; check if t > 1 result (MAC register)
|
||
|
|
NOP IBNE VI15,VI00,vu0RayTriDone
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
;qmfc2 $9, $vf17 # move t to $9
|
||
|
|
;sw $9, 0(%1) # and write out
|
||
|
|
NOP IADDIU VI02, VI00, 1 ; store 1 for return value
|
||
|
|
|
||
|
|
vu0RayTriDone:
|
||
|
|
|
||
|
|
NOP[E] NOP
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
;-----------------------------------------------------------------------------------------------------------------------------
|
||
|
|
; new bounding volume / frustum intersection tests
|
||
|
|
|
||
|
|
|
||
|
|
; VF08 = (x0,y0,z0,r)
|
||
|
|
; VF09 = (bx,by,bz,?)
|
||
|
|
|
||
|
|
; VF10 = (nx)
|
||
|
|
; VF11 = (ny)
|
||
|
|
; VF12 = (nz)
|
||
|
|
; VF13 = (|nx|)
|
||
|
|
; VF14 = (|ny|)
|
||
|
|
; VF15 = (|nz|)
|
||
|
|
; VF16 = (nw)
|
||
|
|
|
||
|
|
; VF17 = (nx)
|
||
|
|
; VF18 = (ny)
|
||
|
|
; VF19 = (nz)
|
||
|
|
; VF20 = (|nx|)
|
||
|
|
; VF21 = (|ny|)
|
||
|
|
; VF22 = (|nz|)
|
||
|
|
; VF23 = (nw)
|
||
|
|
|
||
|
|
; VF24 = (nx)
|
||
|
|
; VF25 = (ny)
|
||
|
|
; VF26 = (nz)
|
||
|
|
; VF27 = (|nx|)
|
||
|
|
; VF28 = (|ny|)
|
||
|
|
; VF29 = (|nz|)
|
||
|
|
; VF30 = (nw)
|
||
|
|
|
||
|
|
; VF31 = (1)
|
||
|
|
|
||
|
|
|
||
|
|
ViewCullTest:
|
||
|
|
|
||
|
|
ADDAx ACC,VF16,VF00x NOP
|
||
|
|
MADDAx ACC,VF10,VF08x NOP
|
||
|
|
MADDAy ACC,VF11,VF08y NOP
|
||
|
|
MADDAz ACC,VF12,VF08z NOP
|
||
|
|
MADDw VF00,VF31,VF08w NOP
|
||
|
|
MADDAx ACC,VF13,VF09x NOP
|
||
|
|
MADDAy ACC,VF14,VF09y NOP
|
||
|
|
MADDz VF00,VF15,VF09z NOP
|
||
|
|
|
||
|
|
ADDAx.xy ACC,VF23,VF00x FMOR VI01,VI00
|
||
|
|
MADDAx.xy ACC,VF17,VF08x NOP
|
||
|
|
MADDAy.xy ACC,VF18,VF08y NOP
|
||
|
|
MADDAz.xy ACC,VF19,VF08z FMOR VI01,VI01
|
||
|
|
MADDw.xy VF00,VF31,VF08w NOP
|
||
|
|
MADDAx.xy ACC,VF20,VF09x NOP
|
||
|
|
MADDAy.xy ACC,VF21,VF09y NOP
|
||
|
|
MADDz.xy VF00,VF22,VF09z NOP
|
||
|
|
|
||
|
|
NOP FMOR VI01,VI01
|
||
|
|
NOP NOP
|
||
|
|
NOP NOP
|
||
|
|
NOP[E] FMOR VI01,VI01
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
OuterCullTest:
|
||
|
|
|
||
|
|
ADDAx.zw ACC,VF23,VF00x NOP
|
||
|
|
MADDAx.zw ACC,VF17,VF08x NOP
|
||
|
|
MADDAy.zw ACC,VF18,VF08y NOP
|
||
|
|
MADDAz.zw ACC,VF19,VF08z NOP
|
||
|
|
MSUBw.zw VF00,VF31,VF08w NOP
|
||
|
|
MSUBAx.zw ACC,VF20,VF09x NOP
|
||
|
|
MSUBAy.zw ACC,VF21,VF09y NOP
|
||
|
|
MSUBz.zw VF00,VF22,VF09z NOP
|
||
|
|
|
||
|
|
ADDAx ACC,VF30,VF00x FMOR VI01,VI00
|
||
|
|
MADDAx ACC,VF24,VF08x NOP
|
||
|
|
MADDAy ACC,VF25,VF08y NOP
|
||
|
|
MADDAz ACC,VF26,VF08z FMOR VI01,VI01
|
||
|
|
MSUBw VF00,VF31,VF08w NOP
|
||
|
|
MSUBAx ACC,VF27,VF09x NOP
|
||
|
|
MSUBAy ACC,VF28,VF09y NOP
|
||
|
|
MSUBz VF00,VF29,VF09z NOP
|
||
|
|
|
||
|
|
NOP FMOR VI01,VI01
|
||
|
|
NOP NOP
|
||
|
|
NOP NOP
|
||
|
|
NOP[E] FMOR VI01,VI01
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
BothCullTests:
|
||
|
|
|
||
|
|
MULAx.zw ACC,VF20,VF09x NOP
|
||
|
|
MADDAy.zw ACC,VF21,VF09y NOP
|
||
|
|
MADDz.zw VF01,VF22,VF09z NOP
|
||
|
|
|
||
|
|
ADDAx ACC,VF16,VF00x NOP
|
||
|
|
MADDAx ACC,VF10,VF08x NOP
|
||
|
|
MADDAy ACC,VF11,VF08y NOP
|
||
|
|
MADDAz ACC,VF12,VF08z NOP
|
||
|
|
MADDw VF00,VF31,VF08w NOP
|
||
|
|
MADDAx ACC,VF13,VF09x NOP
|
||
|
|
MADDAy ACC,VF14,VF09y NOP
|
||
|
|
MADDz VF00,VF15,VF09z NOP
|
||
|
|
|
||
|
|
ADDAx.zw ACC,VF23,VF00x FMOR VI01,VI00
|
||
|
|
MADDAx.zw ACC,VF17,VF08x NOP
|
||
|
|
MADDAy.zw ACC,VF18,VF08y NOP
|
||
|
|
MADDAz.zw ACC,VF19,VF08z FMOR VI01,VI01
|
||
|
|
MADDx.z VF05,VF00,VF00x NOP
|
||
|
|
MADDw.zw VF00,VF31,VF08w NOP
|
||
|
|
MADDw.zw VF00,VF01,VF00w NOP
|
||
|
|
MSUBw.zw VF00,VF31,VF08w NOP
|
||
|
|
MSUBw.zw VF00,VF01,VF00w NOP
|
||
|
|
|
||
|
|
ADDAx ACC,VF30,VF00x FMOR VI01,VI01
|
||
|
|
MADDAx ACC,VF24,VF08x FMOR VI01,VI01
|
||
|
|
MADDAy ACC,VF25,VF08y FMOR VI11,VI00
|
||
|
|
MADDAz ACC,VF26,VF08z FMOR VI11,VI11
|
||
|
|
MSUBw VF00,VF31,VF08w NOP
|
||
|
|
MSUBAx ACC,VF27,VF09x NOP
|
||
|
|
MSUBAy ACC,VF28,VF09y NOP
|
||
|
|
MSUBz VF00,VF29,VF09z NOP
|
||
|
|
|
||
|
|
SUBz.x VF05,VF00,VF05z FMOR VI11,VI11
|
||
|
|
NOP NOP
|
||
|
|
NOP NOP
|
||
|
|
NOP FMOR VI11,VI11
|
||
|
|
|
||
|
|
NOP[E] NOP
|
||
|
|
NOP NOP
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
;-----------------------------------------------------------------------------------------------------------------------------
|
||
|
|
|
||
|
|
|
||
|
|
.EndMPG
|
||
|
|
|
||
|
|
MPGEnd0:
|