AMD Athlon II X2 240 2.80 GHz

[vagxtr]

Thanks. That's seems reasonable cause it was weird enough that AMD suddenly drop down AMD-V support from the newer cpus and ollow intels devious path by not including VT or SpeedStep in their entry cpus just to profit on their more lucrative products.

 

[Omega]

my only wish is that you find a way to find out the CPU's actual power use, as opposed to just the entire PC - find the lowest clocked CPU you can, slap it in and underclock and volt as low as you can (say, a single core sempron at 600MHz) and then do wattage tests so you can figure out how much the CPU itself is using with some degree of accuracy.

I will run some tests with Sempron LE-140... I see what are you getting at, but i fear those measurements will be inaccurate as TDP

 

[Mussels]

I will run some tests with Sempron LE-140... I see what are you getting at, but i fear those measurements will be inaccurate as TDP

i've got my media PC down to 50W idle, and thats with a 10W IDE HDD and 4 sticks of DDR2 ram

you should be able to get something within 10% of the real numbers - moreso if you use say, two PSU's and run the HDD and fans off a second PSU (use onboard VGA if possible, while determining numbers)

 

[Omega]

Please think twice before suggesting something like that :shadedshu
That would mean I'd have to have two platforms for testing Intel/AMD CPU's, and another two setups just to check the power consumption of those CPU's.

What I can try to do is go as low as possible on speed/voltage with Sempron LE-140 on current AMD test platform and that would give me some insight of platform consumption "without" CPU. But still i don't see how could i measure the Load power draw, as you can't run memory and vga on another PSU

 

[Mussels]

its just an idea - if its not practical its not practical.


you can find out VGA and ram with ease - test with one stick of ram at idle and load, stick in another stick. measure the difference - repeat til all slots are filled.

you can use the change as each ram stick was added to determine how much wattage each stick used, with knowledge of any variations based on single/dual channel changes.

as for VGA, find a 512KB PCI card that can barely boot windows XP and do some power testing - then repeat with your real card, to find out how much it uses


(what, you told me to think about this )

 

[Omega]

I appreciate any suggestions... no matter how complicated they sound
Anyways, I will do some testings and let you know the results

 

[FilipM]

Please think twice before suggesting something like that :shadedshu
That would mean I'd have to have two platforms for testing Intel/AMD CPU's, and another two setups just to check the power consumption of those CPU's.

What I can try to do is go as low as possible on speed/voltage with Sempron LE-140 on current AMD test platform and that would give me some insight of platform consumption "without" CPU. But still i don't see how could i measure the Load power draw, as you can't run memory and vga on another PSU

You can run Mobo (mobo/cpu/ram) on one PSU, VGA and peripherals on another PSU. (Those Thermaltake drive bay PSU's - only run VGA on them)

1. Try one stick of ram, then try two, the diference in power consumtion measured will tell you the consumption of the ram (as Musseles suggested).

Take out the ram out of the total thing, and you end up with CPU+mobo, now if we can have a guess at how much the board consumes it would be nice - ending up only with CPU.

2. For VGA is easy, just get another PSU, run only the VGA on it and measure consumption again - you end up with the VGA.

Same process as for the VGA goes for HDD's DVD drives, Floppy, etc - in a few words - you need two PSU's.

 

[Omega]

You are forgetting that VGA draws power from the PCI-E slot. Still leaves the MBO power draw unknown.

 

[Mussels]

Take out the ram out of the total thing, and you end up with CPU+mobo, now if we can have a guess at how much the board consumes it would be nice - ending up only with CPU.

thats where the single core CPU comes in (sempron, in this case) - clock it down really low at really low volts, and its idle will be practically nil, with load <10W

You are forgetting that VGA draws power from the PCI-E slot. Still leaves the MBO power draw unknown.

thus my suggestion of an antique PCI graphics card, the pure 2D kind with <4MB of ram - the kind that use 5W of power
you wouldnt be able to use anything other than XP, but you can do idle and (non 3D) load, to determine the values for CPU, motherboard and ram

 

[Omega]

Aren't we going with this the wrong way?
Why exclude all of the hardware when I need baseline for complete test setup under IDLE/LOAD "without" the CPU.

I just clock down the Sempron as much as it can go, do the consumption idle/load measurements and remove say ~10W from load results. When measuring system power draw with stock CPU, that result is reduced by "baseline" system power draw and it should show the aprox. CPU draw only.

Right?

 

[FilipM]

Yep, if you want a whole system without a CPU, clock the Sempron down as much as it can go, and remove ~10W from the total power consumption - you end up with a baseline to start from for an AMD system

 

[Omega]

Right... and I could do the same thing for LGA775 with Celeron 430. Now, all we need is single core LGA1156 and LGA1336 processors

I'll give it a go on AMD platform for starters, and see what numbers come up.

 

[FilipM]

Hmm, i've just been on Everest, and in the Computer/Sensor section at the bottom there are two interesting things about CPU's - Current Values (as in ampers) and Power Values (as in watts).

Now i've just been idling and it showed 9.78W / 7.11A. Can this be taken as accurate or what? If i calculate the voltage according to those, it looks about right.

 

[eidairaman1]

I expect Athlon II X4s to have 4MB L2 (1MB for every core) shortly

 

[vagxtr]

thats where the single core CPU comes in (sempron, in this case) - clock it down really low at really low volts, and its idle will be practically nil, with load <10W



thus my suggestion of an antique PCI graphics card, the pure 2D kind with <4MB of ram - the kind that use 5W of power
you wouldnt be able to use anything other than XP, but you can do idle and (non 3D) load, to determine the values for CPU, motherboard and ram

You dont need to go that far. Some RivaTNT 16MB or RadeonVE 32MB on PCI slot shouldn't use much more than 3-5W in 2D, hardly 10W in 3D mode (How much power could be dissipated from card with small 40x40x10 passive heatsink thats lukewarm in 3D? Just for peace of mind and server usage stability imho)
But why should we go in such trouble when the whole chipset (IGP+SB) on modern 785G shouldnt use 10W when loaded with 14USB, 6SATA, *IDE* and when comes down to just one SATA HDD (Spinpoint F2 1.5TB claims only 5.2W in write/read) and no external USB activity (legacy Keyboard, mouse shouldnt use much power i guess, and down clocked memory to ddr2-4200 @1.8V (533Hz) shouldn use more than 3-4W (when ddr2-800/1066 @1.8V claims less than 5W consumption). Dont forget to loose timings if possible to 4-4-4 from 3-3-3.
So the whole mobo(excl.cpu-pwm)+hdd+memory setup could use as low as 10-12W with heavy hdd-memory activity. And btw i dont see much of use to downclock AthlonLE/SempronLE so deeply to 600MHz only cause they hardly use 2-3W more on same voltage at 1200MHz. I know that x2 @65nm usually regularly work 1500-1700Mhz@0.825V with default CPU/5 memory divider (600-680MHz) and with these CPU probably didnt go above 20W and cosidering 45nm and only one core these cold be as low as 7W and how low could it be @600MHz <2W.

1. DDR2 memory probably could run at all on AM2 mobo at 600MHz CPU frequency?
2. How could you downclock AMD AM2 cpu beyond 1000MHz eitherway, when lowest allowed multi is x5 and lowest HTT-FSB is 200MHz??
3. Power usage should be lower on AM3 based boards with LV-DDR3 (1.35V) but is that memory supported by AMDs PII? What is lowest memory frequency (divider also) on AM3 boards?

 

[Mussels]

You dont need to go that far. Some RivaTNT 16MB or RadeonVE 32MB on PCI slot shouldn't use much more than 3-5W in 2D, hardly 10W in 3D mode (how much power could be dissipated from card with small 40x40x10 passive heatsink?)
But why should we go in such trouble when the whole chipset (IGP+SB) on modern 785G shouldnt use 10W when loaded with 14USB, 6SATA, *IDE* and when comes down to just one SATA HDD Spinpoit F2 claims (5.2W in write/read) and no external USB activity (legacy Keyboard, mouse shouldnt use much power i guess, and down clocked memory to ddr2-4200 @1.8V (533Hz) shouldn use more than 3-4W (when ddr2-800/1066 @1.8V claims less than 5W consumption). Dont forget to loose timings if possible to 4-4-4 from 3-3-3.
So the whole mobo(excl.cpu-pwm)+hdd+memory setup could use as low as 10-12W with heavy hdd-memory activity. And btw i dont see much of use to downclock Athlons so deeply to 600MHz only cause they hardly use 2-3W more on same voltage at 1200MHz. I know that x2 @65nm usually regularly work 1500-1700Mhz@0.825V with default CPU/5 memory divider (600-680MHz) and with these CPU probably didnt go above 20W and cosidering 45nm and only one core these cold be as low as 7W and how low could it be @600MHz <2W.

1. DDR2 memory probably could run at all on AM2 mobo at 600MHz CPU frequency?
2. How could you downclock AMD AM2 cpu beyond 1000MHz eitherway, when lowest allowed multi is x5 and lowest HTT-FSB is 200MHz??
3. Power usage should be lower on AM3 based boards with LV-DDR3 (1.35V) but is that memory supported by AMDs PII? What is lowest memory frequency (divider also) on AM3 boards?

you make valid points, as for the low clocks... easy, turn the FSB below 200 board depending (i've seen chips with 4x multipliers, however they were older ones)

 

[vagxtr]

you make valid points, as for the low clocks... easy, turn the FSB below 200 board depending (i've seen chips with 4x multipliers, however they were older ones)

Wait, is this possible on regular boards? Is some BIOS mod needed for it?

 

[EnglishLion]

I'm glad to see a good review of this chip - I just advised a mate to buy one!

As for power consumption, I don't think you'll be able to arrive at a CPU only value. IMO it would be better to offer a minimum total system value. In other words try to run the component in a system with minimal power draw.

Choose a low power motherboard, one stick of RAM, no optical drive, 2.5" HDD or SDD, onboard VGA. This would show how low you can go with the CPU, after all no-one is ever going to run just a CPU on it's own!

Some of the socketed ITX boards offer low power running as a priority.

BTW my HTPC idles at 63W (not quite as good as my server @ 26W though)

 

[Omega]

Way too complicated, time consuming and not really that practical.
We will stick with the total system consumption for time being, and for future test I will try to do some experiments.

There is an "suggestions" thread opened, so if you have any good ideas, please present them in THIS thread. We are working hard to bring you best CPU reviews possible, but it will be hard without your support. Thanks!

 

[[I.R.A]_FBi]

dugg

 

[vipz]

is there any comparison table between phenom965 be and athlon II 240.I want know the fps difference between two cpu's with same test setup.

 

[Mussels]

gaah, please edit that post - you dont need a quote of the entire review

 

[Omega]

AMD AM3 processors round-up coming up soon.

 

[bmwracer]

Hi, I just joined this forum... Is there a way to get a printable view of the review of this CPU?

Thanks for your help.

 

[vagxtr]

Hi, I just joined this forum... Is there a way to get a printable view of the review of this CPU?

Thanks for your help.

There were that option until Nov/Dec last year. Then they disabled it and now its totally removed .... It's inspired by other portals bad style to disable that feature recently (in last 2-3yrs)