It is ok to underclock my cpu? Good Reason!

[apatec]

I have a Micron system that has a 200Mhz FSB. The FSB can not be adjusted
on sysbrd (GA-7DX rev 2.2: jp51, j1, and SW1 does not exist) nor in the
BIOS (F1ia).

It is currently using a Athlon 1Ghz 200Mhz FSB cpu. I have a Athlon XP
2600 2.13Ghz 266Mhz FSB processor and curious what would happen if I
swapped the processors. Wouldnt the new processor underclock to 200MHz?

Thanks!

 

[kony]

I have a Micron system that has a 200Mhz FSB. The FSB can not be adjusted
on sysbrd (GA-7DX rev 2.2: jp51, j1, and SW1 does not exist) nor in the
BIOS (F1ia).

If it has the positions for switches, it might be possible
to solder them in and have it work... or maybe not (work)
but you wouldn't know until it's tried.

It is currently using a Athlon 1Ghz 200Mhz FSB cpu. I have a Athlon XP
2600 2.13Ghz 266Mhz FSB processor and curious what would happen if I
swapped the processors. Wouldnt the new processor underclock to 200MHz?

First off you can't even be sure the board would run it
because it's a different family of CPU. Just being same
socket isn't necessarily an indication it'd work as there
were signaling changes and the bios would need be supportive
of it too (at least to the extent of not halting when
unknown CPU was installed).

The new processor would be fine running underclocked, but
the remaining issue is whether the board runs it.

Doesn't it support DDR266 FSB? The bios might it 133MHz,
since MHz terms should be used for the clock rate, not the
FSB rate of "200" posted above, which would be DDR200.

The CPU has pins that tell the board what FSB to run, and
without bios settings or jumpers the board is essentially
forced to run in "auto" mode. If it supports (officially or
unofficially) DDR266 FSB, that's what it'd try to use. If
it doesn't support that then hopefully it'd run at DDR200,
and lower CPU speed 1.6GHz.

The easy answer is to try it and see... shouldn't harm the
CPU if it's not supported. However a newer / faster CPU
puts higher amperage through the motherboard power
subcircuit and that can be stressful to the board, could
reduce it's lifespan.

 

[Paul]

I have a Micron system that has a 200Mhz FSB. The FSB can not be adjusted
on sysbrd (GA-7DX rev 2.2: jp51, j1, and SW1 does not exist) nor in the
BIOS (F1ia).

It is currently using a Athlon 1Ghz 200Mhz FSB cpu. I have a Athlon XP
2600 2.13Ghz 266Mhz FSB processor and curious what would happen if I
swapped the processors. Wouldnt the new processor underclock to 200MHz?

Thanks!

You could always try a AthlonXP-M, which by default has a low multiplier
and should run pretty cool without further hacking. The problem is,
the last retail source I know of, has run dry. You might have to
try Ebay to get one now.

I present this for your amusement purposes only The third link
is the interesting one, if the product is indeed available. I don't know
if anyone has been trying these processors in ordinary S462 motherboards
or not. Buying the Tomcat S2498 would guarantee a working solution,
but it uses DDR ram.

**************

http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=241876

http://store.orbitmicro.com/ccp1966...d-geode-socket-a-motherbo-s2498agn-130655.htm

AMD Geode NX 1500, Socket 462, 1GHz, 133MHz, 128/256KB L1/L2 cache, 6W
http://store.orbitmicro.com/commerce/catalog/product.jsp?product_id=19927

See sample picture with convection cooled heatsink.
http://pc.watch.impress.co.jp/docs/2004/1215/amd.htm

No idea what might happen if that processor was plugged into your
motherboard It might smoke, or it might be icy cool.

Paul

 

[apatec]

Currently, the BIOS has no setting to change FSB, but, there is a bios
update on Gigabyte's website:

http://tw.giga-byte.com/Motherboard/Support/BIOS/BIOS_GA-7DX_More.htm

Gigabyte identifies BIOS version "F8" to support AthlonXP 2600 only if the
Rev of the mobo is 3.0 or higher. Do I try the BIOS update anyway even if
I have rev 2.2?

If no problems with F8 BIOS update then proceed with cpu swap?

If the F8 BIOS update is a problem, should the original Micron BIOS flash
fine?

http://support.mpccorp.com/apps/filelist.asp?ID=3205

Cheers!

 

[kony]

Currently, the BIOS has no setting to change FSB, but, there is a bios
update on Gigabyte's website:

http://tw.giga-byte.com/Motherboard/Support/BIOS/BIOS_GA-7DX_More.htm

Gigabyte identifies BIOS version "F8" to support AthlonXP 2600 only if the
Rev of the mobo is 3.0 or higher. Do I try the BIOS update anyway even if
I have rev 2.2?

Find out why it's limited to rev 3 or higher.

That could be a sign it's not going to work on your board.
However, the bios is generally not the issue so far as a
"newer" version of same bios model, that tends to make a
cosmetic labeling identification of the CPU name when it
posts, not setting the voltage or FSB or being able to
change the signaling which is hard-wired.

If you have the CPU, plug it in and try it. If you don't
have the CPU, get a newer board as there is a serious
performance penalty to running that combo.

If no problems with F8 BIOS update then proceed with cpu swap?

Try the CPU before flashing a new bios.

If the F8 BIOS update is a problem, should the original Micron BIOS flash
fine?

http://support.mpccorp.com/apps/filelist.asp?ID=3205

In theory yes but I've not dealt with Micron boards so I
can't confirm it by having done it myself. If they're using
a standard awards or AMI flasher it should backwards flash
fine. If they've customized a flasher to limit the user
controls more, it's possible you'd have a problem, but I'd
tend to think it more likely that it would still flash to
older revision since that is something a necessary procedure
(rolling back to older bios when newer one is buggy in some
way).

You might also Google search for other users of your board
to see if any have reported success running any kind of
Thoroughbred or newer Athlon rather than a Palomino or
Thunderbird. Since Palomino only went up to XP2100
anything above that would be a sign, though Thoroughbred
started at XP1600 or XP1700 (I forget which at the moment),
so an XP number between XP1600-XP2100 would need further
detail.

I suspect it will work, but also that your board is not
meant to produce the amount of amps that CPU requires and
that it'll be running hot and be shorter lived as a result.

 

[apatec]

Thanks for your post.

Here is what I found out regarding this very SPECIAL GA-7DX REV 2.2 board.
After investigating many forums, there was/are alot of unhappy people out
there when they couldnt set the FSB speed at JP51 jumper for 266Mhz.

Some suggested to carefully solder the JP51 location (which I may try
after possibly getting another system first).

But one person said to press Ctrl+C at POST instead of pressing DEL to
access the BIOS. Well, I tried it and guess what?! It looks very similar
to the standard BIOS menu but it also shows another choice:

Frequency / Voltage Control

when chosen, then next screen looks like this:

ClkGen Spred Spectrum Enabled
CPU Host PCI Clock Default

Im not sure what "ClkGen..." is all about, more research, but the "CPU..."
has the followiing settings: 100/33, 102/34, 104/35, ... 133/33, ...,
to 149/38.

I imagine if the JP51 was closed/jumpered, it would be able to allow the
above settings to be chosen. If not closed/jumpered, choosing these
choices probably would not make any difference except for 100/33; the
default.

So, Im going to update to F2C, Micron's latest BIOS, then pop in the new
processor and see what happens. (Flashing the Gigabyte BIOS F8 is still
very iffy...staying away from that for a while, I would solder first
before doing the update).

 

[kony]

But one person said to press Ctrl+C at POST instead of pressing DEL to
access the BIOS. Well, I tried it and guess what?! It looks very similar
to the standard BIOS menu but it also shows another choice:

Frequency / Voltage Control

Great... I guess... the board should still automatically
adjust the FSB for the new CPU if that FSB is supported, but
this menu would be useful for overclocking.

when chosen, then next screen looks like this:

ClkGen Spred Spectrum Enabled
CPU Host PCI Clock Default

Im not sure what "ClkGen..." is all about, more research, but the "CPU..."

Clkgen spread spectrum reduces EMI on any single frequency
by varying frequency ever so slightly. I recommend always
disabling it, only re-enabling it if you find you have some
piece of (other) equipment that is susceptible to this high
frequency noise.

has the followiing settings: 100/33, 102/34, 104/35, ... 133/33, ...,
to 149/38.

Set to 133/33 if you choose to do this manually, but I
suggest not changing any, and I mean any, settings yet.
Just plug in the CPU and see what it does... you can check
the operational parameters later with software tools like
CPU-Z (for CPU & Memory speed) and Motherboard Monitor 5 for
voltage... though I'm sure the voltage will be right so long
as you leave it on "auto", don't change it.

Also keep in mind that if you were planning on reuse of the
heatsink from the 1GHz Athlon, that might not work well
enough to cool this faster, hotter CPU.

I imagine if the JP51 was closed/jumpered, it would be able to allow the
above settings to be chosen. If not closed/jumpered, choosing these
choices probably would not make any difference except for 100/33; the
default.

So they're greyed out, non-changeable?
Look REALLY closely at the board, front and back, and you
might find that instead of the pins being "open" circuit,
that there might instead be a trace connecting the two.
Sometimes that's how it works, the pins are left off but the
track is there instead and adding the pins to gain
functionality also requires cutting the track first.

Or maybe not, you'd have to determine that, then if you felt
like soldering in the pins.

Keep in mind, you should not need to change ANYTHING to use
the new CPU as it's supposed to run. These bios menus and
board jumpers do not change the speed from 100/33, they
change it from the "default" for any given CPU. The default
on the new cpu will be different than on the present one.
That is, unless they did something wrong in designing this
board, as that's how it works on any other.

So, Im going to update to F2C, Micron's latest BIOS, then pop in the new
processor and see what happens. (Flashing the Gigabyte BIOS F8 is still
very iffy...staying away from that for a while, I would solder first
before doing the update).

You're better off not changing ANYTHING, just putting the
new CPU in first. By flashing the new bios you introduce a
new variable... for all we know the new bios alone could
make changes that effect the system operations... so if you
do change the bios, I suggest NOT changing the CPU yet,
rather running system for awhile to be sure the new bios
keeps entire system running correctly first with old CPU
still installed.

 

[apatec]

CONFIG:
MicronPC Millennia AMD Athlon 1.0ghz 200MHz
GA-7DX Rev 2.2
BIOS F2C

RESULTS:
Ok here are the results of swapping the AMD Athlon 1.0Ghz 200FSF with the
Athlon XP 2600+ 2.13Ghz 266FSB (AXDA2600DKV3C):

1. Shutdown and remove HSF and 1Ghz CPU
2. Insert Athlon XP 2600 with thermal compound and placed compatible HSF
and connected cable.
3. Reboot and pressed Ctrl+C at POST (Hidden key not many know about)
BIOS setting shows: Frequency/Voltage setting
4. Changed CPU HOST PCI CLOCK from "Default" (which is basically 100/33)
to 133/33.

*(I was able to see PC Health setting of CPU with FAN running and CPU
voltage info)

4. Saved BIOS setting and reboot

Result: No video... nor do I hear anything from HD as if it was booting
either.

1. Shutdown and remove CMOS battery for 5min
2. Replace CMOS Battery and reboot normal with CPU HOST back to "Default"
3. At POST press Tab to see listing and it stops:
Detecting unknown CPU
1600Mhz
F1 to continue or DEL to access BIOS

Note: I dont know what else to change in BIOS to rid this interruption at
boot
everytime.

Able to get to Windows noticeably quicker but even CPU-Z reports running
the Thoroughbred, Athlon XP, 1611Mhz at 100 bus speed.

Conclusion:
It appears that the Gigabyte tech was mistaken when he said "the BIOS
setting for FSB does NOT take precedence over the "JP51 -no jumper-
100Mhz setting".

FYI - Micron manual describes JP51 to change FSB from 100 to 133, but when
I looked at the Rev 2.2 board, JP51 nor SW1 (in Gigabytes manual) does not
exist. It does show the description though on mobo)

1. Any thoughts or solution on the interruption at POST to press F1?
2. Do I someday solder JP51 (once I have another computer just in case I
fry it)?

 

[kony]

CONFIG:
MicronPC Millennia AMD Athlon 1.0ghz 200MHz
GA-7DX Rev 2.2
BIOS F2C

Note that this is not a website forum (you post to,
regardless of where you are when you read this), this is
usenet, so your bold control characters ( ) and other
similar formatting characters are not useful here,
actually detract from the readability of the post. Actually
if the web forum is allowing these control characters they
should not be doing so, please ask them to disable it.
Thank you.

RESULTS:
Ok here are the results of swapping the AMD Athlon 1.0Ghz 200FSF with the
Athlon XP 2600+ 2.13Ghz 266FSB (AXDA2600DKV3C):

1. Shutdown and remove HSF and 1Ghz CPU
2. Insert Athlon XP 2600 with thermal compound and placed compatible HSF
and connected cable.
3. Reboot and pressed Ctrl+C at POST (Hidden key not many know about)
BIOS setting shows: Frequency/Voltage setting

It "might" not matter, but you should never try to change
this on the very first boot. The reason is that some boards
always start out at the lowest supported speed when
detecting new CPU, then later adjust. It would be useful to
just post, enter and save the bios with no changes, after
clearing CMOS. I don't claim it will "necessarily" help,
but it does allow the bios to do what it can.

4. Changed CPU HOST PCI CLOCK from "Default" (which is basically 100/33)
to 133/33.

*(I was able to see PC Health setting of CPU with FAN running and CPU
voltage info)

4. Saved BIOS setting and reboot

Result: No video... nor do I hear anything from HD as if it was booting
either.

Do you have PC133 memory? If there is an option to set it
to most conservative timings you might try that. In theory
the bios setting is supposed to work, and the jumpers are
only to override it. This is in general, on your board I
can't guarantee anything not having tried exactly what
you're doing.

1. Shutdown and remove CMOS battery for 5min

The AC power was unplugged during this 5 minutes, yes?

2. Replace CMOS Battery and reboot normal with CPU HOST back to "Default"
3. At POST press Tab to see listing and it stops:
Detecting unknown CPU
1600Mhz
F1 to continue or DEL to access BIOS

Note: I dont know what else to change in BIOS to rid this interruption at
boot
everytime.

There might be a setting such as "half on error" where you
choose all, keyboard or mouse, etc, none (might be worded
differently, and possibly on a different bios menu page).

Able to get to Windows noticeably quicker but even CPU-Z reports running
the Thoroughbred, Athlon XP, 1611Mhz at 100 bus speed.

Conclusion:
It appears that the Gigabyte tech was mistaken when he said "the BIOS
setting for FSB does NOT take precedence over the "JP51 -no jumper-
100Mhz setting".

Maybe, but that is usually the whole point of having the
bios setting... if you had to have the jumper moved for it,
there was no point in having a bios setting at all as the
same jumper movement would just set to/fro 133FSB.

Are you sure your revision of board supports 133MHz FSB?
One thing about different revisions of board that use same
bios version(s) is that the newer bios supports and lists
the options the newer board revision supports but aren't
applicable to the older revision (necessarily).

Regarding the issue I mentioned previously about a possible
trace between the two jumper pins, if you want to try the
jumpers then look closely to be sure there isn't a trace
between them already and if not, try taking a very short
piece of thin wire, tinning a few millimeters of it
solder/flux first, then temporarily tacking it down onto the
corresponding solder points.... that's quicker than having
to pull the board out for access to the back of it to solder
in jumpers, it'll tell you if that would work. Alternately
if you have some precision PCB drill bits, sufficiently
small enough that you can drill stright through the solder
(being careful not to leave behind solder debris on the
board), you can try drilling a very tiny hole, dead-center
where the jumper pins would go being careful not to drill
out the side of the plated-thru hole, and with these two
very tiny holes you can drop in a piece of solid strand wire
bent into a small "u" shape.... not soldering it yet. If it
posts like that, and/or you then can change bios settings
and repost, you can solder the wire in from the top side
rather than the bottom... all this at your own risk of
course, since I have no idea if you're good at soldering.

FYI - Micron manual describes JP51 to change FSB from 100 to 133, but when
I looked at the Rev 2.2 board, JP51 nor SW1 (in Gigabytes manual) does not
exist. It does show the description though on mobo)

Gigabyte does that a lot though, I recall several older
boards where they had pin header positions but didn't even
have the corresponding complimentary surface components on
adjacent traces so even with jumpers the circuit wasn't
complete. It wasn't always that way though, all you can do
is look at what you have and try a few things.

1. Any thoughts or solution on the interruption at POST to press F1?
2. Do I someday solder JP51 (once I have another computer just in case I
fry it)?

If you're really sure that a jumper goes where you see two
particular through-board holes, you might try carefully
holding a bent paper-clip's ends against these two points to
see if that works. I don't recall where the jumper would be
though so I don't know how hard it is, nor how steady your
hands are. Conductive paint or epoxy is yet another option
but a bit messier and if any drips in the wrong place, could
be quite a PITA to clean off, or even worse if you didn't
notice it and it shorted out something.