overhead of extra modules?

[Marc Slemko <ma...@worldgate.com>]

anyone done any benchmarking on how much overhead extra modules add?

In particular:

+Module status_module       mod_status.o
+Module info_module         mod_info.o
+Module proxy_module        modules/proxy/libproxy.a
+Module rewrite_module      mod_rewrite.o
+Module db_auth_module      mod_auth_db.o
+Module digest_module       mod_digest.o
+Module expires_module      mod_expires.o
+Module headers_module      mod_headers.o

...added to the distributed Configuration in 1.2.5.

Re: overhead of extra modules?

[Dean Gaudet <dg...@arctic.org>]

In 1.2.x they have a significant effect because run_method isn't optimized
to remove NULLs from the call lists... like 7% of the time or something
ridiculous like that is spent looking at NULLs since run_method is called
like 7 times per request. 

In 1.3 they have a lesser effect... but we haven't accepted Dmitry's
optimization patch that fixes invoke_handler() in the same way.  Mostly
because I haven't been able to get a test environment that can handle the
load apache can now sustain ;)  So I haven't been able to measure the
actual improvement of Dmitry's patch... 

Dean

On Sun, 8 Feb 1998, Marc Slemko wrote:

> anyone done any benchmarking on how much overhead extra modules add?
> 
> In particular:
> 
> +Module status_module       mod_status.o
> +Module info_module         mod_info.o
> +Module proxy_module        modules/proxy/libproxy.a
> +Module rewrite_module      mod_rewrite.o
> +Module db_auth_module      mod_auth_db.o
> +Module digest_module       mod_digest.o
> +Module expires_module      mod_expires.o
> +Module headers_module      mod_headers.o
> 
> ...added to the distributed Configuration in 1.2.5.
> 
>