The Brazos Performance Preview

[Stukov]

And the heat created by the chip is more or less identical to the power it consumes.

No it is not. While TDP will increase as power increases, rated TDP not a good indicator of power draw except it can tell if it should be low or higher ranges.

[mAJORD]

average power = heat

[JumpingJack]

average power = heat

Power has units of Joules/sec and heat has units of Joules

average power =/= heat

What you said is the equivalent of speed = distance

[Dresdenboy]

Because Zacate platform is expensive.
I have heard of rumours that Intel plans 1C SB against Zacate.

I'm still looking for more data on those SB models.

A 2C Sandy Bridge could be ~158 mm^2 because 2 cores take 2x29.5 mm^2 and the dead space is ~8 mm^2. So 225-59-8 is 158.

Wouldn't 2C Llano be more of a target for lower core SBs?

[duploxxx]

I'm still looking for more data on those SB models.

A 2C Sandy Bridge could be ~158 mm^2 because 2 cores take 2x29.5 mm^2 and the dead space is ~8 mm^2. So 225-59-8 is 158.

Wouldn't 2C Llano be more of a target for lower core SBs?

from AMD perspective yes, but it is Intel that has to position something above Atom

[-Boris-]

No it is not. While TDP will increase as power increases, rated TDP not a good indicator of power draw except it can tell if it should be low or higher ranges.

Every single watt that a circuit consumes is converted to heat. What would these watts else go? Other types of radiation would be very low compared to the heat.

Power has units of Joules/sec and heat has units of Joules

average power =/= heat

What you said is the equivalent of speed = distance

Ok, energy consumption = heat output. Or power consumtion = heat output per second.

[Mechanical Man]

No, average power = average heat produced.

There is no moving parts or anything other. Some might be converted to sound trough vibrations in certain systems but in CPU we can safely say that 99,9999% of used power is converted to heat.

TDP = thermal desing power. So it is at some max that chip will consume. There is no point to give other than coarse TDPs.

In end TDP is just some max that chip will use in some test case. Usually used power is less even in "full" usage situations.

[Calmatory]

TDP doesn't say much about power consumption really. Indirectly yes, but thats all. The products usually fall into certain categories by their TDP, e.g. 9 W, 14 W, 19 W, but that just means that they use the same cooling solution and that cooling solution is designed to dissipate that much of heat without letting the chip being cooled to reach it's maximum junction temperature set by the manufacturer. We can with probability say that a part with TDP of 10 W will consume more than a part with TDP of 6 W, but we can not be certain of it. TDP does not prove it any way. For example it could be possible that for some unknown reason the manufacturer slapped a higher TDP for the part just to make sure it runs cooler.

[-Boris-]

TDP doesn't say much about power consumption really. Indirectly yes, but thats all. The products usually fall into certain categories by their TDP, e.g. 9 W, 14 W, 19 W, but that just means that they use the same cooling solution and that cooling solution is designed to dissipate that much of heat without letting the chip being cooled to reach it's maximum junction temperature set by the manufacturer. We can with probability say that a part with TDP of 10 W will consume more than a part with TDP of 6 W, but we can not be certain of it. TDP does not prove it any way. For example it could be possible that for some unknown reason the manufacturer slapped a higher TDP for the part just to make sure it runs cooler.

Well that's pretty much what I usually say. Now I wasn't talking about TDP. I was only saying that the power a chip consumes is the same as it dissipates in heat.

[alpha0ne]

Being an HTPC tragic this news changes everything for me, at last the right compromise (IMHO) of GPU/CPU power with a very small power draw so true on 24/7 HTPC's are finally doable