PPGA Mendocino to FC-PGA Coppermine

[redbrickhat]

I have an Acer Power SE Series, PN 91.AB744.T05
CPU: Celeron 433 Mhz, FSB: 66 Mhz
Motherboard: Acer V76M, BIOS: Acer V76M BIOS V3.2
Chipset: Sis620/5565


I want to put a PPGA to FC-PGA adapter into the Socket 370, and then
plug in a Coppermine P3.

Questions:


A) What is the highest speed Coppermine P3 I can use?



The Acer support web page
http://www.acersupport.com/desktop/html/pse_specs.html lists for the
Acer Power Series SE the Processor Speeds: 333, 366, 400, 433, 466, 500
or 533MHz Intel Celeron® Socket 370 processor @ 66MHz Front Side Bus
(FSB), 500, 550, 600, 650, or 700MHz Intel PIII @ 100MHz Front Side Bus
(FSB)


But the same page lists the APSe - T800C as an 800 Mhz P3.



B) Do you think the Acer V76M motherboard can support multipliers
higher than 8?


The current CPU is 433 Mhz at FSB = 66Mhz with a multiplier of 6.5,
the Mendocino series had a maximum speed of 533Mhz with a multiplier of
8, so a reasonable choice would be a P3 Coppermine at FSB = 100 x 8 =
800 Mhz.


I was hoping I could try a P3 Coppermine of around 1.1 Ghz which would
require a multiplier of 11 at 100 Mhz.



C) Is it possible to upgrade the BIOS to support a higher multiplier?

The BIOS is called Acer V76M BIOS V3.2

Are there ways to "tweak" the BIOS to increase the multiplier?
..



Thank you.

 

[kony]

I have an Acer Power SE Series, PN 91.AB744.T05
CPU: Celeron 433 Mhz, FSB: 66 Mhz
Motherboard: Acer V76M, BIOS: Acer V76M BIOS V3.2
Chipset: Sis620/5565

First, realize that SIS620 is horrific if using the
integrated video, and still sub-standard without it.

This may not be the ideal system to upgrade.

I want to put a PPGA to FC-PGA adapter into the Socket 370, and then
plug in a Coppermine P3.

One issue here is to be sure it's jumpered correctly, IF it
allows jumper changes. Obviously the documentation for it
is the best source.

Questions:


A) What is the highest speed Coppermine P3 I can use?

It supports 100MHz FSB based on what you show below, so in
theory you could use at least a PIII-1GHz, but only the
version with 100MHz FSB, which is less common than the
133MHz FSB version. More common were the lower speeds, but
the difference overall wouldn't be much on that chipset,
take what you can find easily.

However, the board was designed for lower current
processors, you might find it is a strain and eventual
failure of the board, probably capacitors, to try and run
fastest CPU possible (possible on short-term basis).

The Acer support web page
http://www.acersupport.com/desktop/html/pse_specs.html lists for the
Acer Power Series SE the Processor Speeds: 333, 366, 400, 433, 466, 500
or 533MHz Intel Celeron® Socket 370 processor @ 66MHz Front Side Bus
(FSB), 500, 550, 600, 650, or 700MHz Intel PIII @ 100MHz Front Side Bus
(FSB)


But the same page lists the APSe - T800C as an 800 Mhz P3.

Does it have the latest bios update? You should probably
flash that first, and backup the original bios too before
doing it.

P3 CPUs are multiplier locked, if you were to plug in a
800MHz CPU into a board claiming 700MHz support, it should
work, but perhaps fail to correctly identify the CPU as a
proper model name when it posts.

B) Do you think the Acer V76M motherboard can support multipliers
higher than 8?

yes, there is no "support" per se, it is the CPU that
decides what multiplier to use, independant of the board.

The board merely reports what CPU it is, and the remaining
issue is whether the board unnecessarily (through arbitray
decision of the bios author) halts when it can't positively
ID the CPU. Most boards don't, but you'd have to hear from
an owner of that boad (and using same bios version) to know
for sure about it. In other words, it is expected to work
but not guaranteed, except for that part I mentioned
previously about it using more current which could be a
problem eventually, especially in poorly cooled cases.

The current CPU is 433 Mhz at FSB = 66Mhz with a multiplier of 6.5,
the Mendocino series had a maximum speed of 533Mhz with a multiplier of
8, so a reasonable choice would be a P3 Coppermine at FSB = 100 x 8 =
800 Mhz.

While that seems true, it's mostly non-applicable.
The CPU has a pin that is (voltage) high or low, and this
tells the hardware which FSB speed to use, for example 66 or
100MHz. The CPU itself cannot have it's multiplier changed,
there is nothing the board does here except the frequency of
the FSB, as the CPU's internally generated multiplier
dictates how fast it runs.

I was hoping I could try a P3 Coppermine of around 1.1 Ghz which would
require a multiplier of 11 at 100 Mhz.

You could try one that was 22 x 100MHz if they'd made such a
beast, but does the CPU socket even allow a large enough
heatsink to keep it reasonably cool? You probably can't use
the Celeron 433 heatsink, a 1.1GHz P3 gets significantly
hotter and won't be stable as hot as the Celeron was, in
addition to using more current so the need for airflow
around the CPU VRM subcircuit is even higher that total CPU
wattage (relative to a Mendocino) would suggest.


IMO, since the biggest bottleneck on your platform is the
Sis chipset, you should choose one of the slower P3 that has
the 100MHz FSB, a 700 or 800 MHz model. Also keep in mind
that if your PSU was mATX, and now aged, it might also have
a hard time with the larger current swings of a P3 on an
ACPI enabled OS as 'nix or Windows are.

C) Is it possible to upgrade the BIOS to support a higher multiplier?

It'll work with an infinitely high multiplier already, but
to keep it from stopping when it can' t ID the name of the
CPU, a newer bios might help- there's not much reason NOT to
upgrade the bios, if one is available it's better to do
while you can get ahold of it.

The BIOS is called Acer V76M BIOS V3.2

Are there ways to "tweak" the BIOS to increase the multiplier?

At this point, you need to roll up your sleeves and try it.
Further speculation may be entirely unnecessary.

FWIW, I'd rather have a P3 800 running on a 440BX or 694X
chipset than a P3 1.1GHz on an Sis 620. The best systems
to upgrade are those which started out life as more than
lowest-end systems. Even so, it might be useful as a router
or something like that, but you'd not benefit so much from
fastest CPU as from getting the FSB up to 100MHz.

 

[redbrickhat]

First, realize that SIS620 is horrific if using the
integrated video, and still sub-standard without it.

This may not be the ideal system to upgrade.


One issue here is to be sure it's jumpered correctly, IF it
allows jumper changes. Obviously the documentation for it
is the best source.




It supports 100MHz FSB based on what you show below, so in
theory you could use at least a PIII-1GHz, but only the
version with 100MHz FSB, which is less common than the
133MHz FSB version. More common were the lower speeds, but
the difference overall wouldn't be much on that chipset,
take what you can find easily.

However, the board was designed for lower current
processors, you might find it is a strain and eventual
failure of the board, probably capacitors, to try and run
fastest CPU possible (possible on short-term basis).



Does it have the latest bios update? You should probably
flash that first, and backup the original bios too before
doing it.

P3 CPUs are multiplier locked, if you were to plug in a
800MHz CPU into a board claiming 700MHz support, it should
work, but perhaps fail to correctly identify the CPU as a
proper model name when it posts.



yes, there is no "support" per se, it is the CPU that
decides what multiplier to use, independant of the board.

The board merely reports what CPU it is, and the remaining
issue is whether the board unnecessarily (through arbitray
decision of the bios author) halts when it can't positively
ID the CPU. Most boards don't, but you'd have to hear from
an owner of that boad (and using same bios version) to know
for sure about it. In other words, it is expected to work
but not guaranteed, except for that part I mentioned
previously about it using more current which could be a
problem eventually, especially in poorly cooled cases.




While that seems true, it's mostly non-applicable.
The CPU has a pin that is (voltage) high or low, and this
tells the hardware which FSB speed to use, for example 66 or
100MHz. The CPU itself cannot have it's multiplier changed,
there is nothing the board does here except the frequency of
the FSB, as the CPU's internally generated multiplier
dictates how fast it runs.



You could try one that was 22 x 100MHz if they'd made such a
beast, but does the CPU socket even allow a large enough
heatsink to keep it reasonably cool? You probably can't use
the Celeron 433 heatsink, a 1.1GHz P3 gets significantly
hotter and won't be stable as hot as the Celeron was, in
addition to using more current so the need for airflow
around the CPU VRM subcircuit is even higher that total CPU
wattage (relative to a Mendocino) would suggest.


IMO, since the biggest bottleneck on your platform is the
Sis chipset, you should choose one of the slower P3 that has
the 100MHz FSB, a 700 or 800 MHz model. Also keep in mind
that if your PSU was mATX, and now aged, it might also have
a hard time with the larger current swings of a P3 on an
ACPI enabled OS as 'nix or Windows are.



It'll work with an infinitely high multiplier already, but
to keep it from stopping when it can' t ID the name of the
CPU, a newer bios might help- there's not much reason NOT to
upgrade the bios, if one is available it's better to do
while you can get ahold of it.





At this point, you need to roll up your sleeves and try it.
Further speculation may be entirely unnecessary.

FWIW, I'd rather have a P3 800 running on a 440BX or 694X
chipset than a P3 1.1GHz on an Sis 620. The best systems
to upgrade are those which started out life as more than
lowest-end systems. Even so, it might be useful as a router
or something like that, but you'd not benefit so much from
fastest CPU as from getting the FSB up to 100MHz.

Instead of using the old Acer BIOS, could I use any BIOS for the sis620
chipset?

 

[kony]

Instead of using the old Acer BIOS, could I use any BIOS for the sis620
chipset?

For the core functionality of the chipset, "maybe", but
there may be other features that don't work, and it might
not even post. Generally when one wants to replace the OEM
bios they track down the original manufacturer of the board
and flash the retail board bios. In short, it's not
practical to try it unless you have a socketed EEPROM and
the ability to flash a new EEPROM to the Compaq bios first
in case the different bios doesn't work, then you can swap
in the Compaq bios again.