#codingexercise
for the  problem duscussed yesterday, we see a few more alternatives

parallel processing
sum along all columns
sum along all rows
if any row or column is all zero
split the matrix on either side of the zeros
compute the submatrix as usual

another method may be to merge two sub matrices along a front
the histogram on both sides of the front are computed starting from the front and progressing outwards
for contiguous non-zero paired entries on the histograms on both sides, the max area is calculated for each contiguous zone out of which the largest is chosen

another method for a merge like above would be to find contiguous zones of ones along the front
for each of the contiguous zone, and for each of the ones in the zone find the max area on both sides noting the common width.
retirn the max of all such combined max.