Re: Does Anyone Compile Aspell with -DNDEBUG?

On June 1, 2008 05:18:06 pm Kevin Atkinson wrote:
> without running the tests multiple times.  Since on the second run the
> numbers may become 0.07 0.06.  There are a lot of things that can effect
> cpu time.

Tried again, made wordlist 10x larger, ran several times.
Conclusion == agree.

> If you thing that the 0.01 difference is insignificant (thus running the
> tests again make no sense) than can I summarize your email as following:
>
>    Coming prezip with -DNDEBUG or -g when combined with -O2 does not
>    noticeably effect performance.

Prezip is a poor example to demonstrate speed drop since the amount of 
time spent in file I/O far overshadows the time spent within prezip itself 
to build "1" string of text. It is a small self contained example which did 
show assert on it's own is only a small 1%-2% increase in code, but, -g is 
a large increase in code (11k vs 17k).

Aspell, on the other hand, spends seconds within itself (5 to 37 seconds 
http://aspell.net/test/common-all/ ). Based on the code increase seen in 
prezip (due to all the extra .loc statements), it could be assumed that 
aspell with assert is a small increase in 1%..2% code but (based on the 
code increase seen in prezip) would run far slower with _DEBUG vs DNDEBUG.