Time taken for nnvirtual..

Time taken for nnvirtual..

[Andy Eskilsson]

Are nnvirtual constructed to be slow?? The question arises when I have
a nnvirtual group containing about 5-10 huge newsgroups, (well I only
read the abpe hierarchy for testing purposes *cough* *cough*) about
3500 unread articles..

Ok. I press space on that nnvirtual group.. tumdedumm. wait for a long
while, fires up a slave reads the groups one by one with normal
reading, and are finished *before* the nnvirtual group is finished
with it's 'nntp-reading'.

Why this slowness, is nntp-reading an NP-problem? I know sorting and
scoring is, but reading from the server? My computer isn't slowing
down either, no excessive thrashing of the HD and so on.

	/Andy
-- 
 Hi I am an alien .sig, and at the moment I am having sex to your
 mind, by looking at your smile I can see that you like it.

 Unsolicited commercial email is subject to an archival fee of $400.
 See <http://www.fukt.hk-r.se/~flognat/mail/> for more info.

Re: Time taken for nnvirtual..

[Lars Magne Ingebrigtsen]

Andy Eskilsson <flognat@fukt.hk-r.se> writes:

> Are nnvirtual constructed to be slow??

Yes.  It was faster before, but it wasn't as versatile.

> The question arises when I have
> a nnvirtual group containing about 5-10 huge newsgroups, (well I only
> read the abpe hierarchy for testing purposes *cough* *cough*) about
> 3500 unread articles..
> 
> Ok. I press space on that nnvirtual group.. tumdedumm. wait for a long
> while, fires up a slave reads the groups one by one with normal
> reading, and are finished *before* the nnvirtual group is finished
> with it's 'nntp-reading'.

This sounds odd.  nnvirtual is slow, but not *that* slow.  For
instance, I have the ~10 Emacs groups gathered in one virtual group.
It uses ~2 seconds to scan the group.  Entering the group doesn't take
much time...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Re: Time taken for nnvirtual..

[Han Pilmeyer]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

Lars> Andy Eskilsson <flognat@fukt.hk-r.se> writes:

>> Ok. I press space on that nnvirtual group.. tumdedumm. wait for a long
>> while, fires up a slave reads the groups one by one with normal
>> reading, and are finished *before* the nnvirtual group is finished
>> with it's 'nntp-reading'.

Lars> This sounds odd.  nnvirtual is slow, but not *that* slow.  For
Lars> instance, I have the ~10 Emacs groups gathered in one virtual group.
Lars> It uses ~2 seconds to scan the group.  Entering the group doesn't take
Lars> much time...

I've noticed the same thing. I only have a 3 or 4 groups in my virtual
group and they are rather small too, but it's still quicker to enter
and read the individual groups than it is to enter the virtual
group. In fact this is what's keeping me from creating a few more
virtual groups.

Cheers.
-- 
Han Pilmeyer, han@uto.dec.com, Phone: +31 30 2832671, FAX: +31 30 2832686
Digital Equipment B.V., Europalaan 44, 3526 KS Utrecht, The Netherlands
|         Answers are the easy part, questions raise the doubt   -JB          |
All opinions are those of the author, not of Digital Equipment!

Re: Time taken for nnvirtual..

[Andy Eskilsson]

/ Lars Magne Ingebrigtsen <larsi@ifi.uio.no> wrote:
| 
| This sounds odd.  nnvirtual is slow, but not *that* slow.  For
| instance, I have the ~10 Emacs groups gathered in one virtual group.
| It uses ~2 seconds to scan the group.  Entering the group doesn't take
| much time...

Mmh it is quite fast on small groups, but the time to get the group
seems to grow exponentially with the articles or something the
like.. Any other comments/people who have tried it?

	/Andy

-- 
 Hi I am an alien .sig, and at the moment I am having sex to your
 mind, by looking at your smile I can see that you like it.

 Unsolicited commercial email is subject to an archival fee of $400.
 See <http://www.fukt.hk-r.se/~flognat/mail/> for more info.

Re: Time taken for nnvirtual..

[Steven L Baur]

>>>>> "Andy" == Andy Eskilsson <flognat@fukt.hk-r.se> writes:

Andy> Mmh it is quite fast on small groups, but the time to get the
Andy> group seems to grow exponentially with the articles or something
Andy> the like.. Any other comments/people who have tried it?

My experiences concur with yours, except that I didn't bother to
continue trying with it for very long, it was just too much of a waste
of time.  I have a suspicion that part of the slowdown is related to
crosspost processing, but I'm not certain.
-- 
steve@miranova.com baur
Unsolicited commercial e-mail will be billed at $250/message.
What are the last two letters of "doesn't" and "can't"?
Coincidence?  I think not.

Re: Time taken for nnvirtual..

[Nathan O. Siemers]

nnvirtual is very slow for me also (5.2.40). This is also with very
large groups (comp.os.linux.misc, etc.)

nathan

Re: Time taken for nnvirtual..

[Erik Selberg]

I also see the heinous slowdown... both with a four-group nntp virtual
and a two-group mail/nntp virtual.

Perhaps the sorting algorithm is O(n^2) or something?

-erik
-- 
				Erik Selberg
"I get by with a little help	selberg@cs.washington.edu
 from my friends."		http://www.cs.washington.edu/homes/selberg

Re: Time taken for nnvirtual..

[Andy Eskilsson]

/ Erik Selberg <selberg@cs.washington.edu> wrote:
| 
| I also see the heinous slowdown... both with a four-group nntp virtual
| and a two-group mail/nntp virtual.
| 
| Perhaps the sorting algorithm is O(n^2) or something?

Nahh the sort is quite quick, what I mean is that I can read all the
groups in the virtual group before the text goes over from reading
nntp to scoring.. 

or does it sort at the same time it reads the nntp stream?

BTW this is mostly annoying me when I come to the computer after a
weekend and have 3000-4000 unread in the virtual newsgroup, with the
nightly dosis of ~500 it is ok.

	/andy

-- 
 Hi I am an alien .sig, and at the moment I am having sex to your
 mind, by looking at your smile I can see that you like it.

 Unsolicited commercial email is subject to an archival fee of $400.
 See <http://www.fukt.hk-r.se/~flognat/mail/> for more info.

Re: Time taken for nnvirtual..

[David Moore]

Steven L Baur <steve@miranova.com> writes:

> 
> >>>>> "Andy" == Andy Eskilsson <flognat@fukt.hk-r.se> writes:
> 
> Andy> Mmh it is quite fast on small groups, but the time to get the
> Andy> group seems to grow exponentially with the articles or something
> Andy> the like.. Any other comments/people who have tried it?
> 
> My experiences concur with yours, except that I didn't bother to
> continue trying with it for very long, it was just too much of a waste
> of time.  I have a suspicion that part of the slowdown is related to
> crosspost processing, but I'm not certain.

	I had problems with nnvirtual even on tiny groups.  I wanted to
combine 3 email groups in which two get 1 message a day and the other
gets about 8.  With about 200 old messages in the groups, and this daily
arrival rate, I found entering the nnvirtual group to be over an order
of magnitude slower than a normal group.

	Additionally, I seem to recall that exitting the nnvirtual group
took an extreme amount of time, which no one else seems to have
mentioned.  This virtual group has 'gnus-show-threads' set to nil, as
did one of the three real groups, if that affects anything.  I use nnml
as a backend, and had set the 3 real groups to level 4 or 5 to hide
them.  The virtual group was at level 2.


-- 
David Moore <dmoore@ucsd.edu>       | SysAdmin/Programmer       __o
UCSD Dept. of Anesthesiology - 0629 | Work: (619) 534-3865    _ \<,_
La Jolla, CA 92093-0629             | Fax:  (619) 534-1445   (_)/ (_)
<URL:http://oj.egbt.org/dmoore/>    | Solo Furnace Creek 508 -- 1996!