home *** CD-ROM | disk | FTP | other *** search
Text File | 1995-12-21 | 3.7 KB | 136 lines | [TEXT/CWIE] |
-
-
- -------------------------------------------------------------------------------
- Verify Operations on Matrices
-
-
- ---> Test allocation and compatibility check
- The following matrices have been allocated
-
- Matrix 1:4x1:20 'Matrix m1'
- Matrix 1:4x1:20 'Matrix m2'
- Matrix 1:4x0:19 'Matrix m3'
- Matrix 1:4x1:20 'Matrix m4'
- Status information reported for matrix m3:
- Row lower bound ... 1
- Row upper bound ... 4
- Col lower bound ... 0
- Col upper bound ... 19
- No. rows ...........4
- No. cols ...........20
- No. of elements ....80
- Name Matrix m3
-
- Check matrices 1 & 2 for compatibility
- Check matrices 1 & 4 for compatibility
- m2 has to be compatible with m3 after resizing to m3
- m1 has to be compatible with m5 after resizing to m5
-
- Matrix 1:5x1:25 'Matrix m5'
-
- ---> Test operations that treat each element uniformly
- Writing zeros to m...
- Creating zero m1 ...
- Comparing m1 with 0 ...
- Writing a pattern 8.625 by assigning to m(i,j)...
- Writing the pattern by assigning to m1 as a whole ...
- Comparing m and m1 ...
- Comparing (m=0) and m1 ...
- Clearing m1 ...
-
- Clear m and add the pattern
- add the doubled pattern with the negative sign
- subtract the trippled pattern with the negative sign
-
- Verify comparison operations when all elems are the same
-
- Verify comparison operations when not all elems are the same
-
- Assign 2*pattern to m by repeating additions
- Assign 2*pattern to m1 by multiplying by two
- Multiply m1 by one half returning it to the 1*pattern
-
- Assign -pattern to m and m1
- m = sqrt(sqr(m)); m1 = abs(m1); Now m and m1 have to be the same
-
- Check out to see that sin^2(x) + cos^2(x) = 1
- Element (16,8) with value 1 differs the most from what
- was expected, 1, though the deviation 1.19209e-07 is small
-
- Done
-
-
- ---> Test Binary Matrix element-by-element operations
-
- Verify assignment of a matrix to the matrix
-
- Adding the matrix to itself, uniform pattern 4.25
- subtracting two matrices ...
- subtracting the matrix from itself
- adding two matrices together
-
- Arithmetic operations on matrices with not the same elements
- adding mp to the zero matrix...
- making m = 3*mp and m1 = 3*mp, via add() and succesive mult
- clear both m and m1, by subtracting from itself and via add()
-
- Testing element-by-element multiplications and divisions
- squaring each element with sqr() and via multiplication
- compare (m = pattern^2)/pattern with pattern
-
-
- Comparison of two Matrices:
- Original vector and vector after squaring and dividing
- Matrix 2:11x0:19 ''
- Matrix 2:11x0:19 ''
- Maximal discrepancy 0
- occured at the point (2,0)
- Matrix 1 element is 4.25
- Matrix 2 element is 4.25
- Absolute error v2[i]-v1[i] 0
- Relative error 0
-
- ||Matrix 1|| 850
- ||Matrix 2|| 850
- ||Matrix1-Matrix2|| 0
- ||Matrix1-Matrix2||/sqrt(||Matrix1|| ||Matrix2||) 0
-
-
- Done
-
- ---> Verify norm calculations
-
- Assign 10.25 to all the elements and check norms
- 1. (col) norm should be pattern*nrows
- Inf (row) norm should be pattern*ncols
- Square of the Eucl norm has got to be pattern^2 * no_elems
-
- Done
-
- ---> Check creating some special matrices of dimension 20
-
- test creating Hilbert matrices
- test creating zero matrix and copy constructor
- test creating unit matrices
- check to see that Haar matrix has *exactly* orthogonal columns
- make Haar (sub)matrix and test it *is* a submatrix
-
- Done
-
- ---> Check making/forcing promises, (lazy)matrices of dimension 20
-
- make a promise and force it by a constructor
- make a promise and force it by an assignment
-
- Done
-
- ---> Verify matrix transpose
- for matrices of a characteristic size 20
-
- Check to see that a square UnitMatrix' stays the same
- Test a non-square UnitMatrix
- Check to see that a symmetric (Hilbert)Matrix' stays the same
- Check transposing a non-symmetric matrix
- Check double transposing a non-symmetric matrix
- Done
-