Advanced Search

sirkris

5 Reputation

One Badge

4 years, 290 days
Contact sirkris

MaplePrimes Activity


These are questions asked by sirkris

CodeGeneration in C throws Error: Error,...

sirkris 5 Maple 2019
codegeneration
April 08 2020
0 0

Hi, can someone explain me why CodeGeneration in C throws the error : Error, (in Print) improper op or subscript selector ? Interestingly, CodeGeneration in Python or Matlab works. Source code is attached. Thanks in advance!

membrane_energy.mw
 

 

Membrane Energy

 

restart; with(VectorCalculus); with(LinearAlgebra); with(CodeGeneration)

Define variables

 

Vectors for vertices of current position

v1 := Vector(3, symbol = v1_i) = Vector[column]([[v1_i[1]], [v1_i[2]], [v1_i[3]]], ["x", "y", "z"]) 

v2 := Vector(3, symbol = v2_i) = Vector[column]([[v2_i[1]], [v2_i[2]], [v2_i[3]]], ["x", "y", "z"])NULL

v3 := Vector(3, symbol = v3_i) = Vector[column]([[v3_i[1]], [v3_i[2]], [v3_i[3]]], ["x", "y", "z"])NULL

n := `&x`(v2-v1, v3-v1)

v4 := v1+n/norm(n, 2)^.5

Vector for vertices of next position

v1n := Vector(3, symbol = v1n_i) = Vector[column]([[v1n_i[1]], [v1n_i[2]], [v1n_i[3]]], ["x", "y", "z"])NULL

v2n := Vector(3, symbol = v2n_i) = Vector[column]([[v2n_i[1]], [v2n_i[2]], [v2n_i[3]]], ["x", "y", "z"])NULL

v3n := Vector(3, symbol = v3n_i) = Vector[column]([[v3n_i[1]], [v3n_i[2]], [v3n_i[3]]], ["x", "y", "z"])NULL

nn := `&x`(v2n-v1n, v3n-v1n)

v4n := v1n+nn/norm(nn, 2)^.5``

``

 

Define Transformation

 

V := LinearAlgebra:-Transpose(Matrix([v2-v1, v3-v1, v4-v1]))

Dimension(V) = 3, 3 

Vn := LinearAlgebra:-Transpose(Matrix([v2n-v1n, v3n-v1n, v4n-v1n]))

Dimension(Vn) = 3, 3NULL

Note we have Vn = T*V and if the current triangle is not degenerate, then T = Vn/V. As we can pre-compute 1/V we define a new matrix for it:

Vinv := Matrix(3, 3, symbol = Vinv_ij) =

Matrix(%id = 18446746713267339006)

(1.2.1)

T := MatrixMatrixMultiply(Vn, Vinv)

Dimension(T) = 3, 3``

``

``

Define Energy

 

E := Trace(MatrixMatrixMultiply(T, LinearAlgebra:-Transpose(T)))``

``

``

Gradient and Hessian

 

gradE := Gradient(E, [v1n[1], v1n[2], v1n[3], v2n[1], v2n[2], v2n[3], v3n[1], v3n[2], v3n[3]])

Dimension(gradE) = 9NULL

CodeGeneration[C](gradE, defaulttype = numeric, optimize = tryhard, functionprecision = double, precision = double, deducetypes = false, resultname = 'gradE')t1 = v2n_i[0] - v1n_i[0];
t2 = v2n_i[1] - v1n_i[1];
t3 = v2n_i[2] - v1n_i[2];
t4 = t1 * Vinv_ij[0][0] + t2 * Vinv_ij[1][0] + t3 * Vinv_ij[2][0];
t5 = t1 * Vinv_ij[0][1] + t2 * Vinv_ij[1][1] + t3 * Vinv_ij[2][1];
t6 = t1 * Vinv_ij[0][2] + t2 * Vinv_ij[1][2] + t3 * Vinv_ij[2][2];
t7 = v3n_i[0] - v1n_i[0];
t8 = v3n_i[1] - v1n_i[1];
t9 = v3n_i[2] - v1n_i[2];
t10 = t7 * Vinv_ij[0][0] + t8 * Vinv_ij[1][0] + t9 * Vinv_ij[2][0];
t11 = t7 * Vinv_ij[0][1] + t8 * Vinv_ij[1][1] + t9 * Vinv_ij[2][1];
t12 = t7 * Vinv_ij[0][2] + t8 * Vinv_ij[1][2] + t9 * Vinv_ij[2][2];
t13 = t2 * t9 - t3 * t8;
t14 = fabs(t13);
t15 = t1 * t9 - t3 * t7;
t16 = fabs(t15);
t17 = t1 * t8 - t2 * t7;
t18 = fabs(t17);
t19 = pow(t14, 0.2e1) + pow(t16, 0.2e1) + pow(t18, 0.2e1);
t20 = pow(t19, -0.5e1 / 0.4e1);
t19 = t19 * t20;
t21 = Vinv_ij[0][0] * t13;
t22 = Vinv_ij[1][0] * t15;
t23 = Vinv_ij[2][0] * t17;
t24 = (t23 + t21 - t22) * t19;
t25 = -v2n_i[2] + v3n_i[2];
t26 = fabs(t15) / t15;
t27 = -v2n_i[1] + v3n_i[1];
t28 = fabs(t17) / t17;
t21 = t23 + t21 - t22;
t22 = (t16 * t25 * t26 + t18 * t27 * t28) * t20;
t23 = 0.1e1 / 0.2e1;
t29 = Vinv_ij[0][1] * t13;
t30 = Vinv_ij[1][1] * t15;
t31 = Vinv_ij[2][1] * t17;
t32 = (t31 + t29 - t30) * t19;
t29 = t31 + t29 - t30;
t30 = Vinv_ij[0][2] * t13;
t15 = Vinv_ij[1][2] * t15;
t17 = Vinv_ij[2][2] * t17;
t31 = (t17 + t30 - t15) * t19;
t15 = t17 + t30 - t15;
t17 = t6 + t12;
t30 = t5 + t11;
t33 = t4 + t10;
t13 = fabs(t13) / t13;
t34 = -v2n_i[0] + v3n_i[0];
t35 = (t14 * t25 * t13 - t18 * t34 * t28) * t20;
t35 = t17 * Vinv_ij[1][2] - t24 * (t23 * t35 * t21 - t19 * (t25 * Vinv_ij[0][0] - t34 * Vinv_ij[2][0])) + t30 * Vinv_ij[1][1] - t31 * (t23 * t35 * t15 - t19 * (t25 * Vinv_ij[0][2] - t34 * Vinv_ij[2][2])) - t32 * (t23 * t35 * t29 - t19 * (t25 * Vinv_ij[0][1] - t34 * Vinv_ij[2][1])) + t33 * Vinv_ij[1][0];
t36 = (t14 * t27 * t13 + t16 * t34 * t26) * t20;
t34 = Vinv_ij[2][2] * t17 - t24 * (-t23 * t36 * t21 + t19 * (t27 * Vinv_ij[0][0] - t34 * Vinv_ij[1][0])) + t30 * Vinv_ij[2][1] - t31 * (-t23 * t36 * t15 + t19 * (t27 * Vinv_ij[0][2] - t34 * Vinv_ij[1][2])) - t32 * (-t23 * t36 * t29 + t19 * (t27 * Vinv_ij[0][1] - t34 * Vinv_ij[1][1])) + t33 * Vinv_ij[2][0];
t36 = (t16 * t9 * t26 + t18 * t8 * t28) * t20;
t36 = t24 * (-t23 * t36 * t21 - t19 * (-t8 * Vinv_ij[2][0] + t9 * Vinv_ij[1][0])) + t31 * (-t23 * t36 * t15 - t19 * (-t8 * Vinv_ij[2][2] + t9 * Vinv_ij[1][2])) + t32 * (-t23 * t36 * t29 - t19 * (-t8 * Vinv_ij[2][1] + t9 * Vinv_ij[1][1])) + t4 * Vinv_ij[0][0] + t5 * Vinv_ij[0][1] + t6 * Vinv_ij[0][2];
t37 = (t14 * t9 * t13 - t18 * t7 * t28) * t20;
t9 = t24 * (-t23 * t37 * t21 + t19 * (-t7 * Vinv_ij[2][0] + t9 * Vinv_ij[0][0])) + t31 * (-t15 * t23 * t37 + t19 * (-t7 * Vinv_ij[2][2] + t9 * Vinv_ij[0][2])) + t32 * (-t23 * t29 * t37 + t19 * (-t7 * Vinv_ij[2][1] + t9 * Vinv_ij[0][1])) + t4 * Vinv_ij[1][0] + t5 * Vinv_ij[1][1] + t6 * Vinv_ij[1][2];
t37 = (t14 * t8 * t13 + t16 * t7 * t26) * t20;
t4 = t24 * (t23 * t37 * t21 - t19 * (-t7 * Vinv_ij[1][0] + t8 * Vinv_ij[0][0])) + t31 * (t15 * t23 * t37 - t19 * (-t7 * Vinv_ij[1][2] + t8 * Vinv_ij[0][2])) + t32 * (t23 * t29 * t37 - t19 * (-t7 * Vinv_ij[1][1] + t8 * Vinv_ij[0][1])) + t4 * Vinv_ij[2][0] + t5 * Vinv_ij[2][1] + t6 * Vinv_ij[2][2];
t5 = (t16 * t3 * t26 + t18 * t2 * t28) * t20;
t5 = t10 * Vinv_ij[0][0] + t11 * Vinv_ij[0][1] + t12 * Vinv_ij[0][2] + t24 * (t21 * t23 * t5 + t19 * (-t2 * Vinv_ij[2][0] + t3 * Vinv_ij[1][0])) + t31 * (t15 * t23 * t5 + t19 * (-t2 * Vinv_ij[2][2] + t3 * Vinv_ij[1][2])) + t32 * (t23 * t29 * t5 + t19 * (-t2 * Vinv_ij[2][1] + t3 * Vinv_ij[1][1]));
t6 = (-t18 * t1 * t28 + t14 * t3 * t13) * t20;
t3 = t10 * Vinv_ij[1][0] + t11 * Vinv_ij[1][1] + t12 * Vinv_ij[1][2] + t24 * (t21 * t23 * t6 - t19 * (-t1 * Vinv_ij[2][0] + t3 * Vinv_ij[0][0])) + t31 * (t15 * t23 * t6 - t19 * (-t1 * Vinv_ij[2][2] + t3 * Vinv_ij[0][2])) + t32 * (t23 * t29 * t6 - t19 * (-t1 * Vinv_ij[2][1] + t3 * Vinv_ij[0][1]));
t6 = (t16 * t1 * t26 + t14 * t2 * t13) * t20;
t1 = t10 * Vinv_ij[2][0] + t11 * Vinv_ij[2][1] + t12 * Vinv_ij[2][2] + t24 * (-t21 * t23 * t6 + t19 * (-t1 * Vinv_ij[1][0] + t2 * Vinv_ij[0][0])) + t31 * (-t15 * t23 * t6 + t19 * (-t1 * Vinv_ij[1][2] + t2 * Vinv_ij[0][2])) + t32 * (-t23 * t29 * t6 + t19 * (-t1 * Vinv_ij[1][1] + t2 * Vinv_ij[0][1]));
t2 = 0.2e1;
gradE[0] = -t2 * (-t24 * (t23 * t22 * t21 + t19 * (t25 * Vinv_ij[1][0] - t27 * Vinv_ij[2][0])) + t33 * Vinv_ij[0][0] + t30 * Vinv_ij[0][1] + t17 * Vinv_ij[0][2] - t32 * (t23 * t22 * t29 + t19 * (t25 * Vinv_ij[1][1] - t27 * Vinv_ij[2][1])) - t31 * (t23 * t22 * t15 + t19 * (t25 * Vinv_ij[1][2] - t27 * Vinv_ij[2][2])));
gradE[1] = -t2 * t35;
gradE[2] = -t2 * t34;
gradE[3] = t2 * t36;
gradE[4] = t2 * t9;
gradE[5] = t2 * t4;
gradE[6] = t2 * t5;
gradE[7] = t2 * t3;
gradE[8] = t2 * t1;

``

``

``

``

``

Hessian

 

``

hessE := Hessian(E, [v1n[1], v1n[2], v1n[3], v2n[1], v2n[2], v2n[3], v3n[1], v3n[2], v3n[3]])

Dimension(hessE)

9, 9

(1.5.1)

 

CodeGeneration[C](hessE, optimize = tryhard, deducetypes = false, resultname = 'hessE')

Error, (in Print) improper op or subscript selector

 

``

````

``

``

``

``


 

Download membrane_energy.mw

 

 

 

Page 1 of 1