| View previous topic :: View next topic |
| Author |
Message |
steve
Guest
|
 Posted: Wed Sep 10, 2003 10:57 am Post subject: I hate that....... |
 |
|
My Athlon 1.4GHz system normally does a WU in about 6:10 - 6:30 hours.
The last WU took 53:26.
I hate that.
Please remind me again why a WU would take that long? |
|
| |
|
 Back to top |
Carsten Mueller
Guest
|
| Posted: Wed Sep 10, 2003 1:45 pm Post subject: Re: I hate that....... |
|
|
steve wrote:
[quote]My Athlon 1.4GHz system normally does a WU in about 6:10 - 6:30 hours.
The last WU took 53:26.
I hate that.
Please remind me again why a WU would take that long?
[/quote]
I also have strange results on my machine lately. Normally (my first 1,500
WUs), my system would crunch sweet WUs in 3 hours, normal WUs in 3:45 and
VLAR WUs in 4:00. For 1 month now, the completion of sweet WUs takes 3:45,
VLARs take around 4:50. Strangely, "normal" AR WUs take up to 6 hours!!
Check my SetiQ chart:
http://home.arcor.de/cm66/archart.gif
I>ve checked CPU usage in TaskManager, it>s absolutely normal (~99%). And
since _CPU_ time is measured, even high CPU usage of other task should not
influence _CPU_ time. It just seems the CLI has become up to 50% slower. And
no, I don>t use 3.08, I>m still using CLI 3.03.
I>ve tried closing all other processes, but there has been no improvement.
I>m running the CLI on another machine (P4 2.66 @ 2.84) which returns
absolutely normal results.
Anyone having similar experiences? Any ideas what could be the cause?
The strange system:
P4-M 2.0
512 MB
XP Pro
CLI 3.03
Carsten
--
Carsten Müller
Mobile: +66-9-0699624
Voicemail/Fax: +49-89-1488-206172 |
|
| |
|
| Back to top |
Roger Halstead
Guest
|
| Posted: Wed Sep 10, 2003 10:21 pm Post subject: Re: I hate that....... |
|
|
On Wed, 10 Sep 2003 07:27:04 +0000 (UTC), "Nick M V Salmon"
<spam_dump@btinternet.com> wrote:
[quote]"steve" <steve@unowut.org> wrote
My Athlon 1.4GHz system normally does a WU in about 6:10 - 6:30 hours.
The last WU took 53:26.
I hate that.
Please remind me again why a WU would take that long?
There>s no reason for _any_ work unit to take that much longer than normal
without some problem somewhere - worst 'normal' case is a VLAR work unit
with a sub 0.1 angle range under a Windows '98 based O/S & they are 'only'
circa 50% slower than the average 0.417AR work unit. Windows NT based O/Ss
don>t suffer even that much slowdown for VLARs...
If you get one like that again then have a look in Task Manager (Win2k/XP)
or Iarsn>s TaskInfo (Win98/ME) to see what process is hogging all the
system>s resources. Occasionally some software or other starts hogging
resources for no reason - either killing & restarting that process or
rebooting will clear it...
[/quote]
I had 4 or 5 of these last this past Spring. The only thing that was
heavily using resources was the seti client. These were running on a
..19XP + Athlon. I let one run. It had been projecting 53 hours, but
that kept getting bumped up. Up around 58 hours it was saying 68...So
I flushed it and got a new WU. Over the next few weeks I had several
more. When a WU appeared to be taking longer than about 3X normal,
I>d flush it. All of those WUs were indicating a projected finish
time of some where in the 50 hours range.
I>ve never had one since.
[quote]BTW: I have never had even a ONE work unit take that much longer than normal
here - ie. 743% slower..! I currently have 23,600 results returned and
that figure is increasing at roughly 74 results/day from 7 machines, all
[/quote]
I>m running 4. 2 1900s, a 2000 and a 2800. All Athlon XP +. It>s been
so long since I checked I>m not sure just how many WUs get returned
per week... High 200s "I think". Some where between 1/3 and 1/2 of
what you are getting.
[quote]AMD. Longest (wallclock) time WUs I get at the moment are whilst I>m using
the 'TV on Demand' facility (Like a TiVo.) of my ATI All In Wonder Radeon
8500DV. (+30% or so above normal for that AR.) That one>s a real resource
hog & would probably take all 100% of a 1.4GHz Athlon...
[/quote]
All mine run at 100% most of the time. This one (the 2800) also gets
all my internet work, programming, and even FS9. I use the CLI with
default settings. The other programs including FS9 don>t appear to
even realize the Seti client is even there.
Roger Halstead (K8RI EN73 & ARRL Life Member)
www.rogerhalstead.com
N833R World>s oldest Debonair? (S# CD-2)
[quote]
Ciao...
[UK]_Nick...[/quote] |
|
| |
|
| Back to top |
Nick M V Salmon
Guest
|
| Posted: Wed Sep 10, 2003 10:43 pm Post subject: Re: I hate that....... |
|
|
"Carsten Mueller" <newsman@eishockeygott.de> wrote
[quote]I also have strange results on my machine lately. Normally (my first 1,500
WUs), my system would crunch sweet WUs in 3 hours, normal WUs in 3:45 and
VLAR WUs in 4:00. For 1 month now, the completion of sweet WUs takes 3:45,
VLARs take around 4:50. Strangely, "normal" AR WUs take up to 6 hours!!
Check my SetiQ chart:
http://home.arcor.de/cm66/archart.gif
I>ve checked CPU usage in TaskManager, it>s absolutely normal (~99%). And
since _CPU_ time is measured, even high CPU usage of other task should not
influence _CPU_ time. It just seems the CLI has become up to 50% slower.
And
no, I don>t use 3.08, I>m still using CLI 3.03.
I>ve tried closing all other processes, but there has been no improvement.
I>m running the CLI on another machine (P4 2.66 @ 2.84) which returns
absolutely normal results.
Anyone having similar experiences? Any ideas what could be the cause?
The strange system:
P4-M 2.0
512 MB
XP Pro
CLI 3.03
[/quote]
P4 can 'throttle' itself when it gets too hot, ie. slow itself down. This
last month in Europe has been HOT - has your P4-M been running hot
& self throttling..?
Ciao...
[UK]_Nick...
--
Nick M V Salmon Master Mariner MN(Retd.)
My four initials@btinternet.com |
|
| |
|
| Back to top |
Carsten Mueller
Guest
|
| Posted: Thu Sep 11, 2003 1:58 pm Post subject: Re: I hate that....... |
|
|
Nick M V Salmon wrote:
[quote]P4 can 'throttle' itself when it gets too hot, ie. slow itself down.
This last month in Europe has been HOT - has your P4-M been running
hot & self throttling..?
[/quote]
Damn. That could be it. It would explain inconsistent processing times at
constant CPU usage. Actually I live in Thailand (yes, a pretty hot country)
but my computer survived the hot season without any problems.
But checking with AIDA I just realised that my CPU temp is definitely higher
than it used to be a few months ago. But everytime I check the CPU clock
it>s at 2000 MHz.
Does any of you know a tool which can log CPU temp and clock to a file? This
would help me check on this.
Thanks.
Carsten |
|
| |
|
| Back to top |
Roger Halstead
Guest
|
| Posted: Thu Sep 11, 2003 11:07 pm Post subject: Re: I hate that....... |
|
|
On Wed, 10 Sep 2003 22:43:45 +0000 (UTC), "Nick M V Salmon"
<spam_dump@btinternet.com> wrote:
[quote]"Roger Halstead" <newsgroups@rogerhalstead.com> wrote
"Nick M V Salmon" <spam_dump@btinternet.com> wrote:
There>s no reason for _any_ work unit to take that much longer than
normal
without some problem somewhere - worst 'normal' case is a VLAR work unit
with a sub 0.1 angle range under a Windows '98 based O/S & they are
'only'
circa 50% slower than the average 0.417AR work unit. Windows NT based
snip
I had 4 or 5 of these last this past Spring. The only thing that was
heavily using resources was the seti client. These were running on a
.19XP + Athlon. I let one run. It had been projecting 53 hours, but
that kept getting bumped up. Up around 58 hours it was saying 68...So
I flushed it and got a new WU. Over the next few weeks I had several
more. When a WU appeared to be taking longer than about 3X normal,
I>d flush it. All of those WUs were indicating a projected finish
time of some where in the 50 hours range.
I>ve never had one since.
I>ve never seen a single one..! Weird, were all these on the same 1900+
machine..? I>d suspect some form of hardware fault or a corrupt SETI.exe &
go hunting myself...
You>ve jogged my memory a bit more.[/quote]
After the second *long* WU, I replaced the client exe. It ran about
three days and I had another long WU. This would not be the same
client as I use a script file to run successive clients in different
directories. Again, I just deleted the WU and downloaded another.
Again I ran into another long WU so I replaced the client exe. I
don>t recall replacing all of those exe files that had the long WUs,
but I may have.
I remember now that I did have one other machine that had one
extremely long WU in the same period. I do not remember if I also
replace the client exe on that one or not.
The odds are fairly high on the one machine that something had trashed
a spot on the HD. Those client exe files could have all been located
in the same area even though they are not sequential in the directory.
As the second machine had a problem during the same period it *might*
have been network noise, something trashing the HD, or corrupted WUs
on download. it could also have been coincidence.
Roger Halstead (K8RI EN73 & ARRL Life Member)
www.rogerhalstead.com
N833R World>s oldest Debonair? (S# CD-2)
[quote]
BTW: I have never had even a ONE work unit take that much longer than
normal
here - ie. 743% slower..! I currently have 23,600 results returned and
that figure is increasing at roughly 74 results/day from 7 machines, all
I>m running 4. 2 1900s, a 2000 and a 2800. All Athlon XP +. It>s been
so long since I checked I>m not sure just how many WUs get returned
per week... High 200s "I think". Some where between 1/3 and 1/2 of
what you are getting.
I>ve done 739 in the last ten days (Sept 1st 00.00 to Sept 11th 00.00)
according to SetiTeam - that>s 73.9/day or 517/week. I setup SetiTeam to
record on the 1st of each month so I can keep an eye on how I>m doing...
AMD. Longest (wallclock) time WUs I get at the moment are whilst I>m
using
the 'TV on Demand' facility (Like a TiVo.) of my ATI All In Wonder Radeon
8500DV. (+30% or so above normal for that AR.) That one>s a real
resource
hog & would probably take all 100% of a 1.4GHz Athlon...
All mine run at 100% most of the time. This one (the 2800) also gets
all my internet work, programming, and even FS9. I use the CLI with
default settings. The other programs including FS9 don>t appear to
even realize the Seti client is even there.
Heh, I up the priority to 'normal' on all my WinME machines to cut the 0.2%
'wasted' by the idle loop and even then SETI only causes a barely noticeable
slowdown opening up software Etc. I don>t change the priority on any of
the WinXP machines - all but the WinXP Pro machine lose roughly 2% against
the wallclock somewhere and I have no idea what causes it - irritating...
This Barton machine is running at about 3600+ speed (ie. 11x200 for 2400MHz)
with all very fast disks including a dual 80GB WD Caviar SE RAID0 for video
capture alone yet ATI>s MMC (MultiMediaCentre) takes up to 35% of the CPU
off SETI - a real resource hog..!
Ciao...
[UK]_Nick...[/quote] |
|
| |
|
| Back to top |
Stratcat®
Guest
|
| Posted: Sat Sep 13, 2003 7:47 am Post subject: Re: I hate that....... |
|
|
"Carsten Mueller" <newsman@eishockeygott.de> wrote in message
news:bjpiov$ltnr0$2@ID-200859.news.uni-berlin.de...
[quote]Nick M V Salmon wrote:
P4 can 'throttle' itself when it gets too hot, ie. slow itself down.
This last month in Europe has been HOT - has your P4-M been running
hot & self throttling..?
Damn. That could be it. It would explain inconsistent processing times at
constant CPU usage. Actually I live in Thailand (yes, a pretty hot
country)
but my computer survived the hot season without any problems.
[/quote]
IMHO, Nick>s suggestion is a valid one. When I radically OC, my P4C starts
throttling back @ ~60*C, not the 67 - 70*C commonly referred to in the Intel
specs. This error margin may be due to my monitoring utility (AsusProbe),
but I don>t know for sure.
The common cpu speed/spec utilities (WCPUID, CPU-Z, etc.,) will not
report 'throttling back. They simply report the result of your FSB*cpu
multiplier. When throttling back occurs, the clock speed remains the same,
but your cpu will start eliminating some cycles internally, ie., every 4th
or 5th one, etc. As temps increase the cpu will eliminate more cycles, up to
40 - 60%. IIRC. Meanwhile, your apps may still be running fine, though
slower (mine did). There>s a pretty good technical explaqnation of this on
the Intel site if you do some digging.
Pop open your task manager, and check out your cpu usage graphs. Under
heavy throttling down, you>ll notice that the green cpu usage line, which is
normally flat and clean at the top of the graph, around 100%, will be
fluctuating & spiking down, similar to a 'saw-tooth' wave in an O>scope, if
you>ve ever used one.
Also, in the taskmanger, under 'view', tick 'show kernel times'. It should
normally be a red, slightly wavy line, staying under the 1st green line
(20%), in the cpu graph(s), under normal usage, when not opening apps.
When I>ve overheated, my kernel graphing was extremely erratic and spiked
quite high (50 - 60%). At least these have been my experiences.
If you are overheating, and you use a utility that shows performance times,
like Setspy does, you might note the time when it>s running slow. Then shut
down for a significant period of time, and restart. If the time initially is
as quick as before the problem, but then later slows down, that would lead
me to believe you have an overheating problem.
I guess you could always toss the running laptop in the freezer for a few
minutes while the problem is occurring, and see if it>s faster when you take
it out, but I>d probably check the operating temp parameters for the HDD
first!!!
If you ARE overheating, time to clean up the 'ol fan dust filter(s), heat
sink(s), and check for proper fan operation or air flow obstructions.
Unfortunately, I>ve never owned a notebook, so I can>t help more.
HTH
--
Strat®
---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.516 / Virus Database: 313 - Release Date: 9/1/2003 |
|
| |
|
| Back to top |
|
