intel chip, intel board... but no dice. help?

[Quartz]

Can someone help me try to figure out what's up with this?



I have an intel D815EEA mobo, which has been running a 600EB P3 (flip
chip 370) for some time and everything works fine. I recently picked up
a cheap 1Ghz P3 to make the machine a little more usable.

The problem is, although the 1G passes through the bios and stuff fine,
the machine reboots itself right before XP is about to load (right
before the logo would pop up). Swapping back in the 600EB everything
works fine...


For reference, the chip says:

1000/256/133/1.75v
L123A723-0243
PENTIUM III SL52R


... which ought to work according to...
support.intel.com/support/motherboards/desktop/d815eea/sb/cs-013068.htm
... right? Am I reading something wrong here? Is the chip just bad or
something? I'm using the latest bios version (p11) and there are no
cpu-speed jumpers on the board that I can see.... anyone have any ideas?

____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[kony]

I have an intel D815EEA mobo, which has been running a 600EB P3 (flip
chip 370) for some time and everything works fine. I recently picked up
a cheap 1Ghz P3 to make the machine a little more usable.

The problem is, although the 1G passes through the bios and stuff fine,
the machine reboots itself right before XP is about to load (right
before the logo would pop up). Swapping back in the 600EB everything
works fine...

Are you certain the heatsink is on good, that it wasn't
overheating? If you had tried reinstalling it multiple
times, that's probably not the problem but one time, maybe.

It should use more power too, is your power supply adequate?
You might take voltage readings. Also check it with
memtest86+. IIRC, the "EB" P3 was a 133MHz FSB CPU, so it
shouldn't be an issue of a higher memory bus speed from the
new CPU, though if you had changed no jumpers or bios
settings, we'd be assuming the speed but perhaps not proven
yet? Memtest86+ will also show the operational speed of the
CPU, not just what it's "supposed" to be.

For reference, the chip says:

1000/256/133/1.75v
L123A723-0243
PENTIUM III SL52R


.. which ought to work according to...
support.intel.com/support/motherboards/desktop/d815eea/sb/cs-013068.htm
.. right? Am I reading something wrong here? Is the chip just bad or
something? I'm using the latest bios version (p11) and there are no
cpu-speed jumpers on the board that I can see.... anyone have any ideas?

You could try clearing CMOS. I

 

[Quartz]

Are you certain the heatsink is on good, that it wasn't
overheating?

yeah, that was the first thing everyone thought. but after a couple
reboots when I reach in touch the sink, it's only slightly warm, and if
I shut the machine off and pull the chip out it's not even warm at all.

It should use more power too, is your power supply adequate?

it certainly ought to be. I forget which is in there right now, but it's
somewhere in the 300-500 range. as for cards, I've got a geforce 4
mx440, a scsi card, and a firewire card (which has no devices plugged
into it).... I mean, does the 1g use like 100 watts or something?

You might take voltage readings. Also check it with
memtest86+. IIRC, the "EB" P3 was a 133MHz FSB CPU, so it
shouldn't be an issue of a higher memory bus speed from the
new CPU, though if you had changed no jumpers or bios
settings, we'd be assuming the speed but perhaps not proven
yet?

it's been running with that ram for a while with no problems, I doubt
that's the problem.

Memtest86+ will also show the operational speed of the
CPU, not just what it's "supposed" to be.

well, the bios shows it as a 1g 133 chip and gives me all the stepping
foo and such, I'd assume that it would cycle it down or something if
it's not supported at that speed.

You could try clearing CMOS. I

is there a reliable way to do this from the hardware side? the bios only
has one 'reset settings' type option which I already tried... would
yanking the battery do anything for me?


____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[kony]

yeah, that was the first thing everyone thought. but after a couple
reboots when I reach in touch the sink, it's only slightly warm, and if
I shut the machine off and pull the chip out it's not even warm at all.

If heatsink wasn't making good contact, it would not be
overly warm. A CPU core is rather low mass, partial
heatsink contact would cool it off quite quickly, such that
it didn't feel hot by the time you could get the heatsink
off to check it. Even so, this is no further evidence of an
actual problem and with the P3 having integral thermal
shutoff, I would doubt it's overheating, especially since
one can pull the 'sink and see if there was a clear imprint
of the CPU core. In some cases (with aluminum based 'sink
and optimal/minimal heatsink compound) you can even see the
imprint of the corners of the CPU on the 'sink, meaning it
must be making good contact unless the heatsink itself was
horribly irregular on the bottom.

it certainly ought to be. I forget which is in there right now, but it's
somewhere in the 300-500 range. as for cards, I've got a geforce 4
mx440, a scsi card, and a firewire card (which has no devices plugged
into it).... I mean, does the 1g use like 100 watts or something?

Wattage rating means less if it's a generic. I've a 400W
generic (sitting unused) that is inferior to the typical
name-brand 250W. An upper end P3 might use 40W, but if your
PSU can actually deliver 150W combined 3V/5V, (for whole
system use of course) then it should be ok... but is still
something to investigate, take voltage readings.

it's been running with that ram for a while with no problems, I doubt
that's the problem.


well, the bios shows it as a 1g 133 chip and gives me all the stepping
foo and such, I'd assume that it would cycle it down or something if
it's not supported at that speed.

Probably right then, though 'something' is wrong, best to
get another confirmation when the opportunity arises, IMO.

is there a reliable way to do this from the hardware side? the bios only
has one 'reset settings' type option which I already tried... would
yanking the battery do anything for me?

Disconnect AC to the PSU, then either pull the battery for
~10 minutes or use the clear-cmos jumper (should be shown in
your manual or look around near the battery, though
sometimes further away from it).

 

[Quartz]

XP is known for not working with different chipsets. It is likely that XP is
trying to load the wrong microcode into the CPU, hence the reboot.

hmm... well that's interesting. I'll have to check that.


____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[Quartz]

[heatsink]

hmm... well, lemme buy some more thermal grease and slap it on just to
make sure.

Wattage rating means less if it's a generic.

yeah I know, but what I'm getting at is that I've never had problems
with this before when taxing the system, so...

this mobo also has onboard voltage sensor readings, they've never shown
anything to be wrong before, should I trust them?

Probably right then, though 'something' is wrong, best to
get another confirmation when the opportunity arises, IMO.

yeah... I was hoping someone had a simple answer, but it looks like I
just have to run down the list and check things one by one.

Disconnect AC to the PSU, then either pull the battery for
~10 minutes or use the clear-cmos jumper (should be shown in
your manual or look around near the battery, though
sometimes further away from it).

err... there is a distinct lack of labeled jumpers on this board, other
than the 'special bios config' one.

____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[kony]

[heatsink]

hmm... well, lemme buy some more thermal grease and slap it on just to
make sure.

Wattage rating means less if it's a generic.

yeah I know, but what I'm getting at is that I've never had problems
with this before when taxing the system, so...

While that it a reasonable presumption, now your faster CPU
has a higher PSU load. With that in mind, it's still among
possible culprits.

this mobo also has onboard voltage sensor readings, they've never shown
anything to be wrong before, should I trust them?

yes/no/maybe.

you can trust then within the context of having trusted them
previously? If you want an accurate voltage reading, you
will have to use a multimeter to measure at the connector to
the load (for this purpose, a drive connector, or
motherboard connector, etc), BUT, the deviation of an
onboard sensor is somewhat linear with amperage such that if
it were off and you were aware of it, you can still use the
software report to note a relative change in voltage.

yeah... I was hoping someone had a simple answer, but it looks like I
just have to run down the list and check things one by one.

Well the simple answer is it should work. Who knows, maybe
you received that one CPU out of a million that's defective
in some way? Odds are against it, but that doesn't help
that 1 person figure out what's going on.

err... there is a distinct lack of labeled jumpers on this board, other
than the 'special bios config' one.

Special bios config? Odd, try it.

 

[kony]

hmm... well that's interesting. I'll have to check that.

Since you continue to use the same board, you're using the
same chipset. It's even the same family of CPU, just one
with a higher multiplier and "Maybe" a different stepping.
XP should not be the problem unless there was already some
kind of timing bug that only surfaced due to the higher
performing CPU.

 

[Kadaitcha Man]

Since you continue to use the same board, you're using the
same chipset.

Different CPU, you pathetically stupid, snotfucked moron. The original XP
installation is highly likely to be laoding the wrong CPU microcode, whcih,
you brain-dead shitstain, is a CHIP. I'd say grow a brain, but ****ing
fungus isn't known to be able to do such a thing.

Irrespective of that, XP must be eliminated as the cause first. You can ****
right off and die now, you stupid ****.

 

[Quartz]

It would be
good to try booting safemode

no dice on that either, I don't get any of the text on the screen, it
instantly reboots (same as if I try to boot normally)

I don't have any other OSs handy at the moment... (and I dont' really
feel like messing with linux). lemme ask someone for some other disc and
I'll get back to you on that.

____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[Quartz]

While that it a reasonable presumption, now your faster CPU
has a higher PSU load. With that in mind, it's still among
possible culprits.

bleh.... I don't really feel like going out and dropping $50 on a new
psu...

If you want an accurate voltage reading, you
will have to use a multimeter to measure at the connector to
the load (for this purpose, a drive connector, or
motherboard connector, etc),

speaking of, can I just needle an unused drive power plug, or do I need
to find the mains on the actual mobo itself? (I have no idea how to go
about doing any of this)

Who knows, maybe
you received that one CPU out of a million that's defective
in some way?

I bought it second-hand from a guy who had about 50 of them, he said
they were all tested and working.

Special bios config? Odd, try it.

no, it just enables a hidden menu in the bios that allows me to screw
around with the RAS#/CAS# delay and other low-level stuff. all it lets
me clear is the bios password and the BIS credentials.

oh, it does say this about the 1g CPU though:

CPU Microcode Update Rev. [MU168A01]
CPU Stepping Signature [68A]

I dunno if that helps anyone in any way.

____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[kony]

bleh.... I don't really feel like going out and dropping $50 on a new
psu...

May not be PSU but if it is, does it matter if you feel like
it? Take some voltage measurements.

speaking of, can I just needle an unused drive power plug, or do I need
to find the mains on the actual mobo itself? (I have no idea how to go
about doing any of this)

You could measure an unused plug, but technically the
correct voltage is at the plug, plugged into the load. That
means a used plug, for example the video card (when
applicable) and the motherboard connectors, through the back
while the system is running.

 

[Quartz]

[resintall XP]

setup won't run, it dies before it even gets to the blue screen with
"the file txtsetup.sif is missing or corrupt, status 8192". and before
anyone asks, this is a legit cd that works fine if I put the 600 back in.

any ideas?

anyone?


____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.

 

[Stephen]

[resintall XP]

setup won't run, it dies before it even gets to the blue screen with
"the file txtsetup.sif is missing or corrupt, status 8192". and before
anyone asks, this is a legit cd that works fine if I put the 600 back in.

any ideas?

anyone?

The one time I was getting read errors from a cd for an os install
(Win2k). I found that one of my memory sticks was bad.

Stephen
--

 

[Quartz]

[resintall XP]

setup won't run, it dies before it even gets to the blue screen with
"the file txtsetup.sif is missing or corrupt, status 8192". and before
anyone asks, this is a legit cd that works fine if I put the 600 back in.

any ideas?

anyone?

The one time I was getting read errors from a cd for an os install
(Win2k). I found that one of my memory sticks was bad.

hmm.... is it possible that there's some issue with the bus speed? maybe
I should try swapping out my pc133's with pc100's...


____
- Quartz -
there are 10 types of people in the world: those who get this joke, and
those who don't.