[ hdot ] Use precision-enhanced hdot

[skykongkong8]

hdot : vec(1 x K ) x vec(K X 1 ) -> 1 (scalar)

• Previous hdot was using full-fp16.
• Since this is also one of dimension-shrinking computations, should use inter-fp32 values to enhance precision.
• This has not been detected due to small dimension Tensor usage in unittest. Add bigger-sized test case accordingly.

A. Accuracy w.r.t cblas-f32 (mse)

dim	hdot-f16 (previous)	hdot-f16f32 (now)
1024	0.250061	3.72529e-09
2048	2.07308	0.00362171
4096	11.8475	0.195379

B. Latency

Since this is nanosec-unit result, almost trivial

dim	hdot-f16	hdot-f16f32	cblas-f32
1024	261 ns	280 ns	347 ns
2048	328 ns	359 ns	1473 ns
4096	508 ns	562 ns	909 ns

Conclusion : Better accuracy with almost no latency deterioration

Self evaluation:

1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped

[taos-ci]

:memo: TAOS-CI Version: 1.5.20200925. Thank you for submitting PR #2555. Please a submit 1commit/1PR (one commit per one PR) policy to get comments quickly from reviewers. Your PR must pass all verificiation processes of cibot before starting a review process from reviewers. If you are new member to join this project, please read manuals in documentation folder and wiki page. In order to monitor a progress status of your PR in more detail, visit http://ci.nnstreamer.ai/.