Here is a more advanced answer. It assumes familiarity with how to use the list-oriented nature of Perl, how to use map, and how to access arrays of arrays as described in perlref.

I would do the rearranging of data in a function. Suppose that you can create an array @cells of cells above. I assume that you want it sorted by column and then row. Here is a possible rearrangement functions you could use:

# Takes a number of columns then a list.  Reformats the list into
# an array of arrays, with the original elements sorted down by
# columns, and any missing items missing off of the last column
sub reformat_major_minor {
  my $dim = shift;
  my $row_size = int((@_+$dim - 1)/$dim);
  my @ret;
  foreach my $i (0..$#_) {
    $ret[$i%$row_size][$i/$row_size] = $_[$i];
  }
  return @ret;
}
[download]

And then you can produce your output with:

foreach my $row (reformat_major_minor(3, @cells)) {
  print "<tr>\n", (map "  <td>$_</td>\n", @$row), "</tr>\n";
}
[download]

Should you dislike taking the missing elements on the last column and want to take it on the last row instead, you can use the following reformat function instead:

# Takes a number of columns then a list.  Reformats the list into
# an array of arrays, with the original elements sorted down by
# columns, and any missing items missing off of the last column.
sub reformat_major_minor {
  my $dim = shift;
  my $row_size = int(@_/$dim);
  my @ret;
  push @ret, [splice @_, 0, $row_size] while @_;
  return swap_dims(@ret);
}

# Takes an array of arrays and swaps rows for columns.
sub swap_dims {
  my @ret;
  foreach my $row_ind (0..$#_) {
    my $row = $_[$row_ind];
    foreach my $col_ind (0..$#$row) {
      $ret[$col_ind][$row_ind] = $row->[$col_ind];
    }
  }
  return @ret;
}
[download]

UPDATE
Oops. I put the wrong reformatting function up there in the first slot. Now fixed.