fix: Remove castting on decimals with a small precision to decimal256

[kazuyukitanimura]

Which issue does this PR close?

Part of #670

Rationale for this change

This PR improves the native execution performance on decimals with a small precision

What changes are included in this PR?

This PR changes not to promote decimal128 to decimal256 if the precisions are small enough

How are these changes tested?

Existing tests

[kazuyukitanimura]

OpenJDK 64-Bit Server VM 17.0.11+9-LTS on Mac OS X 14.5
Apple M1 Max
TPCDS Micro Benchmarks:                   Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
add_many_decimals                                 10883          11268         365         26.5          37.8       1.0X
add_many_decimals: Comet (Scan)                   13924          14097         309         20.7          48.3       0.8X
add_many_decimals: Comet (Scan, Exec)             12476          12570         111         23.1          43.3       0.9X

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 53.75%. Comparing base (ded3dd6) to head (a14f888).
Report is 8 commits behind head on main.

Additional details and impacted files

@@              Coverage Diff              @@
##               main     #741       +/-   ##
=============================================
+ Coverage     33.57%   53.75%   +20.17%     
+ Complexity      830      813       -17     
=============================================
  Files           110      107        -3     
  Lines         42608    10273    -32335     
  Branches       9352     1934     -7418     
=============================================
- Hits          14306     5522     -8784     
+ Misses        25347     3774    -21573     
+ Partials       2955      977     -1978     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[andygrove]

I created a criterion benchmark for decimal_div in #743.

The results look great for the changes in this PR.

decimal_div             time:   [5.3281 µs 5.3370 µs 5.3485 µs]
                        change: [-94.767% -94.727% -94.685%] (p = 0.00 < 0.05)
                        Performance has improved.

[kazuyukitanimura]

@comphead @huaxingao @parthchandra @viirya

[andygrove]

I'm not an expert on the decimal math in here, but all tests are passing, and performance is improved, so LGTM

[kazuyukitanimura]

Thank you @andygrove merged