After three years of delays, Merced shipped as Itanium on the 29th of May in 2001. The first OEM systems from HP, IBM, and Dell were shipped in June. Itanium, whose architecture was now referred to as IA-64, was a 6-wide VLIW chip running at either 733 to 800 MHz with a 266 MT/s front side bus. It had 16K of L1 cache, 96K of L2 cache, 2 or 4 MB of L3 cache, and it was a single core chip on socket PAC418 built on a 180nm process. That this chip under performed is an understatement. Given the long development time, multi-billion-dollar development cost, significant hype, and claims that it would out-compete everything on the market… Itanium was a massive failure. The 32 bit x86 chips of the time were able to best it in most workloads. Embarrassingly, the Pentium 4 (whose own performance wasn’t that good) beat Itanium on integer performance and memory bandwidth. Those areas where the chip was strong were in transaction processing and scientific applications. John Crawford, Merced project leader at Intel, reflects: “Everything was crazy. We were taking risks everywhere. Everything was new. When you do that, you’re going to stumble.” The five hundred person team working on the chip was also relatively inexperienced, and disagreements between HP and Intel led to many compromises in design.
↫ Bradford Morgan White
Itanium is the future. This entire article is anti-Itanium propaganda and misinformation. Do your own research.
Itanium was designed to kill of PowerPC, R*000, SPARC, PA-RISC and Alpha, and it did a fine job of that, even before a single chip had been launched.
Imagine what Alpha could have been like if it had been the plattform of choise instead.
NaGERST,
+1!
Alpha had a lot of technical merit with a strong native 64bit implementation, but they were locked out of the market by much stronger incumbents. The partners they would need to succeed were under intel’s control. Microsoft’s support for alpha was incomplete at best. Itanuim would go on to win because of unmatched marketing muscle, but loose over their architecture’s technical debt. 🙁
I think Alpha was great, I loved it, but its spirit didn’t completely die. Its like the BeOs of chips. People who worked on it left the company and went on to work on other chips and made them pretty darn good. Lots of folks ended up at AMD, and then later Pa-Risc/Apple.
People get the corporate maleficence confused with technical failure, I’m no expert on the rise and demise of Itanium, but like many other technical developments o me it seems it’s fate was more sealed by corporate greed than any potential technical shortfall.
As @Alfman points out, Itanium won against competitors due to heavyweight marketing muscle, but I’d assert that same influence was the reason it lost.
Does building to a commercial / corporate plan ever succeed?
cpcf,
Itanium definitely had some interesting technical aspects. Having a compiler schedule CPU instructions might have merit if the software were capable of doing so, but it’s very clear that the compilers of the day were not ready for VLIW. Itanium’s theoretical parallelism on paper wasn’t achievable without assembly and typical applications ended up performing quite badly. Explicit parallelism as we are familiar with it in GPGPU could have been awesome back then, but the VLIW kind of parallelism that itanium was going for never really went anywhere.
The rolling register windows were another interesting idea. Functions can contain a lot of boilerplate logic to push and pop stuff between registers and the stack. Having a register window that rolls around with the stack could make function calls cost less in theory, but I’m not sure if software was ready for this or how efficient itanium’s implementation actually was. Itanium used lots of transistors to implement it’s features, meanwhile software was optimized for a wholly different kind of architecture.
Alpha would have been a much better fit considering that itanium failed to change conventional software practices. It is what it is, but like NaGERST says it would have been interesting to see where alpha could have gone if they had gotten itanium’s share of enterprise investment instead.
If Amiga moved to Itanium, it would’ve been the most widely-used platform today!
In addition to HP, IBM, and Dell, SGI and Fujitsu also had first generation Itanium workstations. All of them were rebadged machines manufactured by Intel, IIRC.
The pharma company I worked for at the time was a big SGI shop, so SGI sent us a few prototype Itanium systems. They put in a lot of technical/engineering resources and worked hard to convince us that Itanium systems were the future, but the systems were enormously disappointing. The rest is history, both for SGI and for Itanium.