some interesting info on asrock's AM3+ page

[moriz]

http://www.asrock.com/news/events/2011AM3+/

seriously funky english aside, there's some tidbits of info that might be worthwhile:
-the AM3+ socket provides for better power management, so it looks like bulldozer's power gating features are based on the socket, not the chipset
-the larger pin holes MIGHT cause issues with AM3 boards that supposedly supports AM3+ cpus, since the pins themselves might be thicker as well. this is very unlikely
-if the AM3 socket is capable of 110A maximum, and the typical voltage of current phenom IIs is 1.375V (150W TDP max), then assuming for the same maximum TDP on AM3+, 145A maximum means that bulldozers operate at 1.034V. of course, its possible that bulldozer operate at the same voltage, which means that they'll have a max TDP of 200W, which is pretty ridiculous. maybe the coolermaster V10 finally have a use after all

so, any thoughts?

[mikeymike]

I find it amusing that ASRock are talking about AM3+ needing higher current as if that's a good thing Point 2 is quite amusing as well

Point 6 about ventilation is making me think "err, wot?" The AM3+ HSF retention system looks like a money-saving design, rather than it helping with ventilation in any significant way. I realise that fluid dynamics is a complex enough topic, and that I've never studied it, but I would be quite surprised if removing a piece of plastic that small would yield a 5C temp reduction. Even if the plastic was completely flush against the side of the HSF (I don't think it is), I don't think it is tall enough to cause that much of a problem.

[just brew it!]

-the larger pin holes MIGHT cause issues with AM3 boards that supposedly supports AM3+ cpus, since the pins themselves might be thicker as well. this is very unlikely

AFAIK it is just the holes, not the pins. Basically they are saying that you're less likely to bend pins of the CPU, since it should be easier to get the pins to align with the (larger) holes. Not that I've ever had much trouble getting the pins in the holes before -- if the CPU doesn't want to seat properly, then you've *already* got bent pins.

-if the AM3 socket is capable of 110A maximum, and the typical voltage of current phenom IIs is 1.375V (150W TDP max), then assuming for the same maximum TDP on AM3+, 145A maximum means that bulldozers operate at 1.034V. of course, its possible that bulldozer operate at the same voltage, which means that they'll have a max TDP of 200W, which is pretty ridiculous. maybe the coolermaster V10 finally have a use after all

I seriously doubt we'll be seeing any 200W desktop parts from AMD. I would bet on the core voltage being enough lower that the TDP will be roughly the same as the current parts.

I find it amusing that ASRock are talking about AM3+ needing higher current as if that's a good thing Point 2 is quite amusing as well

Yeah, next thing you know we'll have enthusiasts trying to overclock their power controllers!

Point 6 about ventilation is making me think "err, wot?" The AM3+ HSF retention system looks like a money-saving design, rather than it helping with ventilation in any significant way. I realise that fluid dynamics is a complex enough topic, and that I've never studied it, but I would be quite surprised if removing a piece of plastic that small would yield a 5C temp reduction. Even if the plastic was completely flush against the side of the HSF (I don't think it is), I don't think it is tall enough to cause that much of a problem.

This one actually makes a certain amount of sense. The "makes the choke 5.4C cooler" comment indicates that they're talking about cooling the voltage regulation components next to the CPU socket, not the CPU itself. (Actually we're more concerned about the temperature of the MOSFETs and capacitors, not the chokes/inductors; but they tend to be located right next to each other, very close to the CPU socket.) By removing the sides of the HSF bracket you're probably improving the airflow across the voltage regulation circuitry, since the air exhausted from the CPU HSF now has an unrestricted path across the surface of the PCB (assuming a traditional downward-blowing HSF). For systems which are being run hard (i.e. close to that 145A limit), this could improve the reliability of the CPU power regulation circuits.