sandybridge E, 180W??

[Rifter]

Xbitlabs - Total system power consumption - idle:

i7 2600k @ 4.7ghz = 42.8 W

Xbitlabs - Total system power consumption (CPU loaded only) - load:

i7-2600k @ 4.7ghz = 190.5 W

CPU Power consumption between load and idle = 190.5 - 42.8 = 147.7 Watts

Looking at less efficient processors, they can easily consume 250W+ of power in overclocked states. Phenom II X6 @ 4.0ghz gets up there.

Here from another source (full system power consumption - only CPU loaded):

Core i7 860 stock - Idle = 150.5 W
Core i7 860 3.7ghz - Load = 425.7 W

Difference between Idle and Load State = 275.2 Watts (no GPU load!). This is coming from overclocking the CPU and loading it.

Here is another CPU Power Consumption analysis:

Core i7-960 3.07ghz (Total System - CPU Loaded only) = 190 W
Core i7-960 4.2ghz (Total System - CPU Loaded only) = 317 W

Guess what? Where did that extra 127 Watts come from? From increased VCore/QPI voltage on the CPU and increased frequency which resulted in a huge CPU power consumption increase. The increase had nothing to do with RAM, Hard Drives, GPU, Mobo since the power consumption of those components is already included in the original 190W.

I agree, ALOT of people underestimate by a very large margin how much power you draw when you go for more than a few hundred Mhz overclock.

The phenom II is especially bad for this it draws massive power(200w+) at over 4Ghz it must be a very leaky chip.

 

[RussianSensation]

We would melt the pins on the board if the chip pulled 200watts and most def burn shit up at 280.

1. This did happen on Socket 1156 with Foxconn sockets, which were later replaced by Lotes by most manufacturers.

2. For extreme overclockers on LN2, top-of-the-line boards deliver up to 600 Watts of power to a single CPU.

EVGA X58 Classified3

lol at 250 watts you would need 22amps of current on the 12v rail going into the cpu

It's more complicated than that. It depends on the CPU / integration of the Northbridge.

Take a look at the Core i7-860 for example. It has an integrated memory controller. The CPU load in overclocking is primarily carried by the 12V CPU line:

Notice, together with the motherboard you are looking at more than 20A by the time you reach 4.0ghz.

But now, look at the i7-950.

The CPU alone is drawing 17A, but the 5V rail has increased from 8.6 to 12A!

"The current going along the 12 V processor line more than doubles during overclocking. At the same time you should remember that the CPU takes some power that goes to the integrated North Bridge from the mainboard. Therefore, the increase in power consumption along the 5 V line should probably also be assigned to the CPU." - Xbitlabs

 

[pantsaregood]

i7 3820 is ever so slightly less expensive than a 2600K? I suppose the lack of overclocking would make it less attractive for enthusiasts. Are all of the above hex, or is the 3820 a quad?

I'm also confused by the naming scheme. I thought the "2" in 2600 meant "2nd generation", so I expected "3" to belong to Ivy Bridge, but instead they are going ahead and giving the "3" designation to high-end Sandy Bridge parts, instead of 28xx and 29xx?

Intel's naming convention makes absolutely no sense. Nehalem and Westmere were the first generation, and Sandy Bridge/Ivy Bridge were set to be the second. For whatever reason, Intel decided to throw logic down the toilet and call SB-E, which uses unmodified Sandy Bridge cores, the third generation of Core i-CPUs.

 

[Nemesis 1]

Xbitlabs - Total system power consumption - idle:

i7 2600k @ 4.7ghz = 42.8 W

Xbitlabs - Total system power consumption (CPU loaded only) - load:

i7-2600k @ 4.7ghz = 190.5 W

CPU Power consumption between load and idle = 190.5 - 42.8 = 147.7 Watts

Looking at less efficient processors, they can easily consume 250W+ of power in overclocked states. Phenom II X6 @ 4.0ghz gets up there.

Here from another source (full system power consumption - only CPU loaded):

Core i7 860 stock - Idle = 150.5 W
Core i7 860 3.7ghz - Load = 425.7 W

Difference between Idle and Load State = 275.2 Watts (no GPU load!). This is coming from overclocking the CPU and loading it.

Here is another CPU Power Consumption analysis:

Core i7-960 3.07ghz (Total System - CPU Loaded only) = 190 W
Core i7-960 4.2ghz (Total System - CPU Loaded only) = 317 W

Guess what? Where did that extra 127 Watts come from? From increased VCore/QPI voltage on the CPU and increased frequency which resulted in a huge CPU power consumption increase. The increase had nothing to do with RAM, Hard Drives, GPU, Mobo since the power consumption of those components is already included in the original 190W.

AT did the power draw of the Cpu . and it wasn't anywhere near those numbers . If your stupid I guess ya can make anything overdraw.
An eletric motor for example . Running a belt . If the belt tension is correct may pull 15 amps . Over tighten the belt it will pull 25 amps.
There be a big differance between Xbits numbers and ATs numbers . So lets use X-bit numbers , Keep the high through out the low . Makes for great negitive hype . You just made me buy a BD . Watch what happens.

 

[grkM3]

Those i7s are 6 core chips,we are talking about quad 2600ks and those watt numbers mean nothing because they could of jammed to much voltage in the chip for those runs.like using 1.4 instead of 1.3

I used watts that were low enough voltage to pass 5 passes of intel burn test.i could of easily ran less voltage to show a lower wattage draw.

You can see in my pic that on idle 4.5 i was jamming 1.4 volts into and this chip can do runs windows at 1.33 at those speeds but not 100% stable

 

[RussianSensation]

Those i7s are 6 core chips,we are talking about quad 2600ks and those watt numbers mean nothing because they could of jammed to much voltage in the chip for those runs.like using 1.4 instead of 1.3

I don't think anyone said that 2500k/2600k CPUs consume 200-400 watts. That doesn't mean it's not possible with other/older generations. 2500k/2600k are extremely efficient even when overclocked. I think the general point was that 180W for a high-end CPU isn't that much, esp. not for a 6-core, since we've been hitting in excess of 180W with overclocked 920/930/860/870s for years.

 

[grkM3]

I don't think anyone said that 2500k/2600k CPUs consume 200-400 watts. That doesn't mean it's not possible with other/older generations. 2500k/2600k are extremely efficient even when overclocked. I think the general point was that 180W for a high-end CPU isn't that much, esp. not for a 6-core, since we've been hitting in excess of 180W with overclocked 920/930/860/870s for years.

go back a few pages,this all started when a poster said 2500/2600s consume 60 watts idle and 280watts on load.

 

[RussianSensation]

go back a few pages,this all started when a poster said 2500/2600s consume 60 watts idle and 280watts on load.

I must have missed that part. My bad!

 

[SHAQ]

My 920 shows 147 watts. If this is 180 without overclocking then it won't be able to overclock much. I don't believe there is an air cooler that can handle it. At least not when stress testing. LinX will shut down pretty quickly I would imagine even at stock speeds.

 

[piesquared]

I wonder what's coming down the pipe that intel needs to push their architecture to the limit at 180W.

 

[OCGuy]

I wonder what's coming down the pipe that intel needs to push their architecture to the limit at 180W.

The OP posts a rumour article and then comes in the thread later and acts as if it is indeed fact, and then suggests Intel must be scared about something AMD is doing.

Come on..

Although I can say that even if it was true about 180w, it wouldnt bother me. I have a feeling I may not be lucky enough to have a SB-E, however, because the advantage over my current setup may just not be worth the $$$$.

 

[blackened23]

I wonder what's coming down the pipe that intel needs to push their architecture to the limit at 180W.

Its all rumor, but if it does turn out to be true these chips will get warmer than lynnfield / bloomfield which is not good.

 

[RussianSensation]

My 920 shows 147 watts. If this is 180 without overclocking then it won't be able to overclock much. I don't believe there is an air cooler that can handle it. At least not when stress testing. LinX will shut down pretty quickly I would imagine even at stock speeds.

147 Watts @ 3.9ghz on a 920 at load is way off. Your CPU can draw 130W at stock 2.66 with Turbo alone. You guys should stop using Gigabyte's software as it's not accurate.

Try SiSoftware Sandra or AIDA64/Everest. They will show much higher / realistic numbers.

Modern air coolers can dissipate 200-300W btw (esp. the top of the line ones).

There is a ton of misconception about CPU power consumption figures in this thread.

Look at this graph:

There is no GPU load here. The massive increases in power consumption are coming almost entirely from the CPU and northbridge on the mobo.

There is absolutely no way that an i7 920 @ 3.9ghz consumes only 147 Watts of power at 100% load unless you have some amazing chip that can do so at 1.1V.

I have provided at least 3 separate websites (Bit-Tech.net, Xbitlabs and Toms) which all show that overclocking alone can add 150+ Watts of power to a 45nm CPU. Now, you still have to add the original amount of power that CPU was using at stock speeds.....

Just think about it logically, your CPU is rated at 130W TDP, while i7 860 is rated at 95W TDP. Both of those CPUs ship with tiny stock heatsinks. You are telling me by only going to 150W on the CPU, your Megahalems is barely able to contain such a processor at 3.9ghz at 63-65*C? It wouldn't even break a sweat if your overclocking just added another 20-30 Watts on top of stock. Your CPU alone is pulling 200W+ of power, easily. Overclocking the i7-950 from 3.06ghz to 4.3ghz alone adds 150Watts. What about the power consumption at load for a stock i7-950?

 

[piesquared]

Its all rumor, but if it does turn out to be true these chips will get warmer than lynnfield / bloomfield which is not good.

True. These rumors surfaced for nehelam too, and many denied it could be possible. Unsurpsingly though, upon release intels assigned TDP was far exceeded. No doubt it will be the case here as well. The interesting part though, is why do they need to push it to such limits?

 

[grkM3]

Iv never seen a stockintel chip ever hit its tdp running stock volts and clocks.

 

[notty22]

Far exceeded ? Only when overclocked.

 

[Rifter]

My 920 shows 147 watts. If this is 180 without overclocking then it won't be able to overclock much. I don't believe there is an air cooler that can handle it. At least not when stress testing. LinX will shut down pretty quickly I would imagine even at stock speeds.

Most top tier coolers will handle 180-200w+, such as NH-D14, Silver Arrow, Megehalems.

My power draw from the wall went up almost 100w with just my CPU overclock so im sure its drawing in the 200w range and my NH-D14 does a great job of keeping it cool. Handles prime or linx overnight with no issues.

 

[blackened23]

147 Watts @ 3.9ghz on a 920 at load is way off. Your CPU can draw 130W at stock 2.66 with Turbo alone. You guys should stop using Gigabyte's software as it's not accurate.

Try SiSoftware Sandra or AIDA64/Everest. They will show much higher / realistic numbers.

Modern air coolers can dissipate 200-300W btw (esp. the top of the line ones).

There is a ton of misconception about CPU power consumption figures in this thread.

Look at this graph:


There is no GPU load here. The massive increases in power consumption are coming almost entirely from the CPU and northbridge on the mobo.

There is absolutely no way that an i7 920 @ 3.9ghz consumes only 147 Watts of power at 100% load unless you have some amazing chip that can do so at 1.1V.

I have provided at least 3 separate websites (Bit-Tech.net, Xbitlabs and Toms) which all show that overclocking alone can add 150+ Watts of power to a 45nm CPU. Now, you still have to add the original amount of power that CPU was using at stock speeds.....

Just think about it logically, your CPU is rated at 130W TDP, while i7 860 is rated at 95W TDP. Both of those CPUs ship with tiny stock heatsinks. You are telling me by only going to 150W on the CPU, your Megahalems is barely able to contain such a processor at 3.9ghz at 63-65*C? It wouldn't even break a sweat if your overclocking just added another 20-30 Watts on top of stock. Your CPU alone is pulling 200W+ of power, easily. Overclocking the i7-950 from 3.06ghz to 4.3ghz alone adds 150Watts. What about the power consumption at load for a stock i7-950?

Great info here, good to know how much power draw overclocking really adds. I remember my old lynnfield systems (overclocked) made a noticable increase in my power bill, lol...those 870s ran crazy hot when oc'ed.

 

[grkM3]

I just put my asus MIVE board back together with my 2600k(sending my ud7 back)and ran cpuid HWmonitor with the chip over clocked to 4.7ghz on all 4 cores and Im running 4 sticks of ram at 1.65 volts with hyperthreading on and I never saw over 114watts running 8 threads of intel burn test.

I lowered the vcore a tad and got it to do 1 pass with 8 threads and it never went over 112 watts.

 

[blackened23]

Right, the sandy bridge is extremely power efficient - most of the discussion is centered on the older lynnfield / bloomfields which were extremely power hungry when overclocked - and it seems like SB-E will be more power hungry than those chips were.

The i7-870 and 920 chips in particular were pretty crazy on power draw.

 

[ed29a]

Ya its been talked about already. and for any to say They know what a particially locked SB-E is good bad or ugly is eating froot loops for breakfest. The money is pretty dang good

My brain just exploded. Can someone translate this?

 

[grkM3]

Right, the sandy bridge is extremely power efficient - most of the discussion is centered on the older lynnfield / bloomfields which were extremely power hungry when overclocked - and it seems like SB-E will be more power hungry than those chips were.

The i7-870 and 920 chips in particular were pretty crazy on power draw.

Lol didnt you say sandys consume 60 watts idle and up to 280 on load when overclocked.

Sandy e will consume a tiny bit more than regular sandys because its got no gpu so adding a few cores will be a wash.

Can someone with a 2500 see how much power it draws at load being stock clocked?

I bet it wont go over 75 watts and its tdo is 95 i think.

 

[toyota]

well this is about all I can give you. at idle my pc uses 72 watts according to my UPS. at full load the entire pc is using 138 watts at the wall for "maximum" test in IBT. btw that's only at 3.3 since I have turbo disabled at the moment.

Xbit said 62 watts under load for the 2500 cpu by itself. http://www.xbitlabs.com/articles/cpu/display/core-i5-2500-2400-2300_10.html#sect0

 

[grkM3]

well this is about all I can give you. at idle my pc uses 72 watts according to my UPS. at full load the entire pc is using 138 watts at the wall for "maximum" test in IBT. btw that's only at 3.3 since I have turbo disabled at the moment.

Xbit said 62 watts under load for the 2500 cpu by itself. http://www.xbitlabs.com/articles/cpu/display/core-i5-2500-2400-2300_10.html#sect0

If you want you can download cpuid hwmonitor for free and it shows how much just the cpu is drawing.Do you have the gpu on on the chip? or are you using a card?

 

[toyota]

If you want you can download cpuid hwmonitor for free and it shows how much just the cpu is drawing.Do you have the gpu on on the chip? or are you using a card?

P67 board so no to the gpu on the chip.

EDIT: 85.01 watts but how accurate is hwmonitor because its only showing 8.27 for my 12V.