Itanium allies fund open-source compiler work

In the newest of a series of moves to try to impart momentum to Intel's Itanium processor, allies backing the chip are funding work to improve a key programming tool.

Some of the $10 billion in five years that members of the Itanium Solutions Alliance are spending on Itanium market and technology development will go to Itanium-specific improvements to the GCC, or GNU Compiler Collection. The Gelato Federation, an organization dedicated to boosting Linux on Itanium computers, is overseeing the work, the allies said this week at the Intel Developer Forum in San Francisco.

Compilers translate human-written software into instructions a computer understands, and compilers are key to Itanium. One core principle behind the Itanium chip, codeveloped by Hewlett-Packard and Intel, is that compilers can sort instructions in advance so the chip can process several with each tick of its clock.

Specifically, the work is geared to improve compiler details: superblock scheduling, instruction scheduling and speculation, and memory disambiguation. The ultimate goal is "to deliver a GCC compiler optimized to support Itanium solutions," the alliance and Gelato Federation said in a statement.

[scdecade]

Too little too late

Yet more good money spent after bad. What the Itanic deckchair rearrangers need to wake up and realize is that these tools needed to be in place 7 years ago. Now it's just plain ridiculous. Compilers are the most essential key to getting maximum performance from EPIC. Who would invest in this platform for their business? At some point Intel's x86/Xeon customers are going to get sick of paying the Itanic premium for commodity products that are inferior to AMD.

Intel's and HP's only hope is Sun. $10 Billion would buy a majority interest in Sun. They'd get an enterprise OS, mature widely supported tools, and the $100+B SPARC installed base. HP was (and seemingly is) too stupid to realize the value and importance of their own PA-RISC/HP-UX installed base. HP could take over Sun's systems and integration units and Intel could take over SPARC developement. This is their one and only realistic hope. Really, the other option would be to buy AMD but that'd never get that past anti-trust review... Otherwise, they're going down with the ship. They've got the money to do it now. In two or three years it might not be so easy.

[fred dunn]

Niche Market, Processor is "dead"...

Why would anyone want to risk their resources on an already unpopular architecture while the x86 64 bit processors will do just fine and there is a HUGE market?

Intel and HP stop wasting your time and move forward with some REAL innovation.