Intel recalls the days of Alpha, PA-RISC and Itanium

TO THE CHARLOTTE STREET HOTEL, in Fitzrovia, to listen to what Intel had to say about its Xeon 5500 (Nehalem) launch and to customer testimonials.

I stayed at Myhotel Bloomsbury, next to a pub, and which used to be a cop shop. I bumped into Paul Hales and Sylvie Barak from Register Two, who were there, too.

Gathered in front of an assembly of British hackdom was Tom Kilroy, from Intel Stateside, who described the launch of the 5500 series as the most important launch since the Pentium Pro.

Ah! The Pentium Pro! I still somewhere have a keyring with a Pentium Pro and cache attached. Intel was forced, as I recall, to re-engineer this chip because there was a problem with the cache. Think someone from Compaq tipped me off on that, all those years ago.

This set me thinking quite a lot. Kilroy wheeled in people from the London Stock Exchange, from Thomson Reuters, and naturally its customers such as Dell and HP – curiously not Big Blue – gave their sales pitches too.

Kilroy was here (left)

And as I thought of the implications of what Kilroy said – I couldn’t help wondering, of course, about the Itanium, a question formed in my mind, from whence it came, no one knows.

After showing us various benchmarks, which appeared to suggest that this was the best microprocessor Intel had ever fabricated, we had to start wondering about the Alpha chip “good until 2025” – said Richard George when he worked for DEC, and the PA-RISC chip. Because Intel seemed to be suggesting that this truly was a “mission critical CPU”. Why else would Mark Reece from the London Stock Exchange be there, otherwise?

The Hidden Agenda

After Kilroy told us that this was part of the “tick tock”  Captain Hook style Intel cadence,  we Brithacks sat patiently, waiting for the Q&A which never seemed to come.

The master of ceremonies eventually allowed a brief Q&A and pointed at me, Mikus Interruptus, saying: “Tom, would you now like to answer Mike’s question about the Itanium?”

Unfortunately my mind had moved on by then and I thought that perhaps a better question was how the financial meltdown had affected Intel’s business.

Said Kilroy: “Certainly there’s been an impact on demand”. The MC said: “Mike, we’re in our quiet period right now.”  Too late!

We finally got a chance to ask our question about whether “Nehalem” was a better chip than the Itanium, but phrased it whether it was a better chip than the PA-RISC chip – obviously with the Power 6 from IBM in mind. IBM was not, officially, represented at this gig.

Mr Kilroy said that the question didn’t really compute, because the Itanium offered stuff like RAS and you couldn’t compare a chip like the Nehalem with the Itanium.

Later, we had a chance to speak to a friend close to Hewlett Packard who told us it had told its customers last September the game was over for PA-RISC. But, we asked, it would have to support customers like the US government on both the PA-RISC chip and the DEC Alpha chip?  Yes, he admitted, that was true. The customers had the latest roadmap.

Do not forget, of course, that Carly Fiorina and the then CEO of Compaq, Mike Capellas, transacted an agreement that meant that, er, er, all things federal about microprocessors – apart from IBM – would belong to HP.

We bumped into a guy called Hugh Jenkins, who now works for the Great Satan of Hardware (Dell Inc). He said that of course Dell still uses AMD microprocessors for some of its server business. Er, BT seems to be an Intel only place, as far as we could tell.

Funny old business this, isn’t it? Intel served bucks fizz (mimosa) at the end. We’d already made our excuses and exeunted stage left before that was served.   ♦

* Spotted from  other Magee spawned websites: Sylvie Barak (INQster), Chris Mellor (Rogister).