CPU performance test Estimated Energy Usage - wrong information?

[snotaap]

On the CPU compare page on this website, there is a very simple 'Estimated Energy Usage' section below the performance comparison.
I see at least two issues with the information as it is presented if if this is supposed to be useful:
1. Few people run their CPU at 100% TDP all of the time, so the energy usage is vastly overestimated, not in any way realistic; and
2. above all, for the same energy usage, a faster CPU gets more work done so the actual energy consumption for a user after choosing a CPU based on this comparison should be divided by the CPU Mark rating. After all, if he would run the slower CPU1 at 100% all of the time, he would use only (mark_cpu1/mark_cpu2)* TDP_CPU2 energy to do the same job.
As it is, I see is no value in the information and especially not when comparing CPUs.

[David (PassMark)]

We don't have statistics for idle power usage. But in most cases it will be pretty insignificant (i.e. it isn't something you would based a purchasing decision on). The motherboard, monitor, etc.. will more just as or more important, when the CPU is idle.

For most people the long duration heavy usage application is gaming. And in that case people don't lower then frame rates on fast CPUs (i.e. the amount of work is greater if the CPU is faster).

But we agree is it a very simple calculator, giving only a very rough indication of cost.

[snotaap]

Well I would have thought the great majority of PC's are slugging away in offices and data centers and that, therefore, 'most people' are actually just writing text, browsing, writing emails and doing other productivity tasks? Perhaps I am hopelessly behind the times and everybody is into heavy gaming these days but I frankly believe the vast majority of CPUs in this world is NOT being used for gaming. Also, your implication that the CPU power feature might be interesting for gamers, aka those people overclocking their CPUs to wring the last performance at enormous power cost and buy expensive video cards that routinely suck up 400W, certainly does not make much sense.
I do not know your target audience but I think that a power feature that shows equivalent power use for a 5yo CPU and a 'hugely faster' recently released one with the same TDP is not a helpful aid in selecting one or the other given the real-work workload of most people.

[David (PassMark)]

As far as I know there is no published data of CPU power usage while the CPU is idle (or near idle). At least not for a wide range of CPUs. (This is because it is near impossible to measure once the CPU is in a socket or soldered to the board). And yes if you only doing EMail, you CPU will be idle for long periods.
Most of the published idle data is for total system power usage (motherboard, PSU losses, RAM, I/O + CPU). Which isn't what we are comparing.

But there is general agreement that the power usage while idle is pretty low, so not very interesting as a point of comparison (except in battery powered laptops, but in that case total system power is more interesting).

We do the best with the data we have.