normally people are concerned about CPU temps being too hot, but I have the opposite concern. I have a new 2.6c Northwood SL6WS with the stock intel cooler. Not overclocked or anything. I have it running on an MSI 865GM2 housed in an Antec Aria with the stock PSU fan and stock cyclone blower. Basically, everything is stock. My IDLE temp is only about 21C. The rooms ambient temperature is about 18C. Is this right? I know the Northwoods are supposd to be cool, but this seems a bit too cool.... I'm thinking it might be something wrong with the temperature sensors on my motherboard maybe?
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Is this 2.6c northwood idle temp right?
- Thread starter pdp76
- Start date
Are you getting this temp from your system bios or from a Windows-based monitoring utility? If the latter, what you're seeing is likely not an idle temp. Look in your system bios instead and compare the two readings.
Modern OS's including WinXP and Win2000 put CPUs in a power-saving state whenever usage falls below a certain level. So the readings you get are artificially low, sometimes 10-15C or more.
Modern OS's including WinXP and Win2000 put CPUs in a power-saving state whenever usage falls below a certain level. So the readings you get are artificially low, sometimes 10-15C or more.
I use a windows monitoring tool, and yes, I've noticed that the BIOS temp readings are higher than the windows readings. That makes sense. So the 2.6c Northwoods are capabale of going into power saving state? I was unaware of that, thanks for the info!Frank DC said:
pdp76 said:
No ,that's BS...
In BIOS Intel cpus run like in DOS mode , aka full load but they aren't doing much work => temp will be higher than in idle mode but lower than full load running a power demanding app.
Idle temp is just what it means : doing nothing.
Idle temp should be measured after half an hour of operation , if you measure temp right after windows started the HSF isn't warm yet and you'll get a lower but false temp.
So are you saying when viewed in the BIOS monitoring tool, the temps will be higher than when viewed in Windows? Because if so, then we agree.savantu said:
But yeah, the temps I'm reporting are after at least an hour of being powered on and everything is "warmed up".
Now that is almost total BS. ACPI-capable operating systems (including WinXP, Win2K etc) implement power saving states for CPUs, one of which takes effect the moment a CPU's usage drops below a certain level. In the ACPI spec they are called C1, C2 and C3. In fact AMD has a utility that displays these current states, so you can watch this throttling as it happens.savantu said:
If you want to know more details see:
http://download.microsoft.com/download/5/7/7/577a5684-8a83-43ae-9272-ff260a9c20e2/Windows%20Native%20Processor%20Performance%20Control.doc
The bottom line is, your CPU's true "idle" temps are displayed in your system BIOS. The ones displayed in Windows are usually artificially low.
Sorry mate - but you are the ignorant one here.Frank DC said:
This document is referring to processors that utilize functions like Intel SpeedStep. The Northwood / Prescott processors do not have any throttling available to them that would allow them to slow down just because Windows says so.
The guy in the previous post was correct in stating that the BIOS loads your CPU and will give you a pretty good idea of what your processor is going to look like under full normal load (load = 100% but not running SETI and Prime 95 at the same time).
That's about the most ridiculous thing I've ever read. These power saving states are part of the ACPI specification, and have zilch to do with what processor one happens to be using. E.g. Windows doesn't care whether the HLT instructions it's sending (which is how the first of these states work) are to an AMD or Intel processor, which processor it is, whether any proprietary functions such as SpeedStep are supported. Etc.
The level of ignorance on this board is nothing short of astonishing. ACPI/OS-directed CPU power saving has been around since Win9x for chrissakes.
The level of ignorance on this board is nothing short of astonishing. ACPI/OS-directed CPU power saving has been around since Win9x for chrissakes.
ACPI 2.0 spec. See section 8 starting on P.223:
http://www.acpi.info/spec20c.htm
This thread deals specifically with C1/the HLT instruction, which is defined in section 8.1.2:
-----
8.1 Processor Power States
ACPI supports placing system processors into one of four power states while in the G0
working state . Processor power states include C0, C1, C2, and C3. The C0 power state is
an active power state where the CPU executes instructions. The C1, C2, and C3 power
states are processor sleeping states where the processor consumes less power and
dissipates less heat than leaving the processor in the C0 state.
...
8.1.2 Processor Power State C1
All processors must support this power state. This state is supported through a native
instruction of the processor (HLT for IA 32-bit processors), and assumes no hardware
support is needed from the chipset. ... Aside from putting the processor in a power state,
this state has no other software-visible effects. In the C1 power state, the processor is able
to maintain the context of the system caches.
http://www.acpi.info/spec20c.htm
This thread deals specifically with C1/the HLT instruction, which is defined in section 8.1.2:
-----
8.1 Processor Power States
ACPI supports placing system processors into one of four power states while in the G0
working state . Processor power states include C0, C1, C2, and C3. The C0 power state is
an active power state where the CPU executes instructions. The C1, C2, and C3 power
states are processor sleeping states where the processor consumes less power and
dissipates less heat than leaving the processor in the C0 state.
...
8.1.2 Processor Power State C1
All processors must support this power state. This state is supported through a native
instruction of the processor (HLT for IA 32-bit processors), and assumes no hardware
support is needed from the chipset. ... Aside from putting the processor in a power state,
this state has no other software-visible effects. In the C1 power state, the processor is able
to maintain the context of the system caches.
Frank DC said:
Oh my...you truly are ignorant.You are contradicting yourself : we are aware of the HLT instruction and precisely this makes the cpu run in "idle"mode.How else can you define idle ??
You simply suport our point : in Windows you are "idle" , in BIOS you are similar to running DOS ( aka full load but not doing any actual work).Do yourself a favour and boot into DOS and you'll hear the fan spinning up...Or run Borland Pascal which acts exactly the same.
The HLT instruction initiates a power saving state, not an "idle mode". Is the difference between the two so difficult to grasp?savantu said:
>> You simply suport our point : in Windows you are "idle" , in BIOS you are similar to
>> running DOS ( aka full load but not doing any actual work).
Not even close. When sitting on a BIOS screen (or DOS prompt), no processing is taking place, and no power saving state has been applied. This is the proper and accurate definition of idle. If we use your definition, it would make just as much sense to turn off a computer altogether, and then claim its CPU runs at ambient room temperature at "idle". Can you understand the difference?
BTW "full load but not doing any actual work" doesn't even make sense. Exactly what do you think is causing your imaginary "full load"? A CPU is either processing instructions or it isn't.
Frank DC said:
So - by power saving state - does it reduce the voltage to the CPU or just clock it down?
I have seen this. For example - if you run SETI on a single CPU - you will peg it at 100%. If you run SETI and Prime 95 and SuperPI all at the same time the temperature will jet up well above what it was when running SETI alone. Does that mean that SETI alone wasn't realy running the CPU at 100%?Frank DC said:
A basic explanation of C1 can be found here:Nomad said:
http://www.heise.de/ct/english/98/20/166/
C1 reduces power consumption by intermittently stopping a CPU's clock and by freezing the CPU's registers during these "off" periods. Temperature reductions caused by C1 vary widely depending on CPU and cooler. E.g. on air-cooled P4 Northwoods and Prescotts it's typically 10C or more.
Internally there are many different ways to "peg a CPU at 100%". A 100% reading in e.g. Windows Task Manager doesn't tell you anything about which areas within the chip are being stressed, and the temp increase you see is largely dependent on exactly what is being done. There's a good explanation of this in CPUBurn's documentation, if you can find it. It goes into detail about the utility's different tests, each of which results in "100% usage", but with very different temperature effects on a CPU.