Subject: first results of the firsts perl tests...
From: Joaquín Cuenca Abela (cuenca@celium.net)
Date: Fri Apr 13 2001 - 20:15:56 CDT
I've written a perl script that writes compulsively (until it reaches
1000000 of words) random english words in Abiword.
I've not yet introduced changes in the format (only random words
separated randomly by spaces, \n, \t, or nothing at all. Spaces are the
most probable separator).
My firsts conclusions is that our memory use doesn't sucks (relatively)
so badly as I thinked at the first time.
top says that abiword is getting:
SIZE RSS SHARE
10188 9.9M 5840
when abi is in his 10 page or so.
Now, the bad news. Abi speed *S*U*C*K*S* when we are far from the
beginning. The script start writting at a very high speed (you know, I
only can see how the words comes one after other), but when he reaches
the 3 or 4 page, his speed decreases visibly.
Right now, it's in his 10 page and its speed is approx. the same speed
that I can get in a good day.
The very good news is that in all that random test, abi don't segfaults,
nor asserts. I will add random formatting changes to the test, and I
will post the results. I will try to see if the slowdown is my test
fault or abi fault.
I think that the test can be run with the slightly outdated version of
the perl bindings that are in cvs.
If somebody is interested, speak up and I will send you the script. If
somebody is *really* interested, I can even commit it to cvs (it's a 25
lines script -- with white lines--).
Cheers,
P.S.: The debug stuff is slowing abi *A LOT* (well, at least I'm seeing
*A LOT* of debug messages in my terminal...).
-- Joaquín Cuenca Abela cuenca@celium.net
This archive was generated by hypermail 2b25 : Fri Apr 13 2001 - 20:16:04 CDT