USB3: Still Not Healthy
It’s early Wednesday morning of Intel Developer Forum week as I write these words, and unless something changes drastically in the next few hours, another IDF will come and go without a public announcement of USB v3 support within the company’s core logic chipsets and integrated CPUs. ‘SuperSpeed’ USB was actually present in abundance at the 2010 IDF, in the form of ICs, test equipment and silicon from notable building block companies such as Fujitsu, NEC, and Texas Instruments (along with a host of smaller suppliers), as well as in mass storage devices and other system implementations. But as I wrote subsequent to last year’s show, a critical mass of USB v3-equipped computers is necessary for the USB v2-to-v3 evolution to begin in earnest. And, in spite of abundant pre-show rumor to the contrary, Intel didn’t pull the trigger on USB v3 again this year.
After last year’s conference, I (and plenty of others) wondered if Intel was intentionally hobbling USB v3 in order to alternatively advance the fortunes of its Light Peak optical interconnect technology. After this year’s IDF, I’m not sure the situation is so black-and-white. Granted, the company had a booth at the product showcase, among other things highlighting partners’ Light Peak-augmented hard drives, video transfer setups and the like. And companies such as PLX were showcasing generic ‘optical interconnect technology’ demos in their booths, on the one hand being careful (due to, I suspect, an Intel-applied muzzle) to not call them ‘Light Peak-capable’, while simultaneously stressing that they weren’t not Light Peak-compliant.
But Intel made no production silicon announcements, nor did it have any news to share about computer OEM adoption plans. And meanwhile, a few blocks away, AMD was trying its darnest not to prematurely reveal Bobcat-based product plans that encompassed integrated USB v3 support…while at the same time telegraphing me via winks and smiles that the company was in good shape in this regard. I frankly don’t know much solid about the situation; Intel representatives uniformly clammed up whenever they saw the ‘press’ Scarlet Letter emblazoned on my attendee badge. But below are some of the rumors I heard at the show:
- I had an enjoyable group dinner with Pericom last night, and one of the company representatives told me that an Intel presenter at a USB v3 class earlier in the week had flatly responded ‘2012′ when asked when Intel would deliver integrated USB v3 support to the PC platform. I interpret that statement to mean that native USB v3 capabilities aren’t planned for Sandy Bridge, nor for that matter will they be in any of Intel’s near-term Atom proliferation plans. Instead, we’ll need to wait for the 22 nm ‘tick’ derivative of Sandy Bridge, currently known as Ivy Bridge, which Intel CEO Paul Otellini said during his Monday morning keynote would begin to appear in the nebulous “second half of next year.”
- An anonymous source told me that USB v3 support was in fact originally planned for Sandy Bridge and that the circuitry remains resident on the companion southbridge chip die, code-named Cougar Point, but that it doesn’t work.
- Reportedly, due to another anonymous source, Intel’s first two Light Peak partners are Apple and Lenovo, with the latter slated to launch its first optical-augmented systems next March.
If it’s any consolation to Intel, and as a recent EDN writeup explained in detail, USB3 host controller design is by no means a simple task. Several other silicon suppliers are struggling to deliver bug-free transceivers; their existing product spins have been rejected both by the USB Implementer Forum and by potential motherboard customers. But the USB-IF announced more than 100 SuperSpeed-certified products at IDF this year. And pragmatically, although I think that 5 Gbps ‘SuperSpeed’ is the end of the road for copper-based USB technology, it’ll coexist for quite some time with 10 Gbps-and-above optical successors. As such, Intel’s USB v3 embrace is simply a matter of ‘when’, not ‘if’.
I recently saw (and ripped out for posterity sake) a frankly horrible advertisement for SuperSpeed USB in an airline in-flight magazine, which among other things referred to USB generation predecessors as ‘Sluggomorbus’ (?!?!). I can’t find the ad online anywhere, perhaps for good reason; when I get back to the home office, I’ll scan it and post it in a follow-up writeup. For now, content yourselves with this promotional video from the USB-IF website:
Followup: Here’s that ‘bad ad’