Itanium Solutions Alliance won't solve the problem
Published: 01 Sep 2005 13:50 BST
It's a good thing for Intel's UK marketing department that the proposed Government crackdown on extreme Internet pornography is not yet in force. With bestiality, necrophilia and severe sadomasochism all off the menu, anyone enthusiastically flogging a dead horse online will be looking at years of porridge for breakfast — and these guys have to sell the Itanium.
To forestall the anguished phone call from Swindon, let's look at the chip's strongest stories first. Sales are up 100 percent year on year, it's getting big wins like the Dresden University of Technology's 6TB, 1,500 core Itanium 2 SGI supercomputer, and Intel's Itanium partners have gathered together to form the Itanium Solutions Alliance club in time for the dual-core Montecito push. The chip does very well at benchmarks, is building solid OS and developer support and is making headway against the RISC competition. "Everyone loves Itanium except the press" was the exasperated refrain at the Intel Developer Forum, and in a world defined by high-end chips for high-end jobs this would be a valid complaint.
Yet Intel itself is doing more than anybody else to redefine that world. At the same IDF, all the buzz surrounded x86 developments. The new Converged Core architecture promises to bring together the best aspects of the Pentium M and the Pentium 4, meaning much more flexibility in creating high performance systems that cost less to buy and run. Everything is or will be 64-bit, neatly appropriating one of Itanium's architectural advantages. Even the next wave of non-64-bit portable chips will be able to address 64GB of RAM. And dual-core is leading inexorably to multicore, with eight-way x86 devices on the radar. With virtualisation and much better management technology, it ceases to matter what's under the hood as long as it delivers raw power at a reasonable price.
Core for core, the Itanium drubs the x86 — but with a five to one or greater cost differential, before whatever discounts Intel offers to club members, it is rarely at an economic advantage. The x86 market is large enough to support genuine competition on price and performance, and as Intel says there are many years left of architectural improvements in the old dog. It's making all the money so where, exactly, is Intel concentrating its development dollars? IDF made that very clear.
The Itanium Solutions Alliance is a good thing, but it reflects the sobering fact that the market for the chip is so small that only by collaborating can ostensible competitors hope to grow it to profitability. Five years ago, such an approach would have had a reasonable chance of succeeding, but with each lurch forward in x86 capabilities the window of opportunity gets closer to closing for good. When even the cheapest jalopy can outrun a thoroughbred racehorse, it's time to make that reservation at the knacker's yard.
Did you find this article useful?
29 out of 51 people found this useful
- Share this article:
