Proximodo CPU Efficency

[Guest]

First up: Thanks! An open source proxomitron has been a wet dream of mine for a while now.

Now on to the report... I have a long list of some pretty intenstive filters and the good news is that they all migrated into proximodo without problem and appear to all be functioning perfectly. The bad news is that the CPU usage is through the roof. With proxomitron a page load generates usage of at max 40% for a split second or two. With proximodo it pegs the needle at 100% for 3-5 seconds per page load.

I hate to be complaining so soon about a project I'm so grateful for, but, well, you can't fix what you don't know about, and this might not be an issue for a lot of people with less filters.

[kuruden]

Try this:

\1=$AV(\1)&&\1=$AV(\1zzz\1)

Pretty useless as is, but I used this pattern to tune \1 and $AV for performance. Result: Proximodo scans pages for than pattern more than twice faster than Proxomitron! What I mean is, it takes some effort to reach performance, but it works. I did it on certain pieces of code, I didn't yet on much of it, but I will.

I am glad Proximodo is no slower than reasonable in its first version Believe me, first versions of Proxomitron were WAYYYY slower than 4.5 Naoko, and that's normal: good development process is
1. write clean code
2. optimize it (only were it needs it)

"Premature optimization is the root of all evil" -- Car Hoare

Btw, give a try to 0.2, I didn't optimize it further yet but I removed bugs that could have explained the CPU consumption. If you still have that kind of speed pb, try determining which of your filters slows down the most and I'll do some speed tuning based on it.

Kuruden

[mozerd]

kuruden Wrote:Btw, give a try to 0.2, I didn't optimize it further yet but I removed bugs that could have explained the CPU consumption.

I went to this sourceforge.net/projects/proximodo but could only see version 0.1 dated November 15, 2004 -- how does one get 0.2?

[atom]

https://sourceforge.net/project/showfiles.p...lease_id=286084