f



Is there still a 100% CPU usage problem?

Hi,

When I tried to use Pegasus about a year ago the cpu usage went up to
100% whenever the program was downloading - does this problem still
exist?

Regards,

-- 

John Latter

Model of an Internal Evolutionary Mechanism (based on an extension to homeostasis) linking Stationary-Phase Mutations to the Baldwin Effect.
http://members.aol.com/jorolat/TEM.html

'Where Darwin meets Lamarck?' Discussion Egroup
http://groups.yahoo.com/group/evomech
0
John
7/22/2004 6:53:32 PM
comp.mail.pegasus-mail.misc 1751 articles. 0 followers. Post Follow

25 Replies
662 Views

Similar Articles

[PageSpeed] 57

[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

It does still exist, but why would that be a significant problem?  Whether
Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
multitask.

In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004 19:53:32
+0100, John Latter <jorolat@msn.com> wrote:

>Hi,
>
>When I tried to use Pegasus about a year ago the cpu usage went up to
>100% whenever the program was downloading - does this problem still
>exist?
>
>Regards,

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/22/2004 7:33:41 PM
On Thu, 22 Jul 2004 19:33:41 GMT, John Navas
<spamfilter0@navasgroup.com> wrote:

>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>
>It does still exist, but why would that be a significant problem?  Whether
>Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>multitask.
>
>In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004 19:53:32
>+0100, John Latter <jorolat@msn.com> wrote:
>
>>Hi,
>>
>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>100% whenever the program was downloading - does this problem still
>>exist?
>>
>>Regards,

When the CPU usage goes up so does the CPU temperature (I use VCool &
MBM)

-- 

John Latter

Model of an Internal Evolutionary Mechanism (based on an extension to homeostasis) linking Stationary-Phase Mutations to the Baldwin Effect.
http://members.aol.com/jorolat/TEM.html

'Where Darwin meets Lamarck?' Discussion Egroup
http://groups.yahoo.com/group/evomech
0
John
7/22/2004 8:59:43 PM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <iha0g09tv571hiqsr94kv0nfuvhc40nksi@4ax.com> on Thu, 22 Jul 2004 21:59:43
+0100, John Latter <jorolat@msn.com> wrote:

>On Thu, 22 Jul 2004 19:33:41 GMT, John Navas
><spamfilter0@navasgroup.com> wrote:
>
>>It does still exist, but why would that be a significant problem?  Whether
>>Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>>multitask.
>>
>>In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004 19:53:32
>>+0100, John Latter <jorolat@msn.com> wrote:
>>
>>>Hi,
>>>
>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>100% whenever the program was downloading - does this problem still
>>>exist?
>>>
>>>Regards,
>
>When the CPU usage goes up so does the CPU temperature (I use VCool &
>MBM)

[yawn]  As in, it's really not an issue.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/22/2004 9:06:38 PM
John Navas <spamfilter0@navasgroup.com> Thu, 22 Jul 2004 19:33:41 GMT 
wrote  ...

>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>
>It does still exist, but why would that be a significant problem?  Whether
>Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>multitask.
>
But is not it like apps take CPU time from free pool
( left by higher priorities ones ) ?
And if there is none ?


-- 
"Libor the Wanderer" <poutnikqdel@atthislas.cz>
Sorry for my english in case of such troubles.
As dialup user I am sorry for late response.
ForPrivateResponseRemoveDelAndThisFromAboveAddress.
0
Libor
7/23/2004 5:09:06 AM
>Subject: Re: Is there still a 100% CPU usage problem?
>From: John Navas spamfilter0@navasgroup.com 
>Date: 22/07/2004 22:06 GMT Daylight Time
>Message-id: <y%VLc.3714$54.52238@typhoon.sonic.net>
>
>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>
>In <iha0g09tv571hiqsr94kv0nfuvhc40nksi@4ax.com> on Thu, 22 Jul 2004 21:59:43
>+0100, John Latter <jorolat@msn.com> wrote:
>
>>On Thu, 22 Jul 2004 19:33:41 GMT, John Navas
>><spamfilter0@navasgroup.com> wrote:
>>
>>>It does still exist, but why would that be a significant problem?  Whether
>>>Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>>>multitask.
>>>
>>>In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004
>19:53:32
>>>+0100, John Latter <jorolat@msn.com> wrote:
>>>
>>>>Hi,
>>>>
>>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>>100% whenever the program was downloading - does this problem still
>>>>exist?
>>>>
>>>>Regards,
>>
>>When the CPU usage goes up so does the CPU temperature (I use VCool &
>>MBM)
>
>[yawn]  As in, it's really not an issue.
>

It hasn't been proposed as an issue and consequently cannot be dismissed as
such :)

>-- 
>Best regards,
>John Navas     <http://navasgrp.home.att.net/>
>WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
>
>
>
>
>
>


Model of an Internal Evolutionary Mechanism 
http://members.aol.com/jorolat/TEM.html

'Where Darwin meets Lamarck?' Discussion Egroup
http://groups.yahoo.com/group/evomech
0
jorolat
7/23/2004 6:43:38 AM
John Navas <spamfilter0@navasgroup.com> wrote:

} [POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
} 
} It does still exist, but why would that be a significant problem?  Whether
} Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
} multitask.

True, but it slows my system to a crawl when I'm doing other CPU intensive
stuff and I notice it, annoyingly, nearly every day [e.g., the flash-player
plugin for IE also uses a lot of CPU time and it is obvious when Peg
decides to wake up and grab the CPU].

Turned around, any decent application ought to be able to figure out how to
do I/O without a CPU spin loop [that was considered incompetent-programming
form when I first learned about dealing I/O, 30+ years ago]

  /Bernie\

-- 
Bernie Cosell                     Fantasy Farm Fibers
bernie@fantasyfarm.com            Pearisburg, VA
    -->  Too many people, too few sheep  <--          
0
Bernie
7/23/2004 3:05:48 PM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <90a2g0pgsh87vas75f80tagb3sr20gie49@library.airnews.net> on Fri, 23 Jul
2004 11:05:48 -0400, Bernie Cosell <bernie@fantasyfarm.com> wrote:

>John Navas <spamfilter0@navasgroup.com> wrote:
>
>} It does still exist, but why would that be a significant problem?  Whether
>} Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>} multitask.
>
>True, but it slows my system to a crawl when I'm doing other CPU intensive
>stuff and I notice it, annoyingly, nearly every day [e.g., the flash-player
>plugin for IE also uses a lot of CPU time and it is obvious when Peg
>decides to wake up and grab the CPU].

As I wrote, I just don't see a problem on Windows XP, even when running other
CPU intensive tasks.  So your problem may be due to:

1.  Different OS.

2.  System tuning.  (Mine is tuned for balanced performance.)

3.  Much slower CPU.  (I'm running an AMD Athlon XP 2400+, and multitasking
performance isn't an issue even with much more CPU-intensive MPEG2 video.)

4.  Other inefficient apps (e.g., flash-player plugin for IE, which really
shouldn't be terribly CPU intensive).

>Turned around, any decent application ought to be able to figure out how to
>do I/O without a CPU spin loop [that was considered incompetent-programming
>form when I first learned about dealing I/O, 30+ years ago]

That depends on the available OS facilities.  The TCP/IP code in Pegasus is
probably a legacy of supporting ancient versions of Winsock that don't have
terribly robust handling of timeouts.  If David were to abandon support of old
systems and take the time to rewrite his code to support only current systems,
then it could probably be improved, but the cost would be high for such a low
payback, and I suspect he has other priorities.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/23/2004 4:39:51 PM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <rbbMc.3850$54.54906@typhoon.sonic.net> on Fri, 23 Jul 2004 16:39:51 GMT,
John Navas <spamfilter0@navasgroup.com> wrote:

>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>
>In <90a2g0pgsh87vas75f80tagb3sr20gie49@library.airnews.net> on Fri, 23 Jul
>2004 11:05:48 -0400, Bernie Cosell <bernie@fantasyfarm.com> wrote:
>
>>John Navas <spamfilter0@navasgroup.com> wrote:
>>
>>} It does still exist, but why would that be a significant problem?  Whether
>>} Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>>} multitask.
>>
>>True, but it slows my system to a crawl when I'm doing other CPU intensive
>>stuff and I notice it, annoyingly, nearly every day [e.g., the flash-player
>>plugin for IE also uses a lot of CPU time and it is obvious when Peg
>>decides to wake up and grab the CPU].
>
>As I wrote, I just don't see a problem on Windows XP, even when running other
>CPU intensive tasks.  ...

Just checked.  DVD MPEG2 playback (which should be much more CPU intensive
than Flash) takes 10-20% of my CPU, depending on player.  Checking of mail by
Pegasus has absolutely no effect on smoothness of playback.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/23/2004 5:15:23 PM
John Navas <spamfilter0@navasgroup.com> organized electrons in article
news:LIbMc.3857$54.54919@typhoon.sonic.net that appeared as follows:
> In <rbbMc.3850$54.54906@typhoon.sonic.net> on Fri, 23 Jul 2004
> 16:39:51 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>> In <90a2g0pgsh87vas75f80tagb3sr20gie49@library.airnews.net> on Fri,
>> 23 Jul 2004 11:05:48 -0400, Bernie Cosell <bernie@fantasyfarm.com>
>> wrote:
>>> John Navas <spamfilter0@navasgroup.com> wrote:
>>>
>>> } It does still exist, but why would that be a significant problem?
>>> Whether } Pegasus is in a spin loop or not, a decent OS (e.g.,
>>> Windows XP) will still } multitask.
>>>
>>> True, but it slows my system to a crawl when I'm doing other CPU
>>> intensive stuff and I notice it, annoyingly, nearly every day
>>> [e.g., the flash-player plugin for IE also uses a lot of CPU time
>>> and it is obvious when Peg
>>> decides to wake up and grab the CPU].
>>
>> As I wrote, I just don't see a problem on Windows XP, even when
>> running other CPU intensive tasks.  ...
>
> Just checked.  DVD MPEG2 playback (which should be much more CPU
> intensive than Flash) takes 10-20% of my CPU, depending on player.
> Checking of mail by Pegasus has absolutely no effect on smoothness of
> playback.

How about the smoothness of writing to a CD-R, DVD+R, or DVD-R, or of
serving other users?

-- 
Thanks and Best Regards,  Jeff G.
My email address in the following is ROT13 encoded to reduce spam:
Guvf vf gur rznvy nqqerff: WIQTWBBBJIAI@fcnzzbgry.pbz

0
Jeff
7/23/2004 5:42:53 PM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <x6cMc.98018$a92.13495@twister.nyc.rr.com> on Fri, 23 Jul 2004 17:42:53
GMT, "Jeff G."
<WIQTWBBBJIAI%fcnzzbgry%pbz%PRECEDING%ROT13%ENCODED%TO%DESPAM%US@example.com>
wrote:

>John Navas <spamfilter0@navasgroup.com> organized electrons in article
>news:LIbMc.3857$54.54919@typhoon.sonic.net that appeared as follows:
>> In <rbbMc.3850$54.54906@typhoon.sonic.net> on Fri, 23 Jul 2004
>> 16:39:51 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>>> In <90a2g0pgsh87vas75f80tagb3sr20gie49@library.airnews.net> on Fri,
>>> 23 Jul 2004 11:05:48 -0400, Bernie Cosell <bernie@fantasyfarm.com>
>>> wrote:
>>>> John Navas <spamfilter0@navasgroup.com> wrote:
>>>>
>>>> } It does still exist, but why would that be a significant problem?
>>>> Whether } Pegasus is in a spin loop or not, a decent OS (e.g.,
>>>> Windows XP) will still } multitask.
>>>>
>>>> True, but it slows my system to a crawl when I'm doing other CPU
>>>> intensive stuff and I notice it, annoyingly, nearly every day
>>>> [e.g., the flash-player plugin for IE also uses a lot of CPU time
>>>> and it is obvious when Peg
>>>> decides to wake up and grab the CPU].
>>>
>>> As I wrote, I just don't see a problem on Windows XP, even when
>>> running other CPU intensive tasks.  ...
>>
>> Just checked.  DVD MPEG2 playback (which should be much more CPU
>> intensive than Flash) takes 10-20% of my CPU, depending on player.
>> Checking of mail by Pegasus has absolutely no effect on smoothness of
>> playback.
>
>How about the smoothness of writing to a CD-R, DVD+R, or DVD-R, or of
>serving other users?

No problem there either -- in fact, I was writing a DVD-R at the time at 4x.
DVD-R burning is I/O intensive, not CPU intensive.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/23/2004 8:57:56 PM
John Navas <spamfilter0@navasgroup.com> wrote:

} [POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
} 
} In <90a2g0pgsh87vas75f80tagb3sr20gie49@library.airnews.net> on Fri, 23 Jul
} 2004 11:05:48 -0400, Bernie Cosell <bernie@fantasyfarm.com> wrote:
} 
} >John Navas <spamfilter0@navasgroup.com> wrote:
} >
} >} It does still exist, but why would that be a significant problem?  Whether
} >} Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
} >} multitask.
} >
} >True, but it slows my system to a crawl when I'm doing other CPU intensive
} >stuff and I notice it, annoyingly, nearly every day [e.g., the flash-player
} >plugin for IE also uses a lot of CPU time and it is obvious when Peg
} >decides to wake up and grab the CPU].
} 
} As I wrote, I just don't see a problem on Windows XP, even when running other
} CPU intensive tasks.  So your problem may be due to:
} 
} 1.  Different OS.

XP/Pro for me...

} 2.  System tuning.  (Mine is tuned for balanced performance.)

Dunno -- mine is pretty much "out of the box" and works satisfactorily for
every *OTHER* app than Peg...

} 3.  Much slower CPU.  (I'm running an AMD Athlon XP 2400+, and multitasking
} performance isn't an issue even with much more CPU-intensive MPEG2 video.)

That could be--  I have a fairly slow system.  And an interesting comment:
that I might have to upgrade my CPU *not* because I'm a heavy-duty gamer,
but because [that part of] Peg is poorly coded...

} >Turned around, any decent application ought to be able to figure out how to
} >do I/O without a CPU spin loop [that was considered incompetent-programming
} >form when I first learned about dealing I/O, 30+ years ago]
} 
} That depends on the available OS facilities.

AFAIK, XP/Pro's TCP facilities are as good as you'd need.  No *other* app
seems to need to run in a CPU spin loop to be able to field a mouse-click
and a TCP packet at the same time...

} ...  The TCP/IP code in Pegasus is
} probably a legacy of supporting ancient versions of Winsock that don't have
} terribly robust handling of timeouts.

That's our point: David has recoded a *TON* of stuff in Peg, but has left
this awful anomaly around [and it _does_ cause some of us trouble].

} ...  If David were to abandon support of old
} systems and take the time to rewrite his code to support only current systems,
} then it could probably be improved, but the cost would be high for such a low
} payback, and I suspect he has other priorities.

Why don't other mail apps have this problem?  [Agent, Eudora, Sylpheed,
etc, etc]?  Is David clinging to the trophy of having Peg be the *ONLY*
mail app left for Windows that'll still run on Windows 3.1 or something
like that?

  /Bernie\
-- 
Bernie Cosell                     Fantasy Farm Fibers
bernie@fantasyfarm.com            Pearisburg, VA
    -->  Too many people, too few sheep  <--          
0
Bernie
7/24/2004 10:00:46 AM
On Thu, 22 Jul 2004 19:33:41 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
> In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004 19:53:32
> +0100, John Latter <jorolat@msn.com> wrote:
>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>100% whenever the program was downloading - does this problem still

> It does still exist, but why would that be a significant problem?  Whether
> Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
> multitask.

Multitasking isn't the issue.  The issue is sluggish response.

It isn't a HUGE problem because of dynamic scheduling priority in XP,
but it is a SIGNIFICANT problem because XP is still noticably more
sluggish when an app does the 100% CPU spin.

I find it very annoying to have pegasus check mail in the background
when it hits the 100% CPU.  My solution was to use hamster to pull down
the mail, and pegasus talks locally to hamster.  Now with huge bandwidth
and no network problems, the 100% CPU never lasts long enough to notice.

Also, the improved IMAP support in pegasus 4 seems to work nicely with
hamster's IMAP server.  Now I can easily play with other mail clients
and access all my old mail in various folders.

sdb
0
Sylvan
7/25/2004 12:03:50 AM
On Fri, 23 Jul 2004 16:39:51 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
> That depends on the available OS facilities.  The TCP/IP code in Pegasus is
> probably a legacy of supporting ancient versions of Winsock that don't have
> terribly robust handling of timeouts.  If David were to abandon support of old

Agreed.  I've been doing winsock programming on windows since well
before Windows NT came out, and those first winsock versions were bad.
They were _very_ difficult to use efficiently.

> systems and take the time to rewrite his code to support only current systems,

No, the "writing a good program" problems with winsock went away with
the arrival of 32 bit windows (NT 3.1 and Windows 95, but not Win32s).
Granted, the actual winsock itself wasn't so great back then, but that
improvement happened over time, and totally transparent to all the code
I've written since 1994.

I don't consider that old of systems to be "current" and I doubt there
is any significant consumer base left who could not run code targetted
to that minimum.

> payback, and I suspect he has other priorities.

Evidently.

sdb
>
0
Sylvan
7/25/2004 3:36:52 PM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <slrncg7ksk.rop.ZsdbUse1+noZs_0407@knoppix.internal> on Sun, 25 Jul 2004
09:36:52 -0600, Sylvan Butler <ZsdbUse1+noZs_0407@Zbigfoot.Zcom.invalid>
wrote:

>On Fri, 23 Jul 2004 16:39:51 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>> That depends on the available OS facilities.  The TCP/IP code in Pegasus is
>> probably a legacy of supporting ancient versions of Winsock that don't have
>> terribly robust handling of timeouts.  If David were to abandon support of old
>
>Agreed.  I've been doing winsock programming on windows since well
>before Windows NT came out, and those first winsock versions were bad.
>They were _very_ difficult to use efficiently.
>
>> systems and take the time to rewrite his code to support only current systems,
>
>No, the "writing a good program" problems with winsock went away with
>the arrival of 32 bit windows (NT 3.1 and Windows 95, but not Win32s).
>Granted, the actual winsock itself wasn't so great back then, but that
>improvement happened over time, and totally transparent to all the code
>I've written since 1994.

My own experience is that the original Winsock code in Windows 95 was pretty
bad.

>I don't consider that old of systems to be "current" and I doubt there
>is any significant consumer base left who could not run code targetted
>to that minimum.

You might well be surprised at how people still run Pegasus on ancient
versions of Windows.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
7/26/2004 2:23:01 AM
John Latter <jorolat@msn.com> wrote in
news:iha0g09tv571hiqsr94kv0nfuvhc40nksi@4ax.com: 

> On Thu, 22 Jul 2004 19:33:41 GMT, John Navas
><spamfilter0@navasgroup.com> wrote:
> 
>>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>>
>>It does still exist, but why would that be a significant problem? 
>>Whether Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP)
>>will still multitask. 
>>
>>In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004
>>19:53:32 +0100, John Latter <jorolat@msn.com> wrote:
>>
>>>Hi,
>>>
>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>100% whenever the program was downloading - does this problem still
>>>exist? 
>>>
>>>Regards,
> 
> When the CPU usage goes up so does the CPU temperature (I use VCool &
> MBM)
> 

Try slowing down the program with the -z 1024 command line option.  FWIW, the 
100% CPU does not exuate to 100% CPU loading at all, it just means there is no  
CPU time given to the idle processor.  My CPU runs 100% all the time and there 
is no affect on any foreground process.

1024	Use blocking sockets; may be needed for some
	defective WINSOCK implementations.


-- 
Thomas R. Stephenson, CPL          
Member Pegasus Mail Support Team

0
Thomas
7/26/2004 5:11:47 PM
On 26 Jul 2004 17:11:47 GMT, "Thomas. Stephenson"
<mercury@mtshasta.lmms.lmco.com> wrote:

>John Latter <jorolat@msn.com> wrote in
>news:iha0g09tv571hiqsr94kv0nfuvhc40nksi@4ax.com: 
>
>> On Thu, 22 Jul 2004 19:33:41 GMT, John Navas
>><spamfilter0@navasgroup.com> wrote:
>> 
>>>[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]
>>>
>>>It does still exist, but why would that be a significant problem? 
>>>Whether Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP)
>>>will still multitask. 
>>>
>>>In <4330g05hp8t23m7g0r2qkbp8g4n143k0br@4ax.com> on Thu, 22 Jul 2004
>>>19:53:32 +0100, John Latter <jorolat@msn.com> wrote:
>>>
>>>>Hi,
>>>>
>>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>>100% whenever the program was downloading - does this problem still
>>>>exist? 
>>>>
>>>>Regards,
>> 
>> When the CPU usage goes up so does the CPU temperature (I use VCool &
>> MBM)
>> 
>
>Try slowing down the program with the -z 1024 command line option.  FWIW, the 
>100% CPU does not exuate to 100% CPU loading at all, it just means there is no  
>CPU time given to the idle processor.  My CPU runs 100% all the time and there 
>is no affect on any foreground process.
>
>1024	Use blocking sockets; may be needed for some
>	defective WINSOCK implementations.

I'll give it a go Thomas - thankyou for your reply :)

-- 

John Latter

Model of an Internal Evolutionary Mechanism (based on an extension to homeostasis) linking Stationary-Phase Mutations to the Baldwin Effect.
http://members.aol.com/jorolat/TEM.html

'Where Darwin meets Lamarck?' Discussion Egroup
http://groups.yahoo.com/group/evomech
0
John
7/26/2004 6:42:24 PM
On Mon, 26 Jul 2004 02:23:01 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
> 09:36:52 -0600, Sylvan Butler <ZsdbUse1+noZs_0407@Zbigfoot.Zcom.invalid>
> wrote:
>
>>On Fri, 23 Jul 2004 16:39:51 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>>> systems and take the time to rewrite his code to support only current systems,
>>
>>No, the "writing a good program" problems with winsock went away with
>>the arrival of 32 bit windows (NT 3.1 and Windows 95, but not Win32s).
>>Granted, the actual winsock itself wasn't so great back then, but that
>>improvement happened over time, and totally transparent to all the code
>>I've written since 1994.
>
> My own experience is that the original Winsock code in Windows 95 was pretty
> bad.

Yup.  That's why I said:

>>Granted, the actual winsock itself wasn't so great back then, but that

And since if you used the API implemented by that winsock, your program
would still run today, and even run better, I said:

>>[winsock] improvement happened over time, and totally transparent to
>>all the code I've written since 1994.

(I've never using the winsock2 API, no need unless you wanted to speak
over something besides IP like IPX or netbeui.)

>>I don't consider that old of systems to be "current" and I doubt there
>>is any significant consumer base left who could not run code targetted
>>to that minimum.
>
> You might well be surprised at how people still run Pegasus on ancient
> versions of Windows.

As long as they are using any win9x, or nt4 or later, pegasus should
work better than it does.  If DH is trying to support NT before 4.0
(doesn't appear to be) or Win32s (doesn't appear to be) or any 16 bit
windows (doesn't appear to be) then I'd be suprised.  But it would still
be fairly simple for pegasus to work better than it does.

sdb
0
Sylvan
8/8/2004 12:13:19 AM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <slrnchas0v.jos.ZsdbUse1+noZs_0408@knoppix.internal> on Sat, 7 Aug 2004
18:13:19 -0600, Sylvan Butler <ZsdbUse1+noZs_0408@Zbigfoot.Zcom.invalid>
wrote:

>On Mon, 26 Jul 2004 02:23:01 GMT, John Navas <spamfilter0@navasgroup.com> wrote:

>> My own experience is that the original Winsock code in Windows 95 was pretty
>> bad.
>
>Yup.  That's why I said:
>
>>>Granted, the actual winsock itself wasn't so great back then,

There's a world of difference between "pretty bad" and "wasn't so great".

>but that
>
>And since if you used the API implemented by that winsock, your program
>would still run today, and even run better, ...

Actually pretty much the same, and that's the issue.

>> You might well be surprised at how people still run Pegasus on ancient
>> versions of Windows.
>
>As long as they are using any win9x, or nt4 or later, pegasus should
>work better than it does.

Pretty much the same.

>If DH is trying to support NT before 4.0
>(doesn't appear to be) or Win32s (doesn't appear to be) or any 16 bit
>windows (doesn't appear to be) then I'd be suprised.  But it would still
>be fairly simple for pegasus to work better than it does.

Doable, but hardly "simple."

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
8/10/2004 5:08:58 AM
On Tue, 10 Aug 2004 05:08:58 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>>If DH is trying to support NT before 4.0
>>(doesn't appear to be) or Win32s (doesn't appear to be) or any 16 bit
>>windows (doesn't appear to be) then I'd be suprised.  But it would still
>>be fairly simple for pegasus to work better than it does.
>
> Doable, but hardly "simple."

Nah.

I've been doing winsock development since before trumpet winsock was in
beta and programming berkeley sockets before that.  I'll stand by
"fairly simple".  16 bit winsock support is a pain, but still fairly
simple to make a proper performing winsock app.  And in the 32 bit
winsock world, it is simple.

There really is no excuse for a 32bit winsock application to soak up the
CPU while waiting for network responses.

sdb
0
Sylvan
8/18/2004 6:16:41 AM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <slrnci5t29.trp.ZsdbUse1+noZs_0408@knoppix.internal> on Wed, 18 Aug 2004
00:16:41 -0600, Sylvan Butler <ZsdbUse1+noZs_0408@Zbigfoot.Zcom.invalid>
wrote:

>On Tue, 10 Aug 2004 05:08:58 GMT, John Navas <spamfilter0@navasgroup.com> wrote:
>>>If DH is trying to support NT before 4.0
>>>(doesn't appear to be) or Win32s (doesn't appear to be) or any 16 bit
>>>windows (doesn't appear to be) then I'd be suprised.  But it would still
>>>be fairly simple for pegasus to work better than it does.
>>
>> Doable, but hardly "simple."
>
>Nah.
>
>I've been doing winsock development since before trumpet winsock was in
>beta and programming berkeley sockets before that.  I'll stand by
>"fairly simple".

I likewise have relevant experience, and I'll stand by, "doable, but hardly
'simple.'"

>16 bit winsock support is a pain, but still fairly
>simple to make a proper performing winsock app.  And in the 32 bit
>winsock world, it is simple.
>
>There really is no excuse for a 32bit winsock application to soak up the
>CPU while waiting for network responses.

See:

-Z: The feature control flag

1024	Use blocking sockets; may be needed for some
	defective WINSOCK implementations.

4096	Use blocking WINSOCK Name Resolution calls; may be
	needed by some defective WINSOCK implementations.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
8/18/2004 4:01:10 PM
John Navas (spamfilter0@navasgroup.com) wrote in comp.mail.pegasus-mail.ms-windows:
> >There really is no excuse for a 32bit winsock application to soak up the
> >CPU while waiting for network responses.
> 
> See:
> 
> -Z: The feature control flag
> 
> 1024	Use blocking sockets; may be needed for some
> 	defective WINSOCK implementations.
> 
> 4096	Use blocking WINSOCK Name Resolution calls; may be
> 	needed by some defective WINSOCK implementations.

Unlike 16-bit Windows where dial-up systems each provided their own Winsock,
there is *one* Winsock on true 32-bit Windows (NT, 2K, XP, Server 2003),
and it is built into the OS.

If blocking sockets needed to be enabled for PM on 32-bit Windows, we would
all have to do it.

-- 
Jeff Rife        |  
SPAM bait:       | http://www.nabs.net/Cartoons/Peanuts/TenPin.gif 
AskDOJ@usdoj.gov |  
spam@ftc.gov     |  
0
Jeff
8/19/2004 2:56:19 AM
[POSTED TO comp.mail.pegasus-mail.ms-windows - REPLY ON USENET PLEASE]

In <MPG.1b8de0d9398d32c09897bc@news.nabs.net> on Wed, 18 Aug 2004 22:56:19
-0400, Jeff Rife <wevsr@nabs.net> wrote:

>John Navas (spamfilter0@navasgroup.com) wrote in comp.mail.pegasus-mail.ms-windows:
>> >There really is no excuse for a 32bit winsock application to soak up the
>> >CPU while waiting for network responses.
>> 
>> See:
>> 
>> -Z: The feature control flag
>> 
>> 1024	Use blocking sockets; may be needed for some
>> 	defective WINSOCK implementations.
>> 
>> 4096	Use blocking WINSOCK Name Resolution calls; may be
>> 	needed by some defective WINSOCK implementations.
>
>Unlike 16-bit Windows where dial-up systems each provided their own Winsock,
>there is *one* Winsock on true 32-bit Windows (NT, 2K, XP, Server 2003),
>and it is built into the OS.

There is also an official Microsoft Winsock for certain versions of 16-bit
Windows.

>If blocking sockets needed to be enabled for PM on 32-bit Windows, we would
>all have to do it.

It of course depends on how Pegasus was programmed.  The documentation above
is about using blocking versus non-blocking Winsock calls in general, not a
given Winsock.  This is clearly legacy code.

Is there some good reason to keep harping on this non-issue?  I think there
are lots of more important things on David's to-do list.

-- 
Best regards,
John Navas     <http://navasgrp.home.att.net/>
WinPMail helpers (PGP/URLPROXY/NsProto/AdrCSV) available on my Home Page
0
John
8/19/2004 2:33:25 PM
I dunno if this is what you're referring to, or you mean just back on
W3.x, but I'm using W95 with PM 4.02a and I get 100%CPU usage during file
send/receive.  It'll still do other things, but that's what FreeMeter
shows for CPU usage.
Is this normal?

On Sat, 7 Aug 2004 18:13:19 -0600, Sylvan Butler
<ZsdbUse1+noZs_0408@Zbigfoot.Zcom.invalid> spewed:
>>>> systems and take the time to rewrite his code to support only current systems,
>>>
>>>No, the "writing a good program" problems with winsock went away with
>>>the arrival of 32 bit windows (NT 3.1 and Windows 95, but not Win32s).
>>>Granted, the actual winsock itself wasn't so great back then, but that
>>>improvement happened over time, and totally transparent to all the code
>>>I've written since 1994.
>>
>> My own experience is that the original Winsock code in Windows 95 was pretty
>> bad.
>
>Yup.  That's why I said:
>
>>>Granted, the actual winsock itself wasn't so great back then, but that
>
>And since if you used the API implemented by that winsock, your program
>would still run today, and even run better, I said:
>
>>>[winsock] improvement happened over time, and totally transparent to
>>>all the code I've written since 1994.
>
>(I've never using the winsock2 API, no need unless you wanted to speak
>over something besides IP like IPX or netbeui.)
>
>>>I don't consider that old of systems to be "current" and I doubt there
>>>is any significant consumer base left who could not run code targetted
>>>to that minimum.
>>
>> You might well be surprised at how people still run Pegasus on ancient
>> versions of Windows.
>
>As long as they are using any win9x, or nt4 or later, pegasus should
>work better than it does.  If DH is trying to support NT before 4.0
>(doesn't appear to be) or Win32s (doesn't appear to be) or any 16 bit
>windows (doesn't appear to be) then I'd be suprised.  But it would still
>be fairly simple for pegasus to work better than it does.
>
>sdb

0
see
8/23/2004 8:56:47 AM
On Sat, 24 Jul 2004 18:03:50 -0600, Sylvan Butler
<ZsdbUse1+noZs_0407@Zbigfoot.Zcom.invalid> spewed:
>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>100% whenever the program was downloading - does this problem still
>
>> It does still exist, but why would that be a significant problem?  Whether
>> Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP) will still
>> multitask.
>
>Multitasking isn't the issue.  The issue is sluggish response.
>
>It isn't a HUGE problem because of dynamic scheduling priority in XP,
>but it is a SIGNIFICANT problem because XP is still noticably more
>sluggish when an app does the 100% CPU spin.
>
>I find it very annoying to have pegasus check mail in the background
>when it hits the 100% CPU.  My solution was to use hamster to pull down
>the mail, and pegasus talks locally to hamster.  Now with huge bandwidth
>and no network problems, the 100% CPU never lasts long enough to notice.
>
Hamster?  What is it, and any use on a single user system?

0
see
5/4/2006 1:34:59 AM
On 26 Jul 2004 17:11:47 GMT, "Thomas. Stephenson"
<mercury@mtshasta.lmms.lmco.com> spewed:
>>>It does still exist, but why would that be a significant problem? 
>>>Whether Pegasus is in a spin loop or not, a decent OS (e.g., Windows XP)
>>>will still multitask. 
>>>
>>>>When I tried to use Pegasus about a year ago the cpu usage went up to
>>>>100% whenever the program was downloading - does this problem still
>>>>exist? 
>>>>
>> When the CPU usage goes up so does the CPU temperature (I use VCool &
>> MBM)
>
>Try slowing down the program with the -z 1024 command line option.  FWIW, the 
>100% CPU does not exuate to 100% CPU loading at all, it just means there is no  
>CPU time given to the idle processor.  My CPU runs 100% all the time and there 
>is no affect on any foreground process.
>
>1024	Use blocking sockets; may be needed for some
>	defective WINSOCK implementations.

What exactly does this do?  How do you know if your winsock is defective?
Or, are you being a little fasicious and referring to ALL M$
implementations ;)

0
see
5/4/2006 1:35:11 AM
Reply: