Comments on: Intel: We’ve always been serious about microservers. No, really

By: Pilgrim Beart

Pilgrim Beart — Thu, 13 Dec 2012 16:36:18 +0000

fGreat piece.
Intel’s own-fab model doesn’t scale, so it is confined to a shrinking “high cost, low volume” niche.
ARM’s IP model does scale, so the ARM ecosystem can ride the cost vs. volume curve all the way down to ubiquity – as it has already done in mobile.
Also, efficiency (vs. brute force) is the law-of-physics “speed limit” which eventually dominates every sector (our brains consist of billions of slow processors for the same reason).
In the history of mankind more cycles have already been executed on ARM cores than on all other processors combined.

By: Jack

Jack — Fri, 07 Dec 2012 11:25:55 +0000

@Lucian Finally somebody brought out the price-point aspect in the Intel v/s ARM+ (ARM eco-system) debate. Usually, it’s always along the lines of how Intel will blow away ARM+ with their next iCore chip :)

By: PJ Morse

PJ Morse — Fri, 07 Dec 2012 02:23:13 +0000

Microservers will usher in a new era of computing and the implications will be huge. One example is the Billion Node Cloud concept http://www.billionnodecloud.com . Without MicroServer technology, this wouldn’t even be possible. How about a cloud on your desk? http://www.crikit.info

And that is just the beginning.

By: Lucian Armasu

Lucian Armasu — Thu, 06 Dec 2012 23:13:03 +0000

Doesn’t sound exactly like big.Little to me. ARM’s low-power and high-performance cores are on the same SoC/chip. This just sounds like they will use both Atom and Xeon chips in a server rack or something. I doubt it will be as efficient switching between them.

Anyways, the problem for Intel is that they are very reluctant to even promote Atom for micro-servers, and it shows from how they talk about it. They have a conflict of interest, because they’d rather sell the much more profitable “bigger” chips.

The reason why this is a problem for Intel is because ARM has absolutely no problem trying to sell ARM chips for servers. In fact they have all the incentive in the world to do it, while Intel has the least incentive to do it. As Clayton Christensen puts it, Intel will be “happy to concede the low-end, non-profitable (for them) market to their disruptive competitors”.

This is why Intel will ultimately lose all markets to ARM (could take a decade or more, though). Because ARM thrives on extremely cheap cores, while for Intel it’s absolutely VITAL for the company’s long term survivability to be able to sell high-margin chips, because that’s how their company is built.

As ARM chips get ever more powerful and “good enough” for most devices (that includes laptops, desktops, in the coming years), that means Intel will need to compete with $20 chips, instead of $200 chips with ARM. Look at their IVB 17W laptop chips right now. They are like $250 a piece right now. That’s absolutely unsustainable for Intel in the long-term. ARM is very close to reaching that performance level (only ~3x behind), and their chips will be an order of magnitude cheaper.

Intel simply can’t survive in that environment – not in the consumer market at least. They’ll probably manage to survive as a company for a decade or more in supercomputers and whatnot, but it’s only a matter of time before ARM chips get them there, too. In fact Nvidia’s Project Boulder is already oriented towards super-computers, too.

By: jjj

jjj — Thu, 06 Dec 2012 22:00:07 +0000

What everybody seems to be ignoring is that microserver is a short term target for ARM and the core is not the only processing unit to play with.