HP's new coke moment
We're doing a real world test of the big three and HP's itanic is sinking fast. It performs well behind Sun's UltraSparc64 VI and IBM's Power5+ and so far, our best numbers are on the Sun/Fijitsu chip.
HP's Itanium server customers face a daunting endurance test over the next two years. We've just learned that Intel's next version of Itanium - Montvale - looks set to disappoint. It will arrive this year at 1.66GHz - just a hair faster than today's 1.60GHz chip. Intel once hoped to kick Montvale up near 3.0GHz, but those …
is for people who want to keep their old hp-ux software running or for developers/managers who are stuck in the world of the original system V and doesn't want to learn how to code on or understand modern systems. Unfortunately when buying servers, my company decided on hp itanics. The funny side is that our linux based development systems (standard intel P4-s with asus motherboards) are faster than the itanic servers with hp unix where we are running the developed software (comparision is done with the same load). The same goes for the storage kit, our home made sata array powered by an industrial pc is faster, bigger and has a better reliability than the expensive storage system the company purchased for the servers.
At what point will Intel+HP give up on IA64? Almost all the customers are HP's ... HP controls the whole software toolchain. How critical is binary compatibility to this community, given that it isn't really all that old?
At some point HP+Intel must see that the resources spent to keep Intanium limping along behind the pack make not sense. It would be better return to spend something to buy customer loyalty through a transition and move on.
Where they move on to is another matter
It is software, indeed, as stated in this article. Mostly low-level OS layers, like LVM, not even application, since still today, a lot of them haven't found the time/expertise to recompile under itanic ! HP-UX + LVM always do the job in a SAN environment and costs nothing as compared to vxVM, its heretic descendant (heretic, because, being exactly the same, it decided to change all of its cmds to absolutely insane names/concepts).
Only HP-UX and its integration to storage env keep itanic afloat. In a world where Linux (with a single distribution) would have ruled, Itanic would have sunk long ago.
PA-RISC is dead...were HP to try to revive it they'd be years behind Sun in R&D...who are themselves effectively years behind IBM (yeah yeah, depends on the workload etc, but in single-thread performance you'd do well to beat a Power6).
HP is stuck with Itanium, and I'm sure they wish they weren't...Their only hope is software.
And HP has been gobbling up sottware companies like gumdrops. And Don't forget, when Ashlee referred to a "Non-Stop" blade, he's talking Non-Stop-KERNEL as in Tandem. If I had my life depending on a server, I'd want it to be a Tandem design running NSK. All though stock exchanges and Telcos and power grids can;t be all wrong.
With a statement like "(Both vendors should keep beating up on Sun/Fujitsu.)", I can't believe you've used a sparc chip since the ultrasparc IV+ 1.5 chips. Ahhhh, the newer utlrasparc64 VI blows away the itanium. Heck, the itanium is put to shame by the newest, fastest ultrasparc IV+ (2.1)