Use sin/cos instead of sind/cosd in DE_2000 calculation

[kimikage]

This is one of the preparations for PR #494.
sind/cosd was one of the speed bottlenecks in the DE_2000 calculation. The degradation in accuracy due to this change is negligible. (For _de2000_t, it is rather an improvement.)

[codecov]

Codecov Report

Merging #498 (a5aca15) into master (9d69f98) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #498      +/-   ##
==========================================
+ Coverage   93.61%   93.64%   +0.02%     
==========================================
  Files           9        9              
  Lines        1097     1101       +4     
==========================================
+ Hits         1027     1031       +4     
  Misses         70       70              

Impacted Files	Coverage Δ
src/differences.jl	99.05% <100.00%> (+0.03%)	⬆️

---

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 9d69f98...a5aca15. Read the comment docs.

[kimikage]

Benchmark

before

julia> @benchmark colordiff(a, b) setup=(a=rand(Lab); b=rand(Lab))
BenchmarkTools.Trial:
  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     165.744 ns (0.00% GC)
  median time:      202.635 ns (0.00% GC)
  mean time:        200.747 ns (0.00% GC)
  maximum time:     880.501 ns (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     759

after

julia> @benchmark colordiff(a, b) setup=(a=rand(Lab); b=rand(Lab))
BenchmarkTools.Trial:
  memory estimate:  0 bytes
  allocs estimate:  0
  --------------
  minimum time:     115.795 ns (0.00% GC)
  median time:      130.392 ns (0.00% GC)
  mean time:        132.842 ns (0.00% GC)
  maximum time:     1.039 μs (0.00% GC)
  --------------
  samples:          10000
  evals/sample:     918

Incidentally, my main machine broke down the other day, and I'm currently using an older generation (Kaby Lake) machine. 😭 So we can't make a quantitative comparison with recent benchmark results.

julia> versioninfo()
Julia Version 1.6.1
Commit 6aaedecc44 (2021-04-23 05:59 UTC)
Platform Info:
  OS: Windows (x86_64-w64-mingw32)
  CPU: Intel(R) Core(TM) i7-7500U CPU @ 2.70GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-11.0.1 (ORCJIT, skylake)

The new machine will have a different CPU type (Tiger Lake).