Despite the barrage of ice storms on the East Coast, we managed to make it out to this year's Spring Intel Developer Forum in San Jose. The warm weather and lack of ice was a welcome change, as was the fact that the opening keynote began at a leisurely 12:30PM PST.

The 13th ever Intel Developer Forum was kicked off in what has become a recent tradition, starting out with an introduction by Pat Gelsinger, Intel's extremely talented CTO and ending with the usual confidence building talk by the well known Intel CEO, Craig Barrett.

Intel's Pat Gelsinger - He'll be back later this week with a bit more to say

Pat didn't have much new to say in his introduction, as we're all expecting the majority of that to take place towards the end of the week during his own keynote. The focus of his introduction continued to be this idea of convergence between the computing and communications industries, and that Intel is still working to bridge the gap between those two sectors.

Intel's "Convergence Cowboy" - Dr. Craig Barrett

Pat quickly segwayed into Dr. Barrett's keynote, which started out as the usual spiel about how Intel is continuing to invest even given the current economic climate. Intel's continued spending in the industry is definitely something to be respected, as the research they are doing into silicon technology has clear benefits for the industry as a whole in the coming decades. Showing off those clear benefits to a group of 4,000 people in a conference hall isn't exactly the easiest thing to do, but during Barrett's keynote there were a handful attempts..

Silicon Photonics Once Again

The first demonstration that was a part of Craig Barrett's keynote was the first public showing of an Intel manufactured silicon based optical modulator.

The idea of integrating CMOS Circuitry and Silicon Photonics is a goal for the future of microprocessor and silicon architecture in general

The idea behind Silicon Photonics, as Intel presented to the world at previous IDFs, is fairly simple; using light to send information within a chip, instead of conventional copper wires and electrons. Some of the major pathways within today's CPUs and platform architectures that currently exist in the form of wires and electrons will be replaced by lasers and paths to channel them. Imagine an extremely high bandwidth FSB that uses lasers to send data between your CPU and chipset, instead of the cumbersome wires that are used today.

Obviously we're not at that point today, but that's where Intel wants to be eventually. One technology that's taking them there is this silicon based optical modulator.

The purpose of the silicon based optical modulator is to take a light source and convert it into electrically compatible 0's and 1's. The idea is that the medium used for transmitting those 0's and 1's should not change the rules of the game that we've come to learn over the past several decades of dealing with CPU and computer architectures.

The optical modulator works by taking an input light source, in this case a laser, and splitting it into two signals - a reference and a phase shifted signal. If you're familiar with wave theory you'll know that you can add two signals of different phases and produce either one amplified signal or a signal with no amplitude - a binary 0 or a binary 1.

Intel used the optical modulator in combination with some conventional processors to send an image of Pat Gelsinger from one point to another; obviously nothing along the lines of an optical FSB as we mentioned earlier, but it's one of a series of first steps in a long journey to that point.

An Unusual Number of Demonstrations

Normally Barrett's keynotes tend to be focused on the overall picture, the grande scheme of things and definitely not a lot of product demos. Although his keynote this year started out on a very similar note to the one we heard 6 months ago, it ended on a much different note - primarily with the number of demonstrations that he introduced.

The first was a quick demo of some of the software Intel is using in their 90nm chip manufacturing. We've mentioned a tool known as a Focused Ion Beam (FIB) in previous articles, but to recap its purpose is to perform surgery on microprocessors. It can be used to cut very small wires through the metal layers of a CPU in order to fix defects among other things. For more information on one particular use of a FIB tool be sure to read our Inside Intel article published last year.

The software Intel demonstrated today was a noise filtering application; when dealing with smaller manufacturing processes, you can either upgrade all of your equipment to deal with the reduced scale of the transistors and wires you'll be working with, or you can try and make what you currently have more flexible. In the case of the 90nm transition and FIB tools, one of the biggest problems you'll encounter is differentiating noise from the actual parts of the circuit that you're applying the FIB to.

The proof is in the pictures from the demonstration, first we have the before shot:

...and now the after shot:

Instead of picking an actual circuit application, here we're greeted with the Intel logo etched into silicon but you get the idea. The filtering algorithm is quite CPU intensive and as you can guess, makes decent use of today's excess processing power.

Newport Mobile Platform Concept

We were quite surprised that the demonstrations didn't end with the Si Photonics and FIB bits, instead they continued on to provide a glimpse into Intel's concepts for mobile and desktop PCs in 2004. For starters we have Intel's 2004 mobile platform concept called Newport.

Here we have the Newport concept unfolded to act as a desktop

The keyboard can also be detached to make it function as a Tablet PC

The platform was running an unusual Windows interface (see below), perhaps a precursor to Longhorn but entirely too much like the current interface in order to be considered the next-generation 3D UI.

The focus of the demonstration was of course the wireless connectivity that the notebook would allow its user, including the ability to interface with a number of other non-PC devices.

In this blurry shot you can see the LCD that enables closed-lid computing

One interesting feature brought up during the demo was the idea of closed-lid computing. The feature is simple; the notebook would have a small LCD screen on the case itself that is visible and functional when the notebook lid is closed. You could use this LCD screen and interface to continue to use your PC to check and send email, including attaching files stored on your notebook's hard drive without having to boot up your PC into Windows.

The benefits are obviously reduced power consumption, since all you're powering is the CPU in a very low power mode, memory and maybe the hard drive, and you're able to continue to use your notebook to perform simple tasks even after you've closed it down. How many times have you shut down your notebook only to realize that you hadn't sent off that email you needed to, with closed-lid computing you wouldn't need to power your entire system back up to do just that.

Marble Falls Desktop Platform Concept

Next we got to see the Marble Falls desktop platform concept:

There isn't much to talk about here other than the small form factor of the machine, which adds significant credibility to the efforts Shuttle has been making in that area.

A part of the Marble Falls concept is the idea of bringing multi-monitor support to the mainstream desktop environment, which was demonstrated in the concept PC we saw during the keynote.

Powersville Desktop Platform Concept

Another 2004 concept was the Powersville desktop platform, also shown off at the keynote. This particular platform employed a Bigwater form factor and is currently using an Intel 865 chipset, although that will probably change as the concept evolves and newer chipsets are made available.

The Powersville concept also includes support for dual displays, which leads us to believe that a future variant of Intel's integrated graphics cores may support multiple display outputs, a feature that's long overdue in our opinion.

Intel also introduced the idea of dual independent audio in the Powersville PC, where you can have two separate audio streams playing to two separate audio outputs on a single PC. We haven't seen much demand for this feature, but we'll be talking with Intel over the coming days to see what their thoughts are on potential uses for it.

Final Words

That finishes up our coverage of the first keynote of the Spring 2003 Intel Developer Forum, we're off to meet with AMD now but when we get back we'll have coverage up from the Technology Showcase at IDF - including some interesting information on Centrino and Serial ATA II. Stay tuned..