Originally Posted by wabard
However, The problem has nothing to do with software bloat, but rather ease of use versus a minor inconvenience of cpu time (which does not really affect the "end" user).
The difference between running chmod 75,000 times and running chmod 750 times is not "minor".
Take a good look at any servers stats and you'll find it spends a lot of time idle.
Unless you install Wordpress. After that, any bit of extra CPU you can squeeze out is precious.
(Adding one instance to a well-populated server bloated it from <0.1 avg load on PIII-600 to >.3 avg load on P4-2800!) IOW these tranquil conditions, while nice, shouldn't be assumed, lest you be the thing that makes them not true...
Also imagine using an expensive licensed OS where time literally is money, whether you're using the time or not. Wouldn't you want to extract as much work for the money as you could get? Many such servers operate under max loads continually.
I would argue that as I type at the keyboard, I'm using less than 1% of one core out of six, with the other 5 cores being used up for a simulation I'm currently running (dynamic flow simulation)...
Congratulations on your 6 cores, if I had so much CPU time sitting idle I might try farming it out for something else...
I just ran my command over 10k of my simulation files in over 400 directories. It returned in ~5 seconds versus the ~2 seconds for the xarg version.
A greater than 100% difference in performance sounds significant to me. Under a machine less fantastic than yours, the difference could easily become tedious.
I was asking why there seems to be a resistance to use built-in options for find where no bias should be found.
Eh? You've proven yourself there's a difference, and obviously know why.
P.S: If you have -print0, try \+ instead of \; to get all the benefits of find+xargs without all the typing.