if you run power dvd, windows media player etc the API routine calls in hardware of your laptop/desktop.

ie sound chips or gpu... then of course your battery or power resources will change accordingly.

its not rocket science.

CPUs use DVFS to lower frequency of the processor when needed= lower power consumption.

Perhaps I misunderstand exactly what you are saying, but running the software will consume power... the flip flops need to change states, when that happens, the flip flops dissipate heat. The software does not "call" hardware, it simply interfaces to it. The software dictates which flip flops are changed, so a power consumption could be attributed to the software, but the ability to measure this is nigh impossible...

 

How many calories does my body burn while reading this thread?

 

How many calories does my body burn while reading this thread?

It all depends of how many other things you are doing at the same time.

Please say what RTOS have you installed and version. Is it the original or somebody reprogrammed you? Better check with parents if possible.

 

ransiluj, you may find these responses to be rude and unhelpful, but the members here are just trying to make a point. What you ask cannot be done. Not "It will be hard to do", but it CAN'T be done. If you run a program on a laptop, the HARDWARE in the laptop will draw a different amount of current than a desktop running the exact same program. It will be impossible to design something to measure the current "drawn by different software". Instead, why don't you try to program a CPU meter? That will give you a better idea of how much stress a program is putting on your computer. Try programming a Rainmeter skin or something. Google it. But determining how much current a piece of software draws cannot be done.

 

The spider in my bathroom just sits there for days and weeks.

How many nanoamperes are running in her brains?

Fascinatingly efficient- About 1cm size, can climb walls, can survive for months without food, totally deadly stingers, can run very fast with hydraulic legs.

 

ransiluj, you may find these responses to be rude and unhelpful, but the members here are just trying to make a point. What you ask cannot be done. Not "It will be hard to do", but it CAN'T be done. If you run a program on a laptop, the HARDWARE in the laptop will draw a different amount of current than a desktop running the exact same program. It will be impossible to design something to measure the current "drawn by different software". Instead, why don't you try to program a CPU meter? That will give you a better idea of how much stress a program is putting on your computer. Try programming a Rainmeter skin or something. Google it. But determining how much current a piece of software draws cannot be done.

I inserted the floppy disk but the computer just shredded it and all the bits blown out on the floor. He said "I have my bad hair day".

 

Whar are you going on about, takao?

 

If you were to actually attempt to measure the amount of current the processor draws from running a specific program, you'd have to figure out the state of every single flip-flop in your CPU. Then you'd have to determine which ones were from the OS, background threads, other programs, interrupts, etc. and remove their current consumption contributions from the total current draw...and do this for each clock cycle....

So, good luck

well that is hard right? we are palning to measure it from outside

 

My question is:

Why on earth would you care how much current a particular software application consumes?

Is this going to be used for marketing???

Our new XYZ software consumes 20ma less than our competitors!

for desing low power consumption software

 

ransiluj, you may find these responses to be rude and unhelpful, but the members here are just trying to make a point. What you ask cannot be done. Not "It will be hard to do", but it CAN'T be done. If you run a program on a laptop, the HARDWARE in the laptop will draw a different amount of current than a desktop running the exact same program. It will be impossible to design something to measure the current "drawn by different software". Instead, why don't you try to program a CPU meter? That will give you a better idea of how much stress a program is putting on your computer. Try programming a Rainmeter skin or something. Google it. But determining how much current a piece of software draws cannot be done.

please try this.. it can say how much power that ur program takes

http://research.microsoft.com/en-us/downloads/fe9e10c5-5c5b-450c-a674-daf55565f794/

 

please try this.. it can say how much power that ur program takes

http://research.microsoft.com/en-us/downloads/fe9e10c5-5c5b-450c-a674-daf55565f794/

Ransiluj, that is NOT measuring how much current a program draws. The concept itself is completely ridiculous, and I don't know why you can't just accept that. That program Estimates the power consumption of your entire system by using CPU usage and other known statistics. It DOES NOT MEASURE the current drawn by a single program (not that it's possible anyway). All above terms in bold are important terms you need to look at. You're completely misunderstanding the entire concept. I suggest you do a little more research about how computers work before even attempting something like this, since it's quite obvious you don't know what your'e talking about (no offense intended).

I've said my part, and if you can't accept that what you're asking is impossible to do, then I have nothing more to say here.

 

well that is hard right? we are palning to measure it from outside

I did not point that out so that you would think this is possible... I pointed this out so you would understand how unbelievably, incredibly hard this would be... It is of no use even if you were to get it working as it would only work for one version of the program...

 

These guys are giving another meaning to the term "green software".
( I hope this comment doesn't encourage them.)

 

Current consumption manifests itself as heat dissipation within the CPU. You could utilize the system management buss to analize the heat generation of the CPU. However, you'll find that every processor will draw a different current to perform a similar task, so you'd have to establish a benchmark of some sort.

Maybe you could tell us how you'd benchmark your conception.

 

Current consumption manifests itself as heat dissipation within the CPU. You could utilize the system management buss to analize the heat generation of the CPU. However, you'll find that every processor will draw a different current to perform a similar task, so you'd have to establish a benchmark of some sort.

Maybe you could tell us how you'd benchmark your conception.

Heat transfer is not an instantaneous representation of power consumed. Heat may be dissipated through various means(convection, conduction, and radiation), so touching a temperature sensor to the CPU will not be representative. Also, you have no control of what processes the OS is running at any time, in order to distinguish the consumption from the various processes...

 

Heat transfer is not an instantaneous representation of power consumed. Heat may be dissipated through various means(convection, conduction, and radiation), so touching a temperature sensor to the CPU will not be representative. Also, you have no control of what processes the OS is running at any time, in order to distinguish the consumption from the various processes...

I won't disagree. Just an idea for the OP, as instantaneous current would, in my view not tell you much. Such an appoach would compensate for background processes on a broader time scale.

I'm thinking that the OP asked a question with some intent. It didn't work out as evidenced by the responses, but maybe he didn't ask the right question. ?

 

I won't disagree. Just an idea for the OP, as instantaneous current would, in my view not tell you much. Such an appoach would compensate for background processes on a broader time scale.

I'm thinking that the OP asked a question with some intent. It didn't work out as evidenced by the responses, but maybe he didn't ask the right question. ?

Considering that a transistor in a CMOS device only draws current when switching states, you would need the instantaneous power consumption for each clock cycle. Each transistor's power dissipation would need to be summed and distinguished between the OS, background threads, and the desired program.

The question seems fine, but shows an inherit lack of understanding as to how a computer works. Aside form creating your own OS, there isn't much chance of determining the power usage of running a certain piece of software, at which point, determining the power consumed from running a program is moot as this is only valid in the one instance of the processor, memory, and OS.