http://en.expreview.com/2008/07/07 [...] -capacity/

I just can't imagine a new dude in the GPU block.

Well, from what Ive read, it has 32 shrunk down PentiumIII cores I believe, uses x86, and can do 2 TFLOPS at DP. That being said, it looks to be an excellent graphics encoder, audio encoder etc or gpgpu. How will it do in a rasterisation enviroment who knows? Im thinking itll do ok, and will need alot of help from game devs to make it shine

More importantly are the financial implications. The price war is hurting the big players, some are them are basically selling cards at a loss so if Larrabee is even just on a par with better margins the manufacturers well be all over it.

Some serious power draw from that beast - not surprising given the raw power it potentially can unleash.

I doubt it. Pentium 3s where manufactured at over 200nm at one stage and only required really low level cooling which indicates a very low power drain. Now factor in Intel's multi core technology where a quad core only draws 50% more power than a dual at the same clock speed + the massive die shrink. Now if it where Pentium 4 cores...

If it were P4, the pipes would be so long, an entire game would be trapped in Larrabee before we saw anything heheh

Personally I think it’s going to struggle against existing video cards as its architecture is completely different from that of the existing ATI and Nivida cards games will need to be seriously re-coded in order to work on video card that uses x86’s chips rather then more traditional shader and pixel pipeline based GPU’s.

300W TDP?

Romulus I thinking more 125-150W TDP.

As we already know, thanks to the E6600/Q6600, dual die chips only take 50% and lets assume that each die takes 8W.

8 dual dies at 12W(dual die takes 50% more power than a single die) hits 96W + that leaves power for the rest of the card including all the memory this beast will need. Plus by then Intel should have its monolithic quad cores sorted out and have done another die shrink offering even more power savings.

Don't forget Intel said using a mid level Larrabee and entire decent performance computer could be built using less than 140W of power.

But it says 300W TDP over there... maybe the 32 cores are really hungry

Im thinking thats a FUD allowed by Intel to throw people off

romulus yeah, 300W TDP with current technology and at 2ghz. Heres an interesting little piece of info from my overclocking. A Q6600 @ 3.2ghz requires 133W, at 3.6ghz it can burn through up to 200W. Small dump in clock speed but the CPU wants a lot more power. So lets assume the die shrink doesn't help, they can still scale back to 1.6ghz which would knock a huge amount of power requirements off. Wait this is Intel, their die shrinks always help.

EDIT: Also they state it will support SIMD which allows for multiple cores from a single physical core much like Hyper Threading does. That means it could only actually have 16 cores halving the power requirements.

The Pentium 3 Arch used roughly 1.45v to get to 1.4 Ghz and that at 130nm. Considering that they used 32 processors each with SIMD features which will increase the utilization of the seperate cores, i think 300 Watt is not that far off. It may be a conservative estimate to stay on the safe side, but once all those processors start running at full speed, it will consume a lot of power.

For those of you concerned with that, please don't believe that intel does this to supply us with the next best graphics card. To them, this is something they can use in the server segment. And having 32 OOO cores capable of x86 code running at 2 Ghz each is quite something. Compare their processing power with that of a 32 processor rack and the power consumption doesn't look bad, not bad at all. Add the new space requirements and ease of installation and you can see where this is going...

I was assuming these are going to be modified 'Atom' processors, since a few sites mentioned that these cores were going to process instructions 'inorder'.

X2_Server I had also heard the same but perhaps Slogogob is right and these are the ones meant for servers much like NVidia's Telsa/Quadro and ATI FireGL ranges are in which case that thermal envelope makes sense.