Itanium 2

[doug mielke]

I had no idea windows was smart enough to tell the difference between AMD and Intel.

My biggest fear of going to Intel is the expectation from management that all my support would then come from

1-(800) RE-BOOT
or
www.reloados.com

[Gregory Fruth]

1) I recall some HP whitepaper saying that they have plans for
keeping PA-RISC around through 2007 or so, though
presumably at a reduced effort level (if they can get
Itanium off the ground).

2) HP-UX 11i v1.5, v1.6 and v2.0 (aka 11.20, 11.22 and 11.23,
respectively) are for Itanium only. I recall reading that v2.0
was going to be the PA-RISC/Itanium merge but evidently
it's been put off until v3.0 or later.

3) That's the plan.

Conclusion A should be changed from "Any 3rd-party..."
to "Many 3rd-party ...". Some classes of PA-RISC
applications WON'T work on Itanium. We're talking about
binary code compatibility here. At the source code level
a recompile ought to take care of it. See:

http://www.hp.com/products1/unix/operating/infolibrary/whitepapers/WP_binary_comp_2_final.pd

[Mike Fisher_5]

Thanks Gregory

Your conclusion & link:
V useful

Seven years on the forum...
Impressed

[Dave Wherry]

The PA-8800 RISC chip is due out late this year. The PA-8900 chip is due out late 2004. So that eliminates rumor A. RISC servers will be available well beyond the spring of 2004.
If history repeats itself, these chips will be late anyway which just pushes their life spans out further.
Of course HP wants every one to move to Itanium. It's very expensive to maintain two chip lines and HP has committed to Itanium. However, current and upcoming RISC chips will continue to run for many years to come. Well beyond their end of sale dates. Just look at some of the posts from users running K, D and other older class systems. They are still doing the job.
Of course, at some point HP will essentially force everyone off of the older RISC chips by making maintenance costs prohibitive and parts availability questionable. All vendors do that. It's how they move you to new technology and reduce their costs of having to stock parts and keep techs trained on the old gear.
As for Itanium, look at the performance improvements from Itanium to Itanium 2. I'm not sure I would move mission critical apps to it yet. Maybe start the migration with some smaller apps. Just like the PA-8800, there is another Itanium in the pipeline for 2004. It's only getting better.
The key for me at this point, if I had to buy new servers, would be to get into the rp server line which will allow in chasis upgrades to new RISC and Itanium chips when the time is right. It will make future upgrades a little less painfull.

[Pete Randall]

Dave,

That sounds disturbingly familiar to the sales pitch I heard when I bought my N class: "board upgradeable to Itanium and future PA-RISC chips". HA!


Pete "Cynical" Randall

[Tim D Fulford]

My 0.02???

The PA-8800 & (PA-8900 roadmapped) are slight insurance policies for HP incase Madison, deefield etc crash (Itanium2 +). PA-8800 is due out Q3 2003 & PA-8900 2004, so there will still be some legs left in PA-RISC

The PA-8800 is a dual core chip, of two PA-8700's sharing 3MB cache (PA-8700 has 2.25MB)... BUT it has 32MB cache on-die. Effectively you should be able to turn a 4-cpu rp54xx into an 8-cpu (rp74xx)

Our company is in a quandry, do we port to Itanium with, quite frankly staggeringly high tpm-c values e.g
rx5670, 4x 1.5 GHz does about 120,000,
rp8400 with 16x750MHz does 140,000.
The cost are ~ $130k & $575k respectievly!!!

BTW to look at chip specs try

http://www.geek.com/procspec/procspec.htm

Tim

[Ian Lochray]

I am in a similar position to Tim. We are an HP reseller. We sell our own application and the hardware that goes with it. When we quote for a HP-UX solution it works out considerably more expensive compared to an equivalent Windows setup. We can try to persuade customers about the reliability and scalability of HP-UX but, when it comes to making a decision, the company FD generally steps in and goes for the much cheaper (in the short term anyway) Windows solution. We can supply on either platform but I want them all to take UNIX because they will have a far happier life in the long term.

HP-UX on Itanium 2 seems to be the perfect solution. The TPM per $ looks to be outstanding and will only get better. So long as HP-UX 11.23 provides the same reliability as HP-UX on PA-RISC I think that we in the UNIX world are on a winner.

We will have to provide PA-RISC support for many years to come but I for one want to produce and sell HP-UX Itanium 2 solutions ASAP.