transpose command in hercmExplorer causes matrix to be displayed incorrectly

[charlesdaniels]

welcome to HeRCM Explorer. Enter "help" for help
> load matricies/bcspwr01.mtx
WARNING: matrix format not specified, assuming mtx 
from filename
done reading matrix
INFO: matrix is symmetric, bottom triangle should be only zeros
> display
INFO: matrix is symmetric, bottom triangle should be only zeros
[[ 1.  1.  0. ...,  0.  0.  1.]
 [ 0.  1.  1. ...,  0.  0.  0.]
 [ 0.  0.  1. ...,  0.  0.  0.]
 ..., 
 [ 0.  0.  0. ...,  1.  0.  0.]
 [ 0.  0.  0. ...,  0.  1.  0.]
 [ 0.  0.  0. ...,  0.  0.  1.]]
> transpose
performing matrix tranpose, please wait...
matrix transpose complete
> display
INFO: matrix is symmetric, bottom triangle should be only zeros
[[ 1.  0.  0. ...,  0.  0.  0.]
 [ 0.  2.  0. ...,  0.  0.  0.]
 [ 0.  0.  2. ...,  0.  0.  0.]
 ..., 
 [ 0.  0.  0. ...,  2.  0.  0.]
 [ 0.  0.  0. ...,  0.  2.  0.]
 [ 0.  0.  0. ...,  0.  0.  3.]]

> load matricies/sample.mtx
WARNING: matrix format not specified, assuming mtx 
from filename
done reading matrix
> display
[[ 0.  0.  0.  0.  0.]
 [ 0.  4.  0.  0.  2.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  0.  0.  5.  7.]
 [ 0.  6.  3.  0.  8.]]
> transpose
performing matrix tranpose, please wait...
matrix transpose complete
> display
[[  0.   0.   0.   0.   0.]
 [  0.  10.   0.   0.   0.]
 [  0.   0.   4.   0.   0.]
 [  0.   0.   0.   5.   0.]
 [  0.   0.   0.   0.  17.]]
> 

[charlesdaniels]

Continuation of the above....

> load matricies/sample.mtx
WARNING: matrix format not specified, assuming mtx 
from filename
done reading matrix
> display
[[ 0.  0.  0.  0.  0.]
 [ 0.  4.  0.  0.  2.]
 [ 0.  0.  1.  0.  0.]
 [ 0.  0.  0.  5.  7.]
 [ 0.  6.  3.  0.  8.]]
> transpose
performing matrix tranpose, please wait...
matrix transpose complete
> display
[[  0.   0.   0.   0.   0.]
 [  0.  10.   0.   0.   0.]
 [  0.   0.   4.   0.   0.]
 [  0.   0.   0.   5.   0.]
 [  0.   0.   0.   0.  17.]]
> range 0 0 2 2
0,0 = 0
0,1 = 0
0,2 = 0
1,0 = 0
1,1 = 4.0
1,2 = 0
2,0 = 0
2,1 = 0
2,2 = 1.0

This does not appear to affect the matrix in-memory, only the display command.

NOTE: I cannot reproduce this issue on x86, it appears to affect ARM

[charlesdaniels]

Tried opening a file which had been transposed on ARM froma n x86 machine and the issue persisted. I suspect there is either something wrong with the hercm write code that is only coming out on ARM, or numpy/scipy are somehow broken subtly on ARM.

[charlesdaniels]

This may have been resolved in 50e7c6c, as an "in-house" matrix transpose function has now been baked into libHercMatrix. This issue will remain open until the new implementation is tested.