The correct way to make this measurement is to run a set of various application tasks, measure the total power used and time how long it took to run the test. Average power is total power used divided by total time.

A Power Angle or KAW is no good because it measures the total power used by the computer. That's effected by the MB, video device used, etc.

Sadly, it's current streetprices where I live. And no, Newegg is not an option for me, due to shipping overseas + ridiculus local taxes and fees.

Yes it does.

Nice calculations. But you didn't take into account the extra time needed to perform the same task due to the lower performance. So you may actually end up using MORE power on the e system.

Lastly, SPCR, are you sure the TDP you are quoting is actually TDP? AMD normally works with ACP (average cpu power) which is actually lower than the TDP (thermal design power)

First, the article clearly states that SPCR used a Power Angel to measure total system consumption.

Second, your calculations ain't right. Power * time = energy. So dividing accumulated KWh by time gives average power.

I'm not sure if SPCR's multimeters can log the CPU/VRM consumption or if they provide only instantaneous results. The Power Angel may be their only way to determine average consumption with varying loads.

I think AMD only uses ACP when they're talking server CPUs like the opterons. And they usually specify both. TDP for engineering maximum thermal cooling required. ACP for estimating power usage such as for data centers. I've never seen it mentioned for desktop CPUs.

Thank you! I'm glad someone appreciated that statistic!

Eh.

200 dollars is pricey considering on board video for Intel sucks. Not to mention the boards are more money to get less features. A LOT LESS features. And of course, you can get a 4290 on board video with amd for less than 110 dollars.

I am not a fan of amd vs intel, but did anyone check out the x3 720? 104 dollars shipped. Beats the pants value wise off of the 750. You can save 150 dollars going amd in that regard. (board + chip price difference) I would guess the wattage saved after a couple of years would negate some of this if you paid a lot for electricity.... I guess? shrug?

I know it wasnt included in the review as quads are quads. just my thoughts though concerning the best actual values.

There's never a question that Mike C has great articles. But there are other questions that always need asking.

Have you seen the non-e voltages? Why were they so high in the first place? Just so that they would be easily overclocked?

Who's going to recoup the price difference of the e version over it's lifetime? That's referring to just electricity, not any possible savings in cooling outlays.

I'd like to see just how low half a dozen non-e equivalent cpus can go before I make such an investment.

Costs at idle matters only those machine that spend most of their time idling. Cost/work done is a more important indicator.


Check the archives for other X3 reviews, you may find the 720 in there for comparison purposes.

I strongly disagree with this sentiment. Overclocking or undervolting is a crapshoot. What happens when you get a processor that doesn't undervolt to your expectations? You can't return it. So you're stuck with a processor that doesn't meet your expectations.

If you get one of these low power processors then you are guaranteed the lower power and lower temperature characteristics.

The lower temperature that these processors exhibit is a big deal. It allows a smaller heatsink, or a passive heatsink.

But most importantly for me is that heat has to be paid for twice in hot climates. You can't calculate the ROI based solely on differences in power draw. You have to factor in cooling costs. A hot CPU really heats up a room and the AC has to work that much harder to compensate in the summer months.

I'm not sure how the 'e' chips would weigh down on me if I hadn't picked up all three of my X2 235/235/240 'e' purchases at 25-30 quid; Sempron 140 prices, 2-3 weeks before the official launch. There was nothing that could touch them at such cost.

Some 'e' designated chips are well binned and not just undervolted to meet spec. The second X2 235e I chose to run in my AIO comes at 29W DC at the 12v at 1.350v stock, and I run it undervolted to 1.03v 2.7GHz. Imagine the power drop... it was not possible with a standard power 245/250 chip available back then, or even now.

I wouldn't have bought it if I wanted OC'ing but default voltage gets it to 3.64GHz stable. As a side note, I typically run it with the HD 4200 IGP but if required, I just add a passive HD 3650, or a passive HD 3850. The one low-speed fan at slo-mo helps significantly. The passive HD 2400/HD 4350/HD 5450 I tried before these were abysmal to replace the IGP with - hardly adding anything except power/heat.

One factor missing in these equations is the cost of one's time. Obviously that varies significantly depending on whether one is a student, unemployed/underemployed, or has fairly high paying job.

Considering all the money I spent previously on fans, controllers, wildly expensive CPU coolers, and the like, and what I perceive to be the value of my time, paying an extra $25 for "e" is well worth the cost, especially if I don't have to worry about how to under-volt a regular CPU or whether my MB can do it easily.

Basically for similar reasons, I quite mowing my lawn a long time ago, and pay someone else to do it. The extra time I spend on my personal life, or my job, pays bigger dividends in the long run. But if I were a student, I obviously would have other prorities (although if I were a student, I probably would not worry about the noise as much).