algorithms/GenericFunctions: x == 0 -> iszero(x)

[nsajko]

Updates #1868

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.15%. Comparing base (5ffdd2d) to head (186416a).
Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1869      +/-   ##
==========================================
- Coverage   88.17%   88.15%   -0.02%     
==========================================
  Files         120      120              
  Lines       30220    30250      +30     
==========================================
+ Hits        26645    26668      +23     
- Misses       3575     3582       +7     

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

[thofma]

Sorry, I should have just written it here once. I don't think this is an improvement.

[fingolfin]

It is measurably an improvement in terms of performance (of course! we know that), albeit a very small one:

julia> x = big(5)
5

julia> @benchmark iszero($x)
BenchmarkTools.Trial: 10000 samples with 1000 evaluations.
 Range (min … max):  2.125 ns … 5.708 ns  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     2.208 ns             ┊ GC (median):    0.00%
 Time  (mean ± σ):   2.196 ns ± 0.049 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

             █          ▇▁
  ▂▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁██▁▁▁▁▁▁▁▁▁▁▃▁▁▁▁▁▁▁▁▁▁▃▁▁▁▁▁▁▁▁▁▁▂ ▂
  2.12 ns        Histogram: frequency by time       2.33 ns <

 Memory estimate: 0 bytes, allocs estimate: 0.

julia> @benchmark $x == 0
BenchmarkTools.Trial: 10000 samples with 1000 evaluations.
 Range (min … max):  2.708 ns … 17.500 ns  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     2.833 ns              ┊ GC (median):    0.00%
 Time  (mean ± σ):   2.830 ns ±  0.168 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

         ▁      ▇      █       ▃      ▃      ▂      ▂        ▂
  ▄▁▁▁▁▁▁█▁▁▁▁▁▁█▁▁▁▁▁▁█▁▁▁▁▁▁▁█▁▁▁▁▁▁█▁▁▁▁▁▁█▁▁▁▁▁▁█▁▁▁▁▁▁▆ █
  2.71 ns      Histogram: log(frequency) by time     3.04 ns <

 Memory estimate: 0 bytes, allocs estimate: 0.

So one can of course argue whether it is worth it in these situations.

Then again, why not take advantage of this, however minor it may be? The code isn't harder to read one way or the other, is it?

Whether the code looks better one way or another is a matter of taste?

[thofma]

Ah sorry, I was not thinking about BigInt case. Everything is fine then.