I did a test with a laptop (XP sp2 512Mbram 1.7GHz single core no HT)
which is similar to the OP systems and I saw the network test issue.
The SB and laptop are wired through a switch so minimal router issues
and no wireless.
Without pipelining enabled - network test was consiustent only 94-96%
This afternoon - for the first time since my last post - while playing
music the SB started to stutter.
Went to the PC where my nugget was using the CBEEBIES website. Closed
firefox, and SB3 was fine... A few minutes later I let him back on and
within seconds the mini-equaliser on the right hand
bpa, u know A LOT. impressive.
IvanSlade;229196 Wrote:
I agree (but I do not think it is computer based or related, Java et
al), and I have been having the same problems. (Do a search for
IvanSlade for what I have reported on other threads.) There is
something wrong. My priority on
I think the problem may be Firefox and the handling of multiple http
request in parallel which may be exacerbated by single core
processors.
Can you check whether pipelining is enabled:
Enter the URL about:config and then type in network.http to the
browser's filter function to identify the
bpa;229099 Wrote:
There is still the unexplained burst of I/O activity in the Slim.exe
properties screen dump you posted. The smaller peaks are most likely
screen refresh as they occur at 30 secs.
I have tried to recreate a similar shape burst with a standard
Slimserver and I cannot -
Triode;229101 Wrote:
I've not read the whole thread, but wanted to point out that the network
test plugin assumes the server cpu is relatively lightly loaded
(depending which version you are running if the server can't create
traffic at the defined rate it either assumes it did or assumes it
NeilF;229075 Wrote:
Can you explain how you come to that conclusion? If you look at the
shots you can see:-
a) There is over 100meg of free memory.
b) If you look at firefox, the memory usage is only changing by couple
of meg depending on page visited.
c) I can have a dozen other programs
Phil Leigh;229157 Wrote:
I know you don't believe this might be memory related, but I wonder if
you are not misinterpreting something. That 100Mb you think is free
isn't free. It is being used by XP. Try opening several large
programs and watch that number stay more or less the same - it may
IvanSlade;229196 Wrote:
I agree (but I do not think it is computer based or related, Java et
al), and I have been having the same problems. (Do a search for
IvanSlade for what I have reported on other threads.) There is
something wrong. My priority on S/server is set at high. That is on
OK I can see something strange going on. If you use process explorer and
monitor slim.exe, the IO rate shoots up from just above zero and
plateaus at about 10kb/s until you move away from the page when it
immediately drops down to near zero again.
This doesn't have any effect on MY server
Phil Leigh;229210 Wrote:
OK I can see something strange going on. If you use process explorer and
monitor slim.exe, the IO rate shoots up from just above zero and
plateaus at about 10kb/s until you move away from the page when it
immediately drops down to near zero again.
This doesn't have
The context switching of both slim and IE goes up when that page is
loaded - they are fighting each other for control of the CPU.
All this switching is going to mess up the level 2 (L2) cache
performance. Also, the ram and pagefile are going to be (potentially)
impacted, depending on how much of
Phil Leigh;229215 Wrote:
The context switching of both slim and IE goes up when that page is
loaded - they are fighting each other for control of the CPU.
All this switching is going to mess up the level 2 (L2) cache
performance. Also, the ram and pagefile are going to be (potentially)
Yes that's it - it is nothing to do with CPU usage - neither is using
very much...they are fighting for control.
The network test is CAUSING the problem...watch what happens to the
context switch count of slim.exe if you stop the test!.
Basically - by having that page open AND running the test
Actually, I've just re-read your original post. You were having
stuttering (occasionally?) with wireless streaming.
Are you still having this?
How are you wired up? - is the PC wired to the router and then wireless
to the SB?
Forget the network test for a moment - you don't need 5000kbs! (1500
Phil, What CPU do you have ?
I tried to reproduce the I/O burst and couldn't see it. I am wondering
whether dual core / Hyperthreading hides the problem.
--
bpa
bpa's Profile:
bpa;229251 Wrote:
Phil, What CPU do you have ?
I tried to reproduce the I/O burst and couldn't see it. I am wondering
whether dual core / Hyperthreading hides the problem.
Single core. AMD Sempron 2800+ 1.6Ghz, 1Gb of RAM.
I would expect that a dual core machine would exhibit very
In my system it is automatically run at system startup.
--
Phil Leigh
You want to see the signal path BEFORE it gets onto a CD/vinyl...it
ain't what you'd call minimal...
...SB3+TACT+Altmann+MF DACXV3/Linn tri-amped Aktiv 5.1 system and some
very expensive cables ;o)
My system is a Intel Prescott with HT.
I saw slimserver lockout behavior with a socketwrapper bug which would
only affect single core CPUs so I think this issues may be related to
single core systems.
The other factor in the socketwrapper bug was whether slim.exe was run
as a user process or
Phil Leigh;229241 Wrote:
Actually, I've just re-read your original post. You were having
stuttering (occasionally?) with wireless streaming.
Are you still having this?
How are you wired up? - is the PC wired to the router and then wireless
to the SB?
Forget the network test for a
Network test is very simple and was written without any support from the
client. All it does is send repeated screen updates to the player and
the appropriate rate over the slimproto tcp session.
The main difference between nettest and streaming actual data is the
size of the buffers being
Triode;229276 Wrote:
Network test is very simple and was written without any support from the
client. All it does is send repeated screen updates to the player at
the appropriate rate over the slimproto tcp session.
The main difference between nettest and streaming actual data is the
Where are you at now? Do you have stuttering or not?
SlimServer is not the problem here. If it was, many many people would
have the same issue.
If you do have stuttering, we should be able to help a bit,
--
Phil Leigh
You want to see the signal path BEFORE it gets onto a CD/vinyl...it
ain't
Phil Leigh;229289 Wrote:
Where are you at now? Do you have stuttering or not?
SlimServer is not the problem here. If it was, many many people would
have the same issue.
If you do have stuttering, we should be able to help a bit,
There's another thread of mine going through this...
I think
NeilF;229198 Wrote:
Had a look and couldn't see any posts of your similar to mine?
QUOTE]
Hi - Try Snap, Crackle and Pop.
Best
Ivan
--
IvanSlade
IvanSlade's Profile:
It's definately nothing to do with traffic!
In the logs over say 10-20 second periods I can see no communication
other than between my pc and the SB!
So what is it about just being on a particular website (eg:
http://www.bbc.co.uk/cbeebies/autumnwatch/game/) that reduces my
5000kbps test from
A couple of points.
1. If network traffic is between a PC app and 127.0.0.1 or the server
IP address and the App is on the same PC - it will not show up with
Ethereal. Slimserver uses AJAX a lot with Web interface and it doesn't
show.
2. If you are using task manager to see what is happening -
Doesn't do it on mine.
512Mb is no way enough to run XP IME.
You need at least 1Gb, preferably 2.
On another angle, Have you checked for viruses and spy/adware?
--
Phil Leigh
You want to see the signal path BEFORE it gets onto a CD/vinyl...it
ain't what you'd call minimal...
NeilF,
People keep pointing you in a different direction yet you are certain
it is the particular web pages, so what advice are we to give?
Don't go to the web pages :) There, fixed it :)
I would like to know about your computer system. What is running in the
background? Using process
brucegrr;229030 Wrote:
NeilF,
People keep pointing you in a different direction yet you are certain
it is the particular web pages, so what advice are we to give?
Don't go to the web pages :) There, fixed it :)
I would like to know about your computer system. What is running in the
bpa;228982 Wrote:
A couple of points.
1. If network traffic is between a PC app and 127.0.0.1 or the server
IP address and the App is on the same PC - it will not show up with
Ethereal. Slimserver uses AJAX a lot with Web interface and it doesn't
show.
2. If you are using task manager
Here are the screen shots from the previous post zipped so you can see
them full size...
+---+
|Filename: procexplorer2.zip|
|Download:
Your sig does not give enough hardware info to be helpful.
According to Process Manager it looks (and I could barely read the
.gif)that you have 5-7 programs running (either using or in background)
All of them use memory. Not talking about cpu usage here. Talking about
ram use. Big difference.
Looks like you are running low on memory and therefore page faulting to
me...
If I were you I'd stick another 512 Mb in there and see what happens.
I've just repeated (as best as I could) your problem on my system and
got no io spikes like yours. Sorry this isn't much help.
--
Phil Leigh
You
Since the problem doesn't happen on other systems and it happens with
both IE and Firefox - it is something about how your system is set up.
1. Have you tried seeing if the problem shows if you turn off Ad-Aware,
Grisoft Antivirus and other apps.
2. What plugins have you installed on Slimserver
Phil Leigh;229055 Wrote:
Looks like you are running low on memory and therefore page faulting to
me...
If I were you I'd stick another 512 Mb in there and see what happens.
I've just repeated (as best as I could) your problem on my system and
got no io spikes like yours. Sorry this isn't
bpa;229067 Wrote:
Since the problem doesn't happen on other systems and it happens with
both IE and Firefox - it is something about how your system is set up.
1. Have you tried seeing if the problem shows if you turn off Ad-Aware,
Grisoft Antivirus and other apps.
2. What plugins have
You must have a very special machine that can run WinXp Pro, 12 programs
and every is fine :)all in 512mg of ram.
No offense, but I don't buy it.
It very well could be something else.but if I was
troubleshooting a customer's system I would start with reducing the
programs are running,
It could be memory but there are a couple of points that don't feel
right fro memory
1. I/O for paging isn't usually attributed to a process so the slim.exe
I/O burst is odd.
2. I have a laptop XP sp2 with 512Mb and it becomes very obvious when
memory is short, reponse time drops so that
BPA,
Agreed.
I can't find where the OP mentioned what version of slimserver he is
running? Plugins?
--
brucegrr
brucegrr's Profile: http://forums.slimdevices.com/member.php?userid=3964
View this thread:
brucegrr;229079 Wrote:
You must have a very special machine that can run WinXp Pro, 12 programs
and every is fine :)all in 512mg of ram.
No offense, but I don't buy it.
It very well could be something else.but if I was
troubleshooting a customer's system I would start with
bpa;229086 Wrote:
It could be memory but there are a couple of points that don't feel
right fro memory
1. I/O for paging isn't usually attributed to a process so the slim.exe
I/O burst is odd.
2. I have a laptop XP sp2 with 512Mb and it becomes very obvious when
memory is short,
By the way, just to put this all back into context...
Here's me spending a short while at www.bbc.co.uk/cbeebies, and then a
short while at www.bbc.co.uk. Notice the two obvious different set of
results in the 5000kbps test...
10 :0 : 0%
20 :2 : 4% ##
30 : 21 : 46%
There is still the unexplained burst of I/O activity in the Slim.exe
properties screen dump you posted. The smaller peaks are most likely
screen refresh as they occur at 30 secs.
I have tried to recreate a similar shape burst with a standard
Slimserver and I cannot - the I/O burst has to be
I've not read the whole thread, but wanted to point out that the network
test plugin assumes the server cpu is relatively lightly loaded
(depending which version you are running if the server can't create
traffic at the defined rate it either assumes it did or assumes it
didn't - meaning it
I've been trying to work out why my SqueezeBox 3 stutters at times
seemingly with poor wireless speed at times. The answer? GMAIL! Yep!
GMAIL!!
Try this if you can if you have a wireless Squeezebox 3?
Run the 5000kbps network test. I get 100%. eg:-
10 :0 : 0%
20 :0 : 0%
NeilF;228845 Wrote:
I've been trying to work out why my SqueezeBox 3 stutters at times
seemingly with poor wireless speed at times. The answer? GMAIL! Yep!
GMAIL!!
Well, no, strictly any application will slow your system down. Windows,
IME, sucks at trying to be a server for SlimServer and
Windows can work fine IF your processor speed and ram is sufficient. I
have two squeezeboxes running right now and I am surfing the net and
doing email. No stutters.
The lower the processor speed and the less ram you have the more likely
it is to tax the system.
If yo uleave gmail open then it
gerph;228867 Wrote:
Well, no, strictly any application will slow your system down. Windows,
IME, sucks at trying to be a server for SlimServer and if you do other
things with the machine whilst you're trying to serve music you're
going to find that it takes away speed. A web browser running
I don't use Gmail but I did load your other page with no problem at
all.
Every action uses memory. What operating system are you using?
Processor speed? ram? Are you using the most current version of
Firefox?
I don't think gmail alone is the problem. Gmail is no different than
any other
brucegrr;228883 Wrote:
I don't use Gmail but I did load your other page with no problem at all.
Every action uses memory. What operating system are you using?
Processor speed? ram? Are you using the most current version of
Firefox?
I don't think gmail alone is the problem. Gmail is no
Install wireshark/Ethereal on the server PC and you can find out what is
being sent from the PC.
--
bpa
bpa's Profile: http://forums.slimdevices.com/member.php?userid=1806
View this thread:
NeilF wrote:
I can reproduce it at will...
Open a browser (FireFox or IE) say to bbc.co.uk, and run the 5000kbps
test in SlimServer... Get 100%.
Move to the 'cbeebies' page (http://www.bbc.co.uk/cbeebies/) results
drop...
Move back to bbc.co.uk, results return to 100%.
All the time
I just can't think of any way that Gmail would be affecting slimserver.
Is slimserver open in a window in Firefox when you access gmail? If so
, how about if you close the slimserver window(not slimserver just the
window) and access gmail? Same problem?
Not trying to give you a hard time here
brucegrr;228894 Wrote:
I just can't think of any way that Gmail would be affecting slimserver.
Is slimserver open in a window in Firefox when you access gmail? If so
, how about if you close the slimserver window(not slimserver just the
window) and access gmail? Same problem?
Not trying
brucegrr;228894 Wrote:
I just can't think of any way that Gmail would be affecting slimserver.
Hmm. GMail is AJAXified, so it's going to be constantly hitting the
network (unlike most pages).
--
SumnerH
SumnerH's
SumnerH;228901 Wrote:
Hmm. GMail is AJAXified, so it's going to be constantly hitting the
network (unlike most pages).
True... But I can be torrenting the hell out of the internet and that
has no effect on SlimServer?
--
NeilF
shuttle sn45g, athlon 3200+, 512mb, xp pro, squeeze box 3
bpa;228889 Wrote:
Install wireshark/Ethereal on the server PC and you can find out what is
being sent from the PC.
OK... I've installed it, and run it with FireFox on a page that does
doesn't affect Slimserver, but I can't make head or tail of what I'm
seeing in the list of stuff?
I've got
Comparing the capture when on bbc.co.uk (no problems) with
bbc.co.uk/cbeebies (problems) there are some red lines in the capture
with the later.
I suspect these are showing the problem happening as there seems to be
a clear delay shown?
See images attached!
First try to get a feel of what is happening such as
1. Give a colour to the packeto/from GMAIL
2. Look at the GMAIL traffic - are the packets very big or very often
compared to other traffic.
3. Have a look at selected GMAIL packets - use top panel to select a
packet and look in bottom panel
AFAICT The only traffic on those screen are to/from SB3 - you can tell
by the port number 3483 is either source or destination.
The dark lines are retransmission - it's not clear why
Where is the GMAIL traffic ?
--
bpa
61 matches
Mail list logo
