Review request

[imported_Holly]

First of all I am sorry if this post should have been placed elsewhere; have not found better place to put it.

I would like to ask for review of CPU economy - price per performance unit (office use, multimedia, gaming) and performance per kWh power. Especially with new Core 2 CPUs with 's' it could be worth comparing. With lower consumption the initial expenses (~+50USD) could be back within few months of running the systems.

Also for C2Q 9xx0s I would like to ask for some comparison with non 's' versions. Performance, temperatures, overclockability etc.

Other than this, thanks AT for keeping me well informed about the news in the industry.

Sincerely,
Jan Hyka

 

[Denithor]

Wikipedia - Power consumption comparison tables

Covers most of the generations of chips from Intel & AMD.

Two things to note on these tables -

1) Power - this is simply the power consumption the chip is -rated- for. Which is very different from what they actually consume depending on several factors (idle or under load, stock or overclocked, etc). Also note that the rating is even a bit deceptive as Intel rates the e8600 (3.33GHz, 6MB cache, 333FSB) and the e5200 (2.5GHz, 2MB cache, 266FSB) for the same 65W TDP. Riiight...these two chips are going to consume exactly the same amount of energy in use. :roll:

2) Speed : Power ratio simply divides raw clockspeed by -rated- power consumption. For dual/quad cores they just multiply the clockspeed by the number of cores. This is not a good comparison for several reasons: different architecture CPUs do vastly different amounts of work/cycle so clockspeed is one of the worst comparisons among different generations/architectures they could make. Also, the whole "# cores x clockspeed" is a ridiculous assumption to make, that's in essence saying that a Q6600 (quad 2.4GHz) will perform at the level of a 9.6GHz single-core CPU (not even close).

Now, even with those qualifications, the tables should give you some idea how the chips compare to each other.

For the second part of your question, comparing a 95W Q9400 to a 65W Q9400s I think is the best idea.

95W - 65W = 30W power savings
30W * 24 hours/day * 365 days/year = 262,800 Wh/year
262,800 Wh/year * 1 kWh/1,000 Wh = 262.8 kWh/year
262.8 kWh/year * $0.10/kWh = $26.28 savings/year

Note1: Assuming CPU is on 24/7/365 and actually uses 65W.
Note2: Plug in your local $/kWh for your savings.

So if the Q9400s is priced around $270 it will cost about $40 more than the standard model and your power is ~$0.10/kWh you could expect payback in about 1.5 years, after which it would slowly pay for itself in energy savings. Plus you're using less total power so it's kind of a "green" thing to do. Or at least that's what Intel's marketing department will tell you.

 

[dbcooper1]

And in the warmer climates you're paying twice for that heat- once to produce it and once to run your A/C to get rid of it so the payback period will be even shorter. My office is typically 5-10 degrees warmer that the rest of the house unless I adjust the vents accordingly by season.

 

[imported_Holly]

Thx for the post Denithor.

Though, the information you provided is quite obvious; all of us can read Intel and AMD spreadsheets. And most of AT readers will know that TDP and real power consumption is different. That's why I asked for the review - to show real values.

Same for power consumption & "spare"... you can do 95-65=30 but that's not the real world honestly. Also, you can barely find any other info about 's' line of C2 cpus other than 65W TDP. No word about voltages, no word about overclocking, no word about 'hidden suprises' in ways of lower performance etc.

The raw clockspeed / power consumption is just laugh because these days performance is not directly proportional to the CPU frequency. Thinking about it it never was. What I think of is kind of benchmarking it the way of
1. copy C: drive to other disk
2. start power consumption measurement
3. winrar the copied volume
4. stop power consumption measurement
5. result... aha, this machine with that CPU took 1.12kWh, that one took 2.3kWh

then do something like that with other usual benchmarks and get results. Honestly if you build 'computer for work' you usualy don't need quad SLI and Extreme Edition CPUs, but more likely you are interested in getting things done reasonably fast and at reasonable cost. And yes, even the gamers (or their bills paying parents) might be interested in knowledge how much power does 1 hour of Quake-Doom-Warcraft-Whatever take. Having 180FPS computation (and another central heating body in a room) on 60FPS LCD screen is really worth nothing (since 2/3rds of the calculations are totaly useless).

Long story short that's why I ask AT to make review to provide real life informations. And yes, I know I ask for few hundred hours of slavery.

edit: examples of benchmarking I had on mind...

 

[NXIL]

Hey Jan Hyka,

Thx for the post Denithor. Though, the information you provided is quite obvious; all of us can read Intel and AMD spreadsheets.

Google is your friend!:

http://www.tomshardware.com/re...-consumption,1707.html

http://www.tomshardware.com/re...t-c-state,2106-11.html

http://www.tomshardware.com/re...efficiency,2069-6.html

http://www.tomshardware.com/re...fficiency,2069-12.html

http://www.tomshardware.com/re...e7200-g31,2039-13.html

http://www.tomshardware.com/re...on-Efficient,1997.html

In our Munich lab?s duel of the energy-savers, the AMD Athlon 64 2000+ beats the Intel Atom 230 in energy consumption and processing power. Each of the systems was based on a desktop platform. The Achilles heel of the Intel system is its old system platform with the 945GC chipset, while AMD offers a more modern 780G platform.

http://www.tomshardware.com/re...md-power-cpu,1925.html

http://www.tomshardware.com/re...-consumption,1750.html

http://www.hardwarezone.com.au...hp?cid=2&id=2304&pg=13


With TDP measurements, manufacturers always specify the maximum power than a chip will possibly draw in a real world scenario to ensure that the power envelop is not exceeded. In the case of the Athlon X2 BE-2350, its 45W TDP is 20W lower than the nearest competition, but the actual savings over time for the whole platform is not as impressive. Processors do not operate at 100% capacity, i.e. maxing out its TDP 100% of the time. We also know that Intel's Core microarchitecture has more power saving technologies built into it than the aging Athlon 64. This is very evidently shown in our power consumption tests in previous articles. Yet for the new Athlon X2 BE-2xxx series, AMD managed to one-up Intel on this front (though not by much). Fully idle, the Athlon X2 BE-2350 could achieve up to 10W power saving over the Pentium Dual Core E2160 or Core 2 Duo E4300. However, once the processors are loaded, this advantage was reduced to a slim 2 - 5W difference.

http://www.pcper.com/article.p...=527&type=expert&pid=4

http://news.softpedia.com/news...Efficiency-90755.shtml

http://blogs.zdnet.com/storage/?p=165

Anyway, the power efficiency of any system is going to depend on:

the power supply, not just the CPU: how efficient it is at different ranges in converting wall juice to PC power.

The CPU, how efficient it is at idle and load, and also how much time the PC spends at idle and load.

The rest of the system: the video card? they suck up a lot of juice too.

Memory...hard drives....etc.

For power efficeincy for an office PC: Mac Mini?

HTH

Edit: for tone.....looks like Jan not a native english speaker? And he speaks Engrish better than I speak anything else.....

NXIL

 

[imported_Holly]

Hello NXIL,
thanks for the links provided. Even though I would have wished for more through testing (longer time, various use, 'score per Wh' for more cpus) the links you provided answered many of my questions.

Ye, I am aware that every power consuming component impacts total consumption of the PC, but that would hardly be possible to test and then present in a readable way. In fact only two important facts are idle work power consumption and score per Wh. Running a process on one machine taking 200W and on other one taking 250W won't save you the 50W, if the second one takes half time to finish it.

Aaaand yep, I am not a native english speaker; I come from Czech republic (central Europe). Learned the most of my english language skills back in days of playing one MMORPG (DAoC).

Once again, thanks for the links.

 

[NXIL]

Hi Jan,

looks like TomsHardWare (apologies to the locals) may be of help to you, as they are interested in the same sorts of questions and answers:

http://www.tomshardware.com/re...ano-power,2036-13.html

Running a process on one machine taking 200W and on other one taking 250W won't save you the 50W, if the second one takes half time to finish it.

I did not realize this article was more than a year old, but, here is one that answers your exact question:

http://www.tomshardware.com/re...-consumption,1707.html

Leading people into buying a low-power, high efficiency processor for desktop use is misleading, because the user could end up in a scenario in which her or his low-power components will have to run at substantial loads all the time. In such a case, all the power savings could be gone, and an average system might turn out to be more efficient.

The Core 2 Duo E6400 system completed the SYSmark 2007 Preview run 14 minutes earlier than the Athlon 64 X2 5000+. As a consequence, the Intel system went back to an idle state earlier, which of course results in considerably less total power consumed. During the same time, the AMD system had to stay at a higher activity level for a longer period of time, which eventually meant that it required as much as 50% more power than the Intel system!

HTH

NXIL

 

[Denithor]

Here's a few newer ones for you to consider.

Power consumption: AMD 780G/790GX - Intel G35/G45 - nVidia Geforce 9300

Note these are run with just the IGP so you get an accurate idea of relative power usage due to the system rather than a huge honkin' GPU. The AMD systems use a PhI 9950 and the Intel/nVidia systems use a Q9300.

IGP Power consumption: 780G, GF8200, G35

Better selection of processor there.

i7 965 versus QX9770: power consumption & benchmark results

For the latest & greatest. i7 tends to get work done faster while simultaneously using less power doing it. This has been the trend over the last couple of generations of chips (especially with die shrinks - the power savings has been significant).

xbitlabs: Comparison of inexpensive quads [and one dual
xbitlabs: Comparison of inexpensive dual-cores

Great lineups here. You could look for a benchmark related to your tasks & divide the performance by the watts in the power consumption charts to get a ratio.

It seems like I've seen someone do a chart like this (performance/watt comparison) but I'm having a hard time digging it up.

Let me know if you want some more.

EDIT: Although kinda old again, here's something closer to what you're looking for, I think: Performance vs Power consumption. Keep in mind these are older chips (e6600 vs several 90nm & 65nm X2 chips) and that performance has increased along with a reduction in watts with Intel's move to 45nm chips. Basically they're getting more & more power efficient all the time, along with boosting performance. Good times!