Horizontal Video Streaks followed by Blue Screen

  • Thread starter Thread starter Bill Eversole
  • Start date Start date
B

Bill Eversole

Hi,

I am working on a Dell Dimension 5150. It is running Windows XP Home
Edition SP3. It has the Intel onboard graphics, Pentium 4 processor,
and 2 512 MB dual in-line memory modules.

First I replaced the hard drive as Dell diagnostics reported some bad
sectors. Now that that problem has been fixed I discovered a new
problem: Horizontal Streaks appear in the display soon followed by
Blue Screen Stop Error: 0x0000008E (0xc000001D, 0xBF87A3F1,
0xAA9F432030x00000000) win32k.sys-address BF87A3F1 base at BF800000,
DateSTamp 48025f2a

These streaks were evident during the XP setup and the Dell
Diagnostics, so I know it is not a driver problem.

Two of the video tests in Dell Diagnostics failed with this error
code: 5300:061F

Right now I am experimenting. I have removed the bottom 512 MB
memory module from the system, and so far I have not seen any
horizontal streaks or any blue screen stop errors. Maybe it's too
soon too say for sure, but could the problems be caused by a bad
memory module--even though it passed the memory diagnostics? Doesn't
onboard video use system memory?

Thanks in advance.

Bill
 
Bill said:
Hi,

I am working on a Dell Dimension 5150. It is running Windows XP Home
Edition SP3. It has the Intel onboard graphics, Pentium 4 processor,
and 2 512 MB dual in-line memory modules.

First I replaced the hard drive as Dell diagnostics reported some bad
sectors. Now that that problem has been fixed I discovered a new
problem: Horizontal Streaks appear in the display soon followed by
Blue Screen Stop Error: 0x0000008E (0xc000001D, 0xBF87A3F1,
0xAA9F432030x00000000) win32k.sys-address BF87A3F1 base at BF800000,
DateSTamp 48025f2a

These streaks were evident during the XP setup and the Dell
Diagnostics, so I know it is not a driver problem.

Two of the video tests in Dell Diagnostics failed with this error
code: 5300:061F

Right now I am experimenting. I have removed the bottom 512 MB
memory module from the system, and so far I have not seen any
horizontal streaks or any blue screen stop errors. Maybe it's too
soon too say for sure, but could the problems be caused by a bad
memory module--even though it passed the memory diagnostics? Doesn't
onboard video use system memory?

Thanks in advance.

Bill

It depends on the memory test used, as to whether the entire memory
is tested or not. If you use memtest86+, the executable "lifts itself"
out of the way, and the test also runs where the code once sat. So the
test is thorough from that point of view. But the other thing to know
about the test, is it consults the BIOS regarding "reserved" memory
sections. It could well be, that the shared memory used by onboard
graphics, doesn't get tested. And so it has never been tested.
(Otherwise, the memory test would end up drawing stuff on the screen.)

Perhaps entering the BIOS and reducing the shared memory setting a
bit, would allow testing more of the memory.

Or, installing a real graphics card, so that shared memory is no longer
needed.

In any case, I think your test case of removing some memory and
seeing the problem disappear, is highlighting that the problem is
related to memory.

Paul
 
It depends on the memory test used, as to whether the entire memory
is tested or not. If you use memtest86+, the executable "lifts itself"
out of the way, and the test also runs where the code once sat. So the
test is thorough from that point of view. But the other thing to know
about the test, is it consults the BIOS regarding "reserved" memory
sections. It could well be, that the shared memory used by onboard
graphics, doesn't get tested. And so it has never been tested.
(Otherwise, the memory test would end up drawing stuff on the screen.)

Perhaps entering the BIOS and reducing the shared memory setting a
bit, would allow testing more of the memory.

Or, installing a real graphics card, so that shared memory is no longer
needed.

In any case, I think your test case of removing some memory and
seeing the problem disappear, is highlighting that the problem is
related to memory.

    Paul

Paul,

Your theory about shared memory not being tested sounds very plausible
to me. I took the computer back to my client yesterday minus the
questionable DIMM. The client will see how it does for now running on
one 512MB DIMM. If it does OK, but needs the extra RAM, I can order a
new 512 MB DIMM later for less than $20 from Crucial.

Crucial tech recommended that I ditch the other DIMM as well and
install a new pair of DIMMs for the fastest speed and guaranteed
compatiblity. Will my client really notice any difference in word
processing and Internet browsing with an identical pair?

Thanks.

Bill
 
Bill said:
Paul,

Your theory about shared memory not being tested sounds very plausible
to me. I took the computer back to my client yesterday minus the
questionable DIMM. The client will see how it does for now running on
one 512MB DIMM. If it does OK, but needs the extra RAM, I can order a
new 512 MB DIMM later for less than $20 from Crucial.

Crucial tech recommended that I ditch the other DIMM as well and
install a new pair of DIMMs for the fastest speed and guaranteed
compatiblity. Will my client really notice any difference in word
processing and Internet browsing with an identical pair?

Thanks.

Bill

Your 5150 is based on 945G and four slots taking DDR2 memory, according
to the Dell site.

Specs
http://support.dell.com/support/edocs/systems/dim5150/en/sm/specs1.htm#wp1052310

Motherboard picture (scroll down)
http://support.dell.com/support/edocs/systems/dim5150/en/sm/techov1.htm

I can give you the benefit of some experiments I did with a previous
generation of board. The board had integrated graphics, and dual channel
memory capability. On the desktop, I could detect the difference in the
update of desktop graphics, based on memory configuration. In other
words, on an 865G, if I ran in single channel mode, the graphics
updated slower, than if a proper dual channel configuration was used.
So in that case, it did make a difference (to me at least - if blindfolded
and put in front of the machine, I could tell you whether the single
channel or the dual channel config was in place). The machine in
question, was supposed to come with a dual channel memory configuration,
but the builder messed up (it was a pre-built being sent as a gift).

I can't tell you whether the 945G is superior in that respect, and
has better characteristics with respect to the memory subsystem. But
you might be able to detect a small difference. If I was the customer,
I'd spend the $20 and fix it.

Paul
 
Back
Top