Why diode temp jumps around?

W

w_tom

1 amp when entire 'system' is working hard and 0.55 amps
when only CPU (no peripherals) is doing work. Those numbers
can be interpreted this way as well. However, take the
heatsink off that CPU when 'system' is only consuming 0.55
amps. Will processor still work 'not hard'? Heatsink cannot
be removed because CPU is working hard at both 1 amp 'system'
consumption and 0.55 amp 'system' consumption. So hard that
heat sink is required during both events. Running at 0.55
amps or sprinting at 1 amp system consumption.

Where may the first 0.55 amps be going? Mostly to CPU when
'system' is doing less.

An overheating Pentium activates an internal emergency
protection circuit that was even found in 486s. No mystery
there. To protect a CPU that is overheating, then CPU cycle
execution time is slowed (in some Pentiums is said to stop).
This is not 'normal computer' operation from a user's
perspective (probably why it is not discussed in datasheets).
From a user's perspective, is CPU 'not working hard' when user
does not execute application programs? No. CPU is 'working
hard' constantly which is why CPU always needs that heatsink.

CPU is always doing something on every clock cycle.
Sometimes that may only be a Halt instruction (reduced
power). After all, how does it know when to stop executing
Halt? Its reduced processing detects the interrupt. Other
times, CPU is executing those other 20+ processes. All this
is still so 'hard' that CPU requires heatsink even when not
executing anything for the user.

Fan tells us little about how much more heat a Pentium
produces. It simply reports when the CPU is 'running' verses
'sprinting'. Again, take the heatsink off any processor that
is 'not doing anything'. See how long it can still 'do
nothing' - normally. Why? Because even when it is 'not doing
anything', it is doing so much as to generate heat that
requires a heatsink.

And so we return to the question. What, from the user's
perspective constitutes hard work? Electrical consumption /
thermal dissipation, or output from CPU to be observed by
user? CPU electronics can still execute and yet not provide
the user with any useful output. Is user then to assume the
CPU is doing almost nothing? If so, they why is CPU still
producing so much heat?


Again, the question - does a laptop tend to execute more
Halt instructions as compared to a desktop?
 
R

Robert Hancock

An overheating Pentium activates an internal emergency
protection circuit that was even found in 486s. No mystery
there. To protect a CPU that is overheating, then CPU cycle
execution time is slowed (in some Pentiums is said to stop).

In anything before the Pentium 4, execution will simply stop when the
THERMTRIP temperature is hit.

Looking at the Pentium 4 datasheets, the power consumption in the auto-HALT
or stop-grant modes is about half of the maximum possible power consumption.

A laptop is not going to execute more halt instructions - the OS will use
the halt instruction when there is nothing to run, and never when there is.
However, some laptop CPUs have features (SpeedStep, LongRun, etc.) that will
reduce the clock speed in some situations to save power.
 
B

Bob

www.sandpile.org
Robert Hancock said:
In anything before the Pentium 4, execution will simply stop when the
THERMTRIP temperature is hit.

Looking at the Pentium 4 datasheets, the power consumption in the auto-HALT
or stop-grant modes is about half of the maximum possible power consumption.

A laptop is not going to execute more halt instructions - the OS will use
the halt instruction when there is nothing to run, and never when there is.
However, some laptop CPUs have features (SpeedStep, LongRun, etc.) that will
reduce the clock speed in some situations to save power.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top