On: Beating C with 70 lines of Go

[jstanley]

Alternatively, here's my entry for "Beating C with 40 lines of C": https://pastebin.com/JzFfE5GB

  $ time wc -w 100m
  2266395 100m
  
  real 0m4.568s

  $ cc -o wc wc.c
  
  $ time ./wc 100m
  2343390 100m
  
  real 0m0.511s

Of course, it disagrees on the answer, because I just used 100M of random data and it doesn't care about wide characters. It gives the same answer as GNU on plain ASCII text.

It's not faster because it's better, it's faster because it is doing less.

[Quekid5 → jstanley]

> It's not faster because it's better, it's faster because it is doing less.

I can't recall who said it, but this reminds me of this idea: If you don't care about correctness then I can make it as fast as you like.

[glandium → jstanley]

Alternatively, here's my entry for "Beating GNU wc with GNU wc":

    $ dd if=/dev/urandom of=100m bs=1M count=100 2> /dev/null

    $ time wc -w 100m
    2319144 100m

    real 0m4.543s
    user 0m4.512s
    sys 0m0.029s

    $ time env LC_CTYPE=C wc -w 100m
    2308032 100m

    real 0m0.842s
    user 0m0.842s
    sys 0m0.000s

Your program is faster, though.