Revision history [back]

Please refer to my question at https://ask.sagemath.org/question/63878/making-code-faster-by-splitting-and-use-of-supercomputer/

According to the suggestion given in the end by Max Alekseyev I have made a complete code. Here I am showing only the initial part , as problem lies in it . I don't know what I am missing or what mistake I have made.

M = matrix(3,3, [1, 1, 1, -1, 1, 1, -1, -1, 1])
List_S=[M^0,M^1,M^2]
n = 3
V = GF(2)^(n^2)
VL1=V.list()[1:]
VL2=VL1[:-1]
P=[]
for v in VL2:    
        P.append(matrix(ZZ,n,n,v))
N=[]
for p in P:
       k=-p
       N.append(k)
U = []
for a in P:
       U.append(a)
for b in N:
      U.append(b)         
show("U=",len(U))
def mat_to_vector(m):
      temp = []
      for i in range(m.ncols()):
               for j in range(m.nrows()):
                      temp.append(m[i][j])
      return(vector(temp))
CAND=[]
for i in range(len(U)):
       List_U=[U[i]]
       List_SU=List_S+List_U
       W=U[:]                 
       W.remove(W[i])         
       T=tuple(List_SU)
       X = ZZ^(n^2)
       Y = X.span(mat_to_vector(T[i]) for i in range(len(T))) 
       for w in W:
              if mat_to_vector(w)  in Y:
                      CAND.append(mat_to_vector(w)) 
       VTM=[matrix(ZZ,n,n,u) for u in CAND]
       for b in VTM:
             b.set_immutable()
       B = set(VTM)
show("B=",len(B))

According to the suggestion, len(B) should come out to be less than len(U). But it is coming out to be same as len(U). Kindly find the error.

Please refer to my question at https://ask.sagemath.org/question/63878/making-code-faster-by-splitting-and-use-of-supercomputer/

According to the suggestion given in the end by Max Alekseyev I have made a complete code. Here I am showing only the initial part , as problem lies in it . I don't know what I am missing or what mistake I have made.

M = matrix(3,3, [1, 1, 1, -1, 1, 1, -1, -1, 1])
List_S=[M^0,M^1,M^2]
n = 3
V = GF(2)^(n^2)
VL1=V.list()[1:]
VL2=VL1[:-1]
VL1=V.list()[1:]    
# Removing the all-zero list because not needed for our purpose
VL2=VL1[:-1]      
#Removing the all-ones list because not needed for our purpose
P=[]
for v in VL2:    
        P.append(matrix(ZZ,n,n,v))
P.append(matrix(ZZ,n,n,v))  
 #Forming n by n matrices from  list in  VL2. Contains all 0,1 matrices.                                               
 N=[]
for p in P:
       k=-p
       N.append(k)
 #Forming n by n matrices from P by taking their negative. Contains all 0,-1 matrices 
U = []
for a in P:
       U.append(a)
for b in N:
      U.append(b)     
#Taking union of P and N   
show("U=",len(U))
def mat_to_vector(m):
      temp = []
      for i in range(m.ncols()):
               for j in range(m.nrows()):
                      temp.append(m[i][j])
      return(vector(temp))
  #Defining a function which converts matrix to vector
CAND=[]
for i in range(len(U)):
       List_U=[U[i]]
       List_SU=List_S+List_U
       W=U[:]                 
       W.remove(W[i])         
       T=tuple(List_SU)
       X = ZZ^(n^2)
       Y = X.span(mat_to_vector(T[i]) for i in range(len(T))) 
       for w in W:
              if mat_to_vector(w)  in Y:
                      CAND.append(mat_to_vector(w)) 
       VTM=[matrix(ZZ,n,n,u) for u in CAND]
       for b in VTM:
             b.set_immutable()
       B = set(VTM)
show("B=",len(B))

According to the suggestion, len(B) should come out to be less than len(U). But it is coming out to be same as len(U). Kindly find the error.