On Tue, 3 Dec 2002 RBramante_at_on.com wrote:

> I am experiencing very poor performance for HTTP PUT requests. I have been
> ignoring these because I have been testing with debug builds and was
> assuming that was the cause but now I am testing release code and not
> seeing much better results. I thought the problems may have been in my
> code, but running a test with curl itself is giving the same exact results.

Ugh. This seems like badness.

> Curl is consistently averaging ~16 seconds around 100k/sec
>
> A simple Java app using HttpURLConnection is consistently transfering the
> same file to the same server in ~2 seconds or less (~900k/sec).

The difference is really HUGE too. We need to get down to the bits with this.

> Needless to say, I was quite suprised by the PUT discrepancy. Any think I
> can tweak to try and improve this? What other info would be of use in
> debugging? This is with the win32 build. Thanks for any thoughts.

Can you compile/builld it with some kind of profiling to see where it spends
the most time?

Another option is to insert infof() calls in lots of places and output the
time to see the relative time between different outputs to see where the
differences are.

Oh, and of course, record as much as possible with a network analyzer to
detect if there's anything weird going on network-wise.

I'll try to setup something like this in my end and run some tests to see if
I can see something similar on a Linux build.

-- 
 Daniel Stenberg -- curl, cURL, Curl, CURL. Groks URLs.
-------------------------------------------------------
This SF.net email is sponsored by: Microsoft Visual Studio.NET 
comprehensive development tool, built to increase your 
productivity. Try a free online hosted session at:
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr0003en