Copying cripples Windows

[Rob]

I thought he said that the disk light is solidly on during these
moments of 'freeze', unless I'm mixing two threads, which is entirely
possible.

ram will do that

 

[Rob]

As a hardware guy, your symptoms don't sound like hardware. Sorry.

If your system crashed, or threw up error dialog boxes, then,
we'd be talking hardware. But a sedate, subsystem limited issue
that you're experiencing, "smells like software".

Windows uses a bit of randomization, when loading applications into
memory. I'd expect some of the structures in Windows though, to have
a degree of consistency in terms of where they're located. But something
like a copy-paste buffer, that would be dynamically allocated (every
new copy, uses fresh memory), and the address of the memory used
would be constantly changing.

Since your symptoms are so "well behaved", I've seen nothing
yet that suggests removing RAM boards. The computer is running,
and could be fetching hundreds of millions of cache lines per
second from main memory. And since it does this without
flinching or fidgeting, "it's working" from a hardware
perspective.

If you really thought the problem was memory related, you have
the option of booting up a Linux LiveCD and testing behavior there.
Since the layout and memory usage are entirely different, the
symptoms seen might give you a second chance to identify
a hardware type issue.

Note that memtest86+ does not test the entire memory. It comes
very close. But, it's not perfect. In particular, it cannot
test the 640K region. Certain areas of memory are marked by the
BIOS as "reserved", and memtest86+ honors those reservations.
It can mean somewhere around one megabyte of memory might not get
tested.

To fix that:

1) Reinstall RAM in single channel mode.
2) Run memtest86+ for one pass.
3) If error free, swap the two sticks sitting on the single
channel, such that the high memory stick, becomes the low
memory stick, and vice versa. By rearranging the two sticks
sitting on a memory channel, you can achieve 100% memory
test coverage. Run memtest86+ again, after the RAM sticks
have been swapped.

I always test one stick at a time takes too long otherwise.

 

[Fokke Nauta]

As a hardware guy, your symptoms don't sound like hardware. Sorry.

If your system crashed, or threw up error dialog boxes, then,
we'd be talking hardware. But a sedate, subsystem limited issue
that you're experiencing, "smells like software".

Windows uses a bit of randomization, when loading applications into
memory. I'd expect some of the structures in Windows though, to have
a degree of consistency in terms of where they're located. But something
like a copy-paste buffer, that would be dynamically allocated (every
new copy, uses fresh memory), and the address of the memory used
would be constantly changing.

Since your symptoms are so "well behaved", I've seen nothing
yet that suggests removing RAM boards. The computer is running,
and could be fetching hundreds of millions of cache lines per
second from main memory. And since it does this without
flinching or fidgeting, "it's working" from a hardware
perspective.

If you really thought the problem was memory related, you have
the option of booting up a Linux LiveCD and testing behavior there.
Since the layout and memory usage are entirely different, the
symptoms seen might give you a second chance to identify
a hardware type issue.

Note that memtest86+ does not test the entire memory. It comes
very close. But, it's not perfect. In particular, it cannot
test the 640K region. Certain areas of memory are marked by the
BIOS as "reserved", and memtest86+ honors those reservations.
It can mean somewhere around one megabyte of memory might not get
tested.

To fix that:

1) Reinstall RAM in single channel mode.
2) Run memtest86+ for one pass.
3) If error free, swap the two sticks sitting on the single
channel, such that the high memory stick, becomes the low
memory stick, and vice versa. By rearranging the two sticks
sitting on a memory channel, you can achieve 100% memory
test coverage. Run memtest86+ again, after the RAM sticks
have been swapped.

Remember to put the sticks back in dual channel mode when
you're finished. If you have four DIMMs, you do the above
procedure twice, using a pair of DIMMs each time. A total
of four memory installation procedures.

HTH,
Paul

I will run the memtest86 in single channel mode. To do that I will have
to take my PC from underneath my desk onto a table. Quite some hassle.

I paid really good attention to what happened again this morning, so I
can now describe the problem in full detail.
After starting up my PC, I extend a range of cells in Excel 2003. It
looked like that went allright, but as soon as I try to do something in
Excel (fill in a cell or save) it freezes and that little circel keeps
turning around. Nothing I can do but kill Excel.
After restarting Excel the same thing happens again and again. In Excel
2010 as well.

OK, now I do something else. I open Notepad, type something in it, and
open an HTML editor. As soon as I copy the text in Notepad, the HTML
editor freezes. Not notepad. When I restart the HTML editor, I can paste
the copied text in it and it works fine. Excel now works fine as well.
The problem will not occur on this day again. Tomorrow morning again,
very likely.

Hardware? Likely, as this problem also happened in the time that this PC
was running XP.

Fokke

 

[Fokke Nauta]

I always test one stick at a time takes too long otherwise.

I will run this test with one stick at a time.

 

[Dave-UK]

I will run the memtest86 in single channel mode. To do that I will have
to take my PC from underneath my desk onto a table. Quite some hassle.

I paid really good attention to what happened again this morning, so I
can now describe the problem in full detail.
After starting up my PC, I extend a range of cells in Excel 2003. It
looked like that went allright, but as soon as I try to do something in
Excel (fill in a cell or save) it freezes and that little circel keeps
turning around. Nothing I can do but kill Excel.
After restarting Excel the same thing happens again and again. In Excel
2010 as well.

OK, now I do something else. I open Notepad, type something in it, and
open an HTML editor. As soon as I copy the text in Notepad, the HTML
editor freezes. Not notepad. When I restart the HTML editor, I can paste
the copied text in it and it works fine. Excel now works fine as well.
The problem will not occur on this day again. Tomorrow morning again,
very likely.

Hardware? Likely, as this problem also happened in the time that this PC
was running XP.

Fokke

This free program might give you a clue.
One of its features is a history tab. Try starting it and then create your problem.
Any slowdown should show up in the timings of the history tab.
http://systemexplorer.net/

 

[Fokke Nauta]

As I suggested before, I think I'd look for errors, such as in
Event Viewer first.

In the Event viewer I can see a series of Program hang errors. Event
ID=1002. It says: "The program EXCEL2003.EXE version 11.0.5612.0 stopped
interacting with Windows and was closed. To see if more information
about the problem is available, check the problem history in the Action
Center control panel."

In the Action Center there is nothing useful I can find.

Next, using Task Manager or Resource Monitor, I look for
high CPU usage, while things aren't working, to see if
there is evidence of a culprit.

There is no high CPU usage. The only thing I see in the Resource
monitor, is a red text with something like "Excel is not responding".
Everything else looks normal.

 

[Fokke Nauta]

This free program might give you a clue.
One of its features is a history tab. Try starting it and then create
your problem.
Any slowdown should show up in the timings of the history tab.
http://systemexplorer.net/

Thanks, interesting application.
I just installed it. I'll wait for this problem to reoccur.

 

[Rob]

On 23/12/2012 11:16 PM, Fokke Nauta wrote:


out of interest what MB/CPU/memory ?

 

[Paul]

Hardware? Likely, as this problem also happened in the time that this PC
was running XP.

Fokke

OK, and that's the kind of evidence you're looking for.

Anything that points to hardware specifically.

Normally, bad RAM would give blue screens, hex dumps, that sort
of thing.

But in your case, only certain operations fail, and it isn't obvious
that it is RAM.

But if your previous OS also did the same thing, that makes
it more likely to be hardware.

You can also boot a Linux LiveCD, and gather symptoms there
as well. Such as start gedit, copy and paste in there, and
see if it freezes. Then try "dmesg" in another window
and see if anything interesting appears near the end of the log.

Paul

 

[Gene E. Bloch]

When I heard of Ditto I decided to give it a try but my 64-bit Win 7 Pro
PC started behaving oddly. One look at Ditto and I didn't like the
layout either. So I immediately uninstalled the damn thing which wasn't
showing up as one of the installed programs in START> CONTROL PANEL>
PROGRAMS AND FEATURES either which I thought was odd. Can't remember now
how I undeleted it but I certainly did not use the RESTORE facility.
Can't remember whether DITTO had its own UNINSTALL facility.

All programs have their own install facility - or none. Windows doesn't
uninstall a program, it calls the program's own facility.

It would seem that Ditto didn't register it with Windows, so you'd have
to go to the program files folder and start the uninstaller, or perhaps
it was a menu entry in Ditto.

Or maybe you had to find a program on the web, like some antivirus
programs require (on their own site).

 

[Fokke Nauta]

OK, and that's the kind of evidence you're looking for.

Anything that points to hardware specifically.

Normally, bad RAM would give blue screens, hex dumps, that sort
of thing.

But in your case, only certain operations fail, and it isn't obvious
that it is RAM.

But if your previous OS also did the same thing, that makes
it more likely to be hardware.

You can also boot a Linux LiveCD, and gather symptoms there
as well. Such as start gedit, copy and paste in there, and
see if it freezes. Then try "dmesg" in another window
and see if anything interesting appears near the end of the log.

Paul

I booted up with a Gentoo Live CD. Didn't had any problems with copying
and pasting in gedit. I ran dmesg. You can find the results in
www.pc3.nl/test. The file is called dmesg.txt. I don't know where to
look for. The end of the log file shows an error: "cannot submit urb 0,
error -28: not enough bandwidth". Googling on this leads to USB audio
issues. Has nothing to do with my problem.

Fokke

 

[Fokke Nauta]

Hi all,

W7 Pro 64b.
When I start my PC to get working, the first thing I do is copy a few
cells in Excel. Once I've done that, Excel freezes and there's nothing I
can do but kill the application. It happens in Excel 2003 as well as 2010.
When I copy tekst from Word to Notepad, the same thing happens and I can
kill Notepad. When I copy tekst from Notepad into an HTML editor, the
same thing happens again and I can kill the HTML editor.
This can go on for minutes and all of a sudden it works again, and I can
work for the rest of the day.
This happens very frequently and is very annoying.

I don't think it's Windows, as it also happened a few times in the days
that my PC was running with XP. But now it happens very frequently.

What can it be? Memory problem? CPU?
Any recognizes this or has an idea as to what it may cause?

Thanks in advance.

Fokke

Thanks for all hints and tips. I have not been able to check the mem
modules yet but I will. I will come back on this.
In the mean time I have learned to live with this.
What I do:
- Start an HTML editor,
- Start notepad and type some text in it,
- Copy the text and paste it in the HTML editor,
- This one freezes. I kill it and start it again,
- Paste the text again. Now it works and for the rest of the day
everything works fine.

This has become a ritual I have to perform almost every morning I start
to work with my PC.
I can live with it during the time I have not found a solution.

Fokke

 

[Paul]

I booted up with a Gentoo Live CD. Didn't had any problems with copying
and pasting in gedit. I ran dmesg. You can find the results in
www.pc3.nl/test. The file is called dmesg.txt. I don't know where to
look for. The end of the log file shows an error: "cannot submit urb 0,
error -28: not enough bandwidth". Googling on this leads to USB audio
issues. Has nothing to do with my problem.

Fokke

OK, now go back to Windows and disconnect that USB audio device

The USB audio device could be interfering with USB HID (mouse/keyboard)
devices. Just a wild guess.

Part of testing, is to "simplify" things. Removing your
USB audio device temporarily, would be an example of that
approach.

Paul