Oh, so now superpi doesn't matter, where as it's intel's fanboi's de facto claim of superiority over AMD?

hypothetical question....would a 900mhz 45nm single core conroe consume more power?

at 2400mhz the new 45nm quad core conroe's are consuming 37 watts under load...... how much power is needed to drive 1 core at 1000mhz?... and if it will spank the crap out of the Atom why is Atom being made at all?

seems like alot of hassle for nothing.

It seems to me that a single thread of the two threads this thing will run at the same time is roughly as fast as a Willamette P4, in a 2w MAX power budget, averaging much less.

seriously, check the clock for clock performance against the 2.4Ghz Northwood.. it actually comes out looking pretty good (or the P4 comes out looking pretty terrible)

For reference a 1Ghz Via C7-M ULV should score around 1:40.
It has a max TDP of 3.5w so i don't know if Intel's Atom makes any sence.
BTW the C7 is built on a 90nm process while the Atom at 45nm.

"A 1.6GHz Atom chip seems to sit right between a 900MHz Celeron M and 1.13GHz Pentium 3-M—slower than the former, and a bit speedier than the latter."

Are you sure that's not backwards? Is a 1.13GHz Pentium 3-M really slower than a 900MHz Celeron M?

Always nice to have news but a Super PI benchmark doesn't say much really, especially given the expected target of such a processor.

Ok, my bad. I should have followed the link and read the article. I didn't understand the different architectures involved in the Celeron M vs. the Pentium 3-M. Now that I think about it, I think the 1.13GHz Pentium 3-M was exactly the processor in my old Compaq Presario.

Is this for cellphone or PDA?

The Eee PC has a Celeron 900, but its underclocked to 650, so I wouldn't consider it necessarily the best example for comparisons.

Would be interesting to see integer and SSE FP performance benchmarked, especially in a multithreaded manner. I'm sure more information will come out in the near future with regards to this.

Handheld PC

The only trouble with this benchmark is that it also only tests x87 floating point, the chip designers may have focused more on SSE2 FPU operations and integer operations.

This is a nice taster, but I think we need a wider spread to be sure.

Thats great performance for a chip that is so small 11 can fit on the face of a penny. I would love to have a cell phone powered by this thing (Why? I don't know either but it would be cool).