Intel Ramps Up 45nm Chip Production, Announces 'Atom' Line

Multiple readers have written to tell us of the latest developments out of Intel. Earlier this week, Intel announced the Atom brand of low cost, low power consumption processors. The CPUs, measuring only 25 square millimeters, are the result of the Silverthorne and Diamondville projects. The announcement has caused this CNet columnist to question whether Intel can "spur innovation in ultrasmall devices the way it has in the PC and server industry." Concurrently, Intel has increased its production of 45nm processors to a rate of roughly 100,000 chips per day. As TG Daily notes, the massive investments Intel has made into chip production will make it difficult for AMD to catch up.

Comments

Laughably high power consumption for handheld

[pslam]

The Atom architecture is intended to give Intel a foothold in handheld devices that have traditionally been the sole domain of very low-power RISC processors. The chip itself is tiny at less than 25mm square, and, according to Santa Clara, has a TDP of 0.6W - 2.5W, as compared to a 35W TDP for a "typical" Core 2 Duo.

Sigh. They do this every year or two - Intel announces a new core that will get them into more handhelds. They're still an order of magnitude short. Typical "very low-power RISC processors" you see in a device such as a mobile phone or MP3/video player are more like 0.01W - 0.25W, or even less. They're way more efficient clock-for-clock (and MIP-for-MIP) than any x86 core Intel has ever churned out.

Unless they have a funny definition of hand-held device we don't normally use, of course.

Re:Laughably high power consumption for handheld

[pslam]

These chips aren't designed to go into cellphones, and Intel frankly says they are not going into cellphones. They are instead designed for MIDs that will predominantly run Linux.

That's funny, because according to the link, MIDs are a class of hand-held device invented by Intel. So I'm right - they have a different definition of hand-held to everyone else.

The next generation of Atom at 32nm will have the proper power envelope to run your cellphone BTW.

They will be 10 times more power efficient than their 45nm version? Extremely unlikely. Also consider that the real lower power processor market isn't standing still either - they're managing about a 25%-50% power efficiency improvement per year. Also consider that the current high-end low-power CPUs you find in mobiles are comparable in performance to the first-Gen Centrino chips.

The kind of "hand-held" devices Intel are talking about have big batteries and are held with two hands. 1 Watt is not a lower power device in this market. The real hand-held device chip market measures their power in milliwatts not watts. They idle at a single milliwatt and average a 20-50mW in use. Intel is still running orders of magnitude higher than that.

Re:Laughably high power consumption for handheld

[pslam]

I was under the impression that most of these had extremely lousy performance, and relied on dedicated decoding chips specificly designed for the task they do.

It's a common impression but it's wrong. For example, most of the CPUs you find powering MP3 players do decoding entirely in software. Even many CPUs powering hand-held video players do decoding entirely in software, but to be honest you'll find most of the high-end video players use hardware decode because a) it's faster and b) it's more power efficient.

The performance of the CPUs you find in a most high-end hand-held devices these days is surprisingly good. Well, it's surprising to people who haven't worked in the field, at least. We're constantly somewhat annoyed that the rest of the world hasn't worked it out yet. A high end ARM11 (common in high-end mobile phones) is actually quite competitive to the performance of a Via C3, for example.

Re:Laughably high power consumption for handheld

[bhima]

From Linux Devices: http://www.linuxdevices.com/news/NS5492118276.html

ARM vs. Atom

There's much to like about the Intel Atom, writes Williston in EETimes. Yet, he suggests, the media and its readers may have been overwhelmed by the hype machine. Williston offers the following responses to typical arguments from the atomic power lobbyists, at times quoting analysts such as Forward Concepts's Will Strauss to back him up:
Atom will beat ARM because it can run Vista. -- No it can't, says Williston. Atom can run Windows CE and Linux, but ARM can do the same.

Only Atom offers a "real" Internet experience with Flash video, YouTube, etc. -- "Wrong," writes Williston, pointing to ARM Flash players from BSquare, and an ARM-based YouTube decoder from On2. He might also have noted that Nokia's ARM- and Linux-based Internet tablets use a Mozilla-based browser, with plugins for Flash, Windows Media files, and even Microsoft's Flash-like Silverlight technology.

Intel dominates every market it enters. Here, the writer refers the reader to the history books, especially two years ago when Intel sold its PXA line of embedded processors to Marvell after failing to dominate the market for ARM-based SoCs.

Atom will win because ARM is proprietary technology. Nope, he writes. ARM chips are available from a number of semiconductor vendors.

Intel will win on cost. Not likely, he writes. Using a 65nm process, the Cortex-A8 occupies less than 3mm x 3mm, he notes, while the Atom core probably takes up about 9mm x 9mm of Atom's 25mm x 25mm die size, despite its smaller 45nm process. "With such a huge area disadvantage, it's hard to see how Intel will win on cost," he writes.

Intel will win on power. Once again, not likely, he argues. Intel quotes a thermal design power (TDP) of 0.6W to 2W for Atom, he writes, but doesn't specify clock speeds. ARM offers only "typical" power measurements, making comparison difficult. But at best, he suggests, Intel matches ARM on power usage, while "in most scenarios, Atom burns more power."

Intel will win because it has the most advanced fabs. Perhaps, he writes, but who cares? "Consumers focus on cost, power and speed," he writes.

Re:Laughably high power consumption for handheld

[SlashWombat]

Sorry, if you RTFA carefully, the Atom chip size is 25 square mm. This means its about 5mm by 5mm.

While intel did sell its PXA line of ARM uP's, it still makes a fairly large range of ARM processors, most of which clock at fairly impressive speeds. (Faster than most of the competitors ARM uP's) Easy to check, just go to the Intel site.

Even ARM processors start requiring a fair bit of power when the clock rate gets high.

ARM IS proprietary. The fact that every semi vendor appears to have ARM in its lineup just means they have licensed the arm core.
The ARM's that draw milliwatts from the supply are NOT the ARMs used in mobile phones ... They are the "embedded" arms, running at less than 80 MHz. (they are still very nice chips, but, no go in a phone. These slower ARM's simply don't have enough grunt to perform the relevant speech compression algorithms. Most of these slower ARM's are ARM7 devices which do not have MMU's, making multi threaded apps significantly slower.
If Intel offer these new chips at a low enough price point, they probably will eventually drive ARM out of the market, but only if the ARM is denied access to 45/32/smaller geometries. This is not really likely. BUT, the development tools for the X86 architecture are MUCH better than dev tools for the ARM. This in its own right will ensure future development for many apps will use low power x86 devices.

The Ars Performance Judgement

[gnutoo]

Atom's performance in shipping hardware isn't something we've been able to test, yet, but given the architecture's simple, in-order nature, you shouldn't expect Atom to match even a Pentium M in raw performance.

AMD is supposed to feel threatened by that?

Re:The Ars Performance Judgement

[TheRaven64]

ARM's small but tenacious market share Last figures I read (from early 2007, admittedly) showed that ARM was the most widely deployed CPU architecture in the world. Considering that mobile phones (which outnumber PCs by about 3:1) and set top boxes almost all use ARM cores, I think calling ARM's market share 'small' is quite funny.

Re:AMD doesn't HAVE to compete in this market.

[ocirs]

That statement isn't made towards atom, rather Intel's ability to mass produce 45nm which costs less, performs better and generates less heat. AMD is just starting 45nm production and by the time we see it hit full production, you can expect Intel to be transitioning 32nm.

Re:subsidies anyone?

[that this is not und]

People need to remember that AMD is only in the x86 business at all because they got their foot in the door as a second-source producer of Intel chips decades back. Without those old agreements, they wouldn't be making an x86 processor at all.

Re:I don't think it does

[TheRaven64]

It's not the high-level runtimes that are making x86 obsolete, it's the low-level ones. Emulator technology has come a long way in the last few years. Rosetta (which Apple licensed from a little start up from Manchester University in the UK, by the way), has really shown how unimportant the ISA is. I run a few PowerPC apps on my MacBook Pro and don't even notice that they aren't native.