How bad is SMP with 2 different processor speeds?

[Norm Dresner]

I have a dual-capable SBC CPU board which accepts Socket 370 CPUs and I
happen to have 2 Pentium III cpus, one is 700MHz and the other is 800MHz.
What are the chances that I'd be able to run, say, Linux or Win2K with this
combination?

TIA
Norm

 

[Andrzej Popowski]

I have a dual-capable SBC CPU board which accepts Socket 370 CPUs and I
happen to have 2 Pentium III cpus, one is 700MHz and the other is 800MHz.
What are the chances that I'd be able to run, say, Linux or Win2K with this
combination?

This is not a proper setup, but IMHO should work OK.

There could be problems if this CPUs have different supply voltage or
FSB frequency. In this case results would depend on motherboard
hardware.

 

[Tony Hill]

I have a dual-capable SBC CPU board which accepts Socket 370 CPUs and I
happen to have 2 Pentium III cpus, one is 700MHz and the other is 800MHz.
What are the chances that I'd be able to run, say, Linux or Win2K with this
combination?

Assuming they run at the same voltage and bus speed, chances are about
50/50. Some motherboards will do it, some will not. The operating
systems themselves should be able to handle it, but some drivers might
be problematic.

 

[Triffid]

I have a dual-capable SBC CPU board which accepts Socket 370 CPUs and I
happen to have 2 Pentium III cpus, one is 700MHz and the other is 800MHz.
What are the chances that I'd be able to run, say, Linux or Win2K with this
combination?

TIA
Norm

I have experimented with mis-matched PIII CPUs a few times and tend to
agree with Tony - the chances of it working reliably are no better than
50/50.

Most dual boards display a BIOS message indicating the family, core
speed, and number of processors detected on boot. IME it works better if
the BIOS reports the core speed of the slower processor, i.e. switch the
CPUs around if the BIOS sees the faster one.

 

[Norm Dresner]

I dug the CPUs out of the carton and looked closely at them.
One CPU is clearly marked
750/256/100/1.65v
which I assume is a 750MHz w/256 K cache, 100MHz

The other is marked
RB80526P4600256
7007A330-1672
I beleive this is a 600 MHz and I'm interpreting the 256 to be the same as
in the 750. But the voltage? I'm guessing 3.30v -- Is this correct?

If one is 3.3v and the other is 1.65v then these two CPUs will definitely
not work on a dual board, right?

TIA
Norm

 

[Triffid]

I dug the CPUs out of the carton and looked closely at them.
One CPU is clearly marked
750/256/100/1.65v
which I assume is a 750MHz w/256 K cache, 100MHz
Yup.

The other is marked
RB80526P4600256
7007A330-1672

That's an Intel order code, which AFAIK doesn't tell you any more than
core speed and cache (which you've correctly interpreted) unless you can
find it at intel.com (I couldn't).

All Intel CPUs should have an sSpec number on them somewhere - 5
character alphanumeric, first character "S". Find those, then plug them
in at processorfinder.intel.com for all the gory details.

I beleive this is a 600 MHz and I'm interpreting the 256 to be the same as
in the 750. But the voltage? I'm guessing 3.30v -- Is this correct?

Can't be. A 600Mhz PIII has either a Katmai (2.0v) or Coppermine
(1.6-1.8v) core.

If one is 3.3v and the other is 1.65v then these two CPUs will definitely
not work on a dual board, right?

IME a voltage difference is not a problem as dual boards typically have
separate Vcore regulators for each CPU and are therefore able to honour
mismatched VID requests.

 

[Norm Dresner]

That's an Intel order code, which AFAIK doesn't tell you any more than
core speed and cache (which you've correctly interpreted) unless you can
find it at intel.com (I couldn't).
All Intel CPUs should have an sSpec number on them somewhere - 5
character alphanumeric, first character "S". Find those, then plug them
in at processorfinder.intel.com for all the gory details.

Thanks for the URL. I found the sSpec # on the chip and it's
600MHz / 256K / 100 MHz / 1.65v
so the voltages match after all.

As for the board -- since it's a PICMG CPU board and not a motherboard I
wouldn't count on having enough room to have two independent VRs -- but it's
possible. I can't get to the board right now to check but I will, though
it's no longer an issue.

Thanks
Norm

 

[Triffid]

this combination?


Thanks for the URL. I found the sSpec # on the chip and it's
600MHz / 256K / 100 MHz / 1.65v
so the voltages match after all.

As for the board -- since it's a PICMG CPU board and not a motherboard I
wouldn't count on having enough room to have two independent VRs -- but it's
possible. I can't get to the board right now to check but I will, though
it's no longer an issue.

Thanks
Norm

They're both Coppermines, speed discrepancy is only 150Mhz - it might work!

Let us know...