This is more about Ellison and Hurd poking HP in the eye than technical merit. Itanium is a niche product for sure, but its selling, its sales are growing, and it provides a huge wealth of r&d "ip" (in the talented staff sense) when it comes to the problem intel has to solve every year "we have twice the transistors we did for the last design... what are we gonna do with them".
The loss of Intel compiler support is really telling. IA-64 leans very heavily on its compiler to get good performance -- it's almost as much a compiler project as a silicon one.
Discontinuing a support for a product developers don't use isn't really telling at all. Most Itanium development is done using compilers from the operating system vendor, not with ICC. Intel still provides the code generators for those compilers.
It is if the only selling point for the chip was it's advanced capabilities if you used the custom compiler.
If people are only buy Itaniums to run legacy x86 code THAT is significant - it would be like Apple announcing sales of Macbooks but admitting everyone just used them to run Windows
It also provides a wealth of "ip" in the sense that patent lawyers were involved from the beginning of the Itanium project to maximize the probability of Intel's remaining the monopoly supplier of Itanium-compatible processors.
Is there anyone on HN who can speak in Itanium's defense? Anyone used it and been impressed? We've all heard the Itanic jokes -- but surely there was some gold buried in there?
I've never worked with it, but the idea is that it's unbelievably fast if you have a smart enough compiler. Intel's x86 chips have a lot of logic onboard that will optimize code as it's running, in real-time. Itanium uses all of its silicon to do whatever it's told really, really fast and just trusts that the compiler has figured out all the optimizations ahead of time. Also, it has mountains of cache memory, so that helps too :)
The problem was coming up with the Sufficiently Smart Compiler to turn branchy code into explicitly parallel instructions. It never really turned up for C/C++ and Fortran code.
It would probably fly for languages like Haskell or Erlang.
I did software development for NonStop Kernel on Itanium for a few years. 5 years ago, Itanium was FAST. Our software was twice as fast as competitors who used twice as many processing cores. Problem was that, unlike x86, Intel's Itanium roadmap moves slow and continually plagued by problems. About 2-3 years ago, with cost reduction in mind, we began porting software over to Linux on x86. Initially we needed 2.5x as many x86 cores to reach the performance of the 1.6Ghz Itanium 2. By the time Intel began shipping Westmere-EP processors early in 2010, each core was nearly 1:1 with Itanium, and at half the cost and 3x the density.
I can understand Microsoft and Red Hat, because it takes actual work for an OS to be intimately aware of the architecture. But for Oracle wouldn't it just be running a cross compiler and testing on another set of boxes?
Doesn't Oracle build Java now? Building a competitive JVM requires a fair amount of knowledge of the underlying architecture.
On top of that, it's hardly trivial or cost-free as "just be running a cross compiler and testing on another set of boxes" sounds. My guess is they've done the calculations, and Itanium doesn't earn them more money.
If a customer was screwed by the Itanium end-of-life, and responds by choosing SPARC (under Oracle's custody) to migrate to, they almost deserve what's happening to them.
19 comments
[ 5.0 ms ] story [ 50.9 ms ] threadhttp://en.wikipedia.org/wiki/Itanium
If people are only buy Itaniums to run legacy x86 code THAT is significant - it would be like Apple announcing sales of Macbooks but admitting everyone just used them to run Windows
It would probably fly for languages like Haskell or Erlang.
On top of that, it's hardly trivial or cost-free as "just be running a cross compiler and testing on another set of boxes" sounds. My guess is they've done the calculations, and Itanium doesn't earn them more money.