Use MulAdd's CZeros to remove duplicate zero-fill

[jishnub]

With this, multiplying a BandedMatrix by a OneElementVector would be as performant as slicing.

julia> B = brand(6000,6000,4000,4000); O = OneElement(1, 3, size(B,2));

julia> @btime $B * $O;
  3.826 μs (2 allocations: 46.92 KiB)

julia> @btime $B[:,$O.ind...];
  4.013 μs (4 allocations: 47.00 KiB)

[codecov[bot]]

Codecov Report

Attention: 2 lines in your changes are missing coverage. Please review.

Comparison is base (ff75189) 89.55% compared to head (213a970) 89.59%. Report is 2 commits behind head on master.

Files	Patch %	Lines
src/banded/gbmm.jl	89.47%	2 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #424 +/- ## ========================================== + Coverage 89.55% 89.59% +0.03% ========================================== Files 25 25 Lines 3544 3557 +13 ========================================== + Hits 3174 3187 +13 Misses 370 370 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.