Slashdot Mirror
SGI Lives On, In Name At Least
Hugh Pickens writes "In a surprise corporate move, after Rackable Systems received bankruptcy court approval on April 30 to close its purchase of SGI, the company announced on Monday that the deal had closed and that the combined company would be called SGI — short for Silicon Graphics International instead of the original Silicon Graphics Inc. The revival of the SGI brand will certainly please people in Silicon Valley with a historical bent, as SGI has been one of the area's true icons. However, some consider this a curious turn of events, considering that Rackable has come to represent the new guard in the server market, while SGI has struggled for years. Executives hope the name change will help it expand its business overseas, where SGI is a better-known brand. The new SGI will also continue to develop and support the high-performance computing systems that Silicon Graphics was known for, says Rackable's president and CEO. 'There should be no disruption to Silicon Graphics customers.'"
Posting anonymously 'cos I was involved.
The killing of the Alpha was a rather long and tortured affair. It also was a really sad day when the people who made the decision had to accept that their architecture was dead.
Let's be honest here, I loved the Alpha and it was a phenomenal architecture for its time. However, by 1998 its time was actually ending anyway. Realistically, other CPU architectures like MIPS were chomping at our heels in our opinion. They were rapidly catching up with us in terms of performance for the cost of the CPU fabbing (note that I said cost, not retail price).
We had started work on a next generation architecture shortly before Compaq came in, but we never could quite get the performance out of it that we felt we needed to keep ahead of the next generation of competitor's CPU's. We wanted to define a 5 year architecture; one that could scale rapidly in performance while keeping our development costs relatively low... but as we started to work on it we realized already that it was going to be really hard to create even a 3 year architecture, let alone 5.
By the time Compaq came in, there was already rumbling in the halls that Alpha was in trouble... this wasn't news to anyone. As a result, when Compaq did arrive and told us their plans we basically had only one realistic choice; shutter the Alpha division.
If we'd moved ahead with the next gen architecture, it would've been the Alpha equivalent of a Pentium 4; kludged to improve clock speed at the expense of creating an architecture that was fundamentally a dead end at day 1. Intel could soak this up because of volume... DEC never had a chance. We didn't sell that kind of volume, ever... and we didn't have the marketing muscle to catch up. As a result, we couldn't create an "Alpha IV"... it would've killed us anyway.
In the opinion of those who were present at the meetings when the future direction was set, Alpha wasn't killed, rather it was euthanized. It was better to go out in a blaze of glory than to fade into irrelevance.
On the bright side, many of the engineers I used to work with now have great jobs that they love... some working for Intel, but many working in what they love; small boutique CPU's and embedded systems.