cube_rotate.vs.asm: move all sin/cos values to temp[3]

This commit is contained in:
Zack Buhman 2025-10-23 19:52:38 -05:00
parent 8594bc4a38
commit 7ac88f0ed3
2 changed files with 44 additions and 44 deletions

View File

@ -1,65 +1,65 @@
# CONST[0] = {0.159155, 0.5, 6.283185, -3.141593} -- CONST[0] = {0.159155, 0.5, 6.283185, -3.141593}
# CONST[1] = {theta1, theta2, 0.2, 0.5} -- CONST[1] = {theta1, theta2, 0.2, 0.5}
# each instruction is only allowed to use a single unique `const` address -- each instruction is only allowed to use a single unique `const` address
# --
# instructions may use multiple `temp` addresses, so const[1] is moved to -- instructions may use multiple `temp` addresses, so const[1] is moved to
# temp[0]: -- temp[0]:
# --
temp[0].xy = VE_ADD const[1].xy__ const[1].00__ ; temp[0].xy = VE_ADD const[1].xy__ const[1].00__ ;
# ME_SIN and ME_COS are clamped to the range -π to +π prior to the sin/cos -- ME_SIN and ME_COS clamp their inputs to [-π,+π] prior to the sin/cos
# calculation. -- calculation.
# --
# This 3-instruction sequence linearly remaps the range [-,+] to [-π,+π] -- This 3-instruction sequence remaps the range [-,+] to [-π,+π]
temp[0].xy = VE_MAD temp[0].xy__ const[0].xx__ const[0].yy__ ; temp[0].xy = VE_MAD temp[0].xy__ const[0].xx__ const[0].yy__ ;
temp[0].xy = VE_FRC temp[0].xy__ ; temp[0].xy = VE_FRC temp[0].xy__ ;
temp[0].xy = VE_MAD temp[0].xy__ const[0].zz__ const[0].ww__ ; temp[0].xy = VE_MAD temp[0].xy__ const[0].zz__ const[0].ww__ ;
# sin and cos -- sin and cos
temp[3].x = ME_SIN temp[0].___x ; temp[3].x = ME_SIN temp[0].___x ;
temp[3].y = ME_COS temp[0].___x ; temp[3].y = ME_COS temp[0].___x ;
temp[4].x = ME_SIN temp[0].___y ; temp[3].z = ME_SIN temp[0].___y ;
temp[4].y = ME_COS temp[0].___y ; temp[3].w = ME_COS temp[0].___y ;
# temp[3] now contains: -- temp[3] now contains:
# temp[3] = {sin(theta1), cos(theta1), sin(theta2), cos(theta2)} -- temp[3] = {sin(theta1), cos(theta1), sin(theta2), cos(theta2)}
######################################################################### -------------------------------------------------------------------------
# first rotation: X-axis rotation: -- first rotation: X-axis rotation:
######################################################################### -------------------------------------------------------------------------
# y_ = (-z0 * st1) -- y_ = (-z0 * st1)
# z_ = ( z0 * ct1) -- z_ = ( z0 * ct1)
temp[1].yz = VE_MUL input[0]._-zz_ temp[3]._xy_ ; temp[1].yz = VE_MUL input[0]._-zz_ temp[3]._xy_ ;
# x1 = (x0 * 1 + 0) -- x1 = (x0 * 1 + 0)
# y1 = (y0 * ct1 + nz0st1) -- y1 = (y0 * ct1 + nz0st1)
# z1 = (y0 * st1 + z0ct1) -- z1 = (y0 * st1 + z0ct1)
temp[1].xyz = VE_MAD input[0].xyy_ temp[3].1yx_ temp[1].0yz_ ; temp[1].xyz = VE_MAD input[0].xyy_ temp[3].1yx_ temp[1].0yz_ ;
######################################################################### -------------------------------------------------------------------------
# second rotation: Y-axis rotation: -- second rotation: Y-axis rotation:
######################################################################### -------------------------------------------------------------------------
# x_ = (-z1 * st2) -- x_ = (-z1 * st2)
# z_ = ( z1 * ct2) -- z_ = ( z1 * ct2)
temp[2].xz = VE_MUL temp[1].-z_z_ temp[4].x_y_ ; temp[2].xz = VE_MUL temp[1].-z_z_ temp[3].z_w_ ;
# x2 = (x1 * ct2 + nz1st2) -- x2 = (x1 * ct2 + nz1st2)
# y2 = (y1 * 1 + 0) -- y2 = (y1 * 1 + 0)
# z2 = (x1 * st2 + z1ct2) -- z2 = (x1 * st2 + z1ct2)
temp[2].xyz = VE_MAD temp[1].xyx_ temp[4].y1x_ temp[2].x0z_ ; temp[2].xyz = VE_MAD temp[1].xyx_ temp[3].w1z_ temp[2].x0z_ ;
######################################################################### -------------------------------------------------------------------------
# scale -- scale
######################################################################### -------------------------------------------------------------------------
temp[3].xyz = VE_MAD temp[2].xyz_ const[1].zzz_ const[1].00w_ ; temp[3].xyz = VE_MAD temp[2].xyz_ const[1].zzz_ const[1].00w_ ;
######################################################################### -------------------------------------------------------------------------
# output -- output
######################################################################### -------------------------------------------------------------------------
out[0].xyzw = VE_MUL temp[3].xyzz temp[3].11-z1 ; out[0].xyzw = VE_MUL temp[3].xyzz temp[3].11-z1 ;
out[1].xyzw = VE_ADD input[1].xyzw input[1].0000 ; out[1].xyzw = VE_ADD input[1].xyzw input[1].0000 ;

View File

@ -4,12 +4,12 @@
0x00300004, 0x01f90000, 0x01fa4002, 0x01fb6002, 0x00300004, 0x01f90000, 0x01fa4002, 0x01fb6002,
0x00106050, 0x003fe000, 0x01ffe000, 0x01ffe000, 0x00106050, 0x003fe000, 0x01ffe000, 0x01ffe000,
0x00206051, 0x003fe000, 0x01ffe000, 0x01ffe000, 0x00206051, 0x003fe000, 0x01ffe000, 0x01ffe000,
0x00108050, 0x007fe000, 0x01ffe000, 0x01ffe000, 0x00406050, 0x007fe000, 0x01ffe000, 0x01ffe000,
0x00208051, 0x007fe000, 0x01ffe000, 0x01ffe000, 0x00806051, 0x007fe000, 0x01ffe000, 0x01ffe000,
0x00602002, 0x05d2e001, 0x01c8e060, 0x01ffe060, 0x00602002, 0x05d2e001, 0x01c8e060, 0x01ffe060,
0x00702004, 0x01c90001, 0x01c1a060, 0x01d18020, 0x00702004, 0x01c90001, 0x01c1a060, 0x01d18020,
0x00504002, 0x03d74020, 0x01cf0080, 0x01ffe080, 0x00504002, 0x03d74020, 0x01df4060, 0x01ffe060,
0x00704080, 0x01c10020, 0x01c52080, 0x01d40040, 0x00704080, 0x01c10020, 0x01d56060, 0x01d40040,
0x00706004, 0x01d10040, 0x01d24022, 0x01dc8022, 0x00706004, 0x01d10040, 0x01d24022, 0x01dc8022,
0x00f00202, 0x00910060, 0x0955a060, 0x01ffe060, 0x00f00202, 0x00910060, 0x0955a060, 0x01ffe060,
0x00f02203, 0x00d10021, 0x01248021, 0x01ffe021, 0x00f02203, 0x00d10021, 0x01248021, 0x01ffe021,