Fix exploding values

[roeezis]

Fix non convergence of the local smoothing filter iterations, as sdot(d,q) would be Infinity for large numbers, eventually resulting in NaN rnorm.
the fix changes float to 64bit double precision for scalars, sdot and saxpy calculations.

[codecov]

Codecov Report

Merging #18 (d729f43) into master (d8ec668) will decrease coverage by 0.00%.
The diff coverage is 29.26%.

@@             Coverage Diff              @@
##             master      #18      +/-   ##
============================================
- Coverage     40.85%   40.85%   -0.01%     
- Complexity     7221     7222       +1     
============================================
  Files           264      264              
  Lines         63167    63167              
  Branches      10299    10299              
============================================
- Hits          25809    25806       -3     
- Misses        35659    35663       +4     
+ Partials       1699     1698       -1     

Impacted Files	Coverage Δ	Complexity Δ
...n/java/edu/mines/jtk/dsp/LocalSmoothingFilter.java	25.17% <29.26%> (ø)	19.00 <0.00> (ø)
core/src/main/java/edu/mines/jtk/mesh/TriMesh.java	60.83% <0.00%> (-0.23%)	244.00% <0.00%> (ø%)
core/src/main/java/edu/mines/jtk/mesh/TetMesh.java	51.43% <0.00%> (ø)	388.00% <0.00%> (ø%)
...n/java/edu/mines/jtk/interp/CubicInterpolator.java	72.41% <0.00%> (+0.43%)	51.00% <0.00%> (+1.00%)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d8ec668...d729f43. Read the comment docs.