[Pan-users] Re: Strange issue with missing headers when downloading new headers...and a request

[Duncan]

fred <address@hidden> posted
address@hidden, excerpted below, on  Wed, 11 Oct 2006
00:15:56 -0500:

> For whatever reason 0.116 seems much faster than 0.115. Fc5-32bit  X2
> 3880+. Thanks Charles

I've not loaded 0.116 yet, but I've gone thru the bugs in the
announcement, and it /should/ be faster, in at least two ways, which will
affect different users differently.

1) For servers allowing many connections (it's easiest to see if it's over
four connections, for which you have to edit servers.xml appropriately),
<0.116 would take some time to reach the allowed number of connections. 
I know I noticed this, but hadn't filed a bug as I thought that's just
the way it was (out of pan's control). This has been fixed according to the
announce.

2) A partial-buffer timing issue has been fixed that was previously being
worked around with a short (200 ms) sleep.  With the root issue fixed, the
sleep could be and was removed.  I believe those micro-sleeps were the big
reason each connection wasn't reaching its full potential speeds, and it
took more connections to achieve the same speeds compared to old-pan.

The combination of the two of these fixes should make 0.116 much faster,
hopefully now able to achieve the efficiency of old-pan, which from my
observation very nearly approached the max theoretical connections speeds
allowed, when the per-connection speed was capped at something lower than
one's total internet pipe speed.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman