jws@hpcllf.HP.COM (John Stafford) (02/19/88)
The following is what I think I remember from when we encountered the
same problems with contrib programs and 5.2... This is not an official
statement from HP...
Some versions of the contrib programs used a shared library which
changed at 5.2. The solution is to simply get the latest CONTRIB tape
and reinstall them (the ones on the latest tape don't use the shared
library). The CONTRIB tape is part number 97080-19200 (the one I have
is dated 7/86 and works fine).
John Stafford -- Hewlett Packard Computer Language Lab
{allegra,decvax,ihnp4,ucbvax}!hplabs!hpda!jws
{fortune,sun,thirdi,ucbvax} !hpda !jwsjv@mhres.mh.nl (Johan Vromans) (03/03/88)
In article <4760006@hpfcls.HP.COM> mcb@hpfcls.HP.COM (Mike Berry) writes: >Re: Support life of s500 >I believe our support life is 10 years. But with HPUX frozen at 5.2 for the next 10 years, there will be no System V rel 3 featurs, no System V rel 4 and so on. Although the HP9000 and it's HPUX operating system will be supported, they are technically dead.
chan@hpfcmr.HP.COM (Chan Benson) (03/16/88)
> We don't even think an 800 series can compete with a 500 with 2 or > 3 computers in terms of multi-user capablility. I assume you mean 2 or 3 cpus. I don't know what your applications are, but having gone from a 500 to an 840 here at work, I have to say that the 840 blows the doors off a 500 for multi-user central system. Well maybe that's a little strong, but it is considerably faster, even with a greater number of users. I don't think that HP has ruled out doing a multi-processor 800, but it is a lot of work to do right. Chan Benson HP Fort Collins This does not represent any official HP position, just the opinions of a lowly employee (oops I mean loyal employee).
wunder@hpcea.CE.HP.COM (Walter Underwood) (03/17/88)
>> ... We don't even think an 800 >>series can compete with a 500 with 2 or 3 computers in terms of multi-user >>capablility. > >I agree. Well, I've used both, and I disagree. We replaced FOUR s500s with one 840, and everybody was happy. Comnpile, link, and install of notes went from 45 minutes to 9.5 minutes. Interactive response was far better. The s500 was killed by a few uucicos, but we just don't notice them on the 840. Our s500 had two processors -- we did some tests, and decided that more memory was more useful than the third processor. The s800 has a faster filesystem and a faster CPU. Everything that we have tried on our 840 is much, much faster than it was on our s500. I think I've heard about some pretty good trade-in programs for the s500. I'd look at those, if I were you. >>So the logical question here is: Why doesn't HP have a >>multi-CPU 800? Was the experience with the 500 series so bad that HP has >>given up that idea? The "HP Precision Architecture Handbook" (part number 5954-9906) notes that the architecture supports "multiprocessing for fault-tolerance or increased performance." It also specifies the cache-consistancy requirements. The multi-processor support is already designed in. When will HP ship a multiprocessor 800? Heck, I have no idea. We just announced a 14 MIPS workstation (the 835), so that should keep the CPU hogs happy for a little while. Wish I had one. wunder
chuck@hpunsca.HP.COM (Chuck Munro) (03/20/88)
The comments I've read from Bo Thide, and the various responses to this note string have got me to thinking about the issue of older-machine obsolescence vs. new-machine development. Many of the comments struck a nerve with me, since I was a customer of HP for over 7 years before I joined the company as an SE. Bo's problem is certainly not unique - I think we've all felt the rather helpless sensation as technology marches on, and we all race to keep up. From my perspective as a former customer, I still remember the frustration when HP would announce a newer/better/cheaper replacement product, and then proceed to give the new product all the goodies which I wanted to see added to my older system. The standard HP answer was "you should upgrade your system .. the cost is minimal". That was ok for the rich guys, but I worked in a hospital clinical engineering environment, and we never had *any* upgrade money available! So we put up with the old system, and waited for 2 or 3 product cycles until we could afford the change. But wait - the upgrade was now no longer available ... too late! I'd bet that universities are in the same financial situation as hospitals. From the same perspective, I suppose the situation is actually getting a bit more difficult, since product life-cycles tend to be getting very short, no matter who the manufacturer is. Well, I subsequently put on an HP hat (that was 9 years ago), and soon I got to see the other side of the coin. Lucky for all of us, HP began to see the need for a unified system architecture, and elimination of the many operating systems which we had, in favor of a smaller and more closely-focused number of O/S's (for multi-user systems, we're down to RTE, MPE, and HP-UX, with really good migration paths from RTE and MPE to HP-UX). At one point, I was supporting six flavors of RTE at one time ! Believe me, it wasn't an easy thing to do ! Now, for the HP9000 'family' we saw the results of a rapid swing of the technical community to UN*X-based systems, combined with a rather sudden realization that we (HP) also had to curb the proliferation of CPU architectures before things got too far out of hand. The Series-500 was a real engineering miracle when it was introduced, but if we stayed with it, the competition would eventually eat us up. Why ? Because making the FOCUS chip set faster and faster would become exponentially expensive, largely due to its CISC design. HP was at a crossroads, and a decision had to be made ... even more new architectures to suit every market, or a single, manageable machine design (which had made DEC so successful). So, we 'cut the cord' and spent many millions on what has already proven to be a *great* new computer. So, what about we Series-500 users? (I'm using one to write this note.) Classical catch-22 !! HP could easily upgrade HP-UX continually for the s500, and add all those goodies which many of us really want ... BUT ... are you willing to pay for this service? It costs real money to hire the people to upgrade the O/S's for the new CPU's *and* all of the old ones at the same time. That money has to come from customers. So it becomes simply a question of economics. If enough s500 customers indicate to HP that they're willing to pay for continued upgrades and enhancements, then I bet you'll get just that. When you mention this to your local sales rep, make sure you get him/her to pass your feelings on to the factory. The biggest problem with the s500, from a selfish perspective, is that it's just too damned reliable, so justification for replacement is more difficult! I've *never* had a single hardware failure on this beast in two years. It runs 24 hours per day, on unfiltered AC power, and has even been used at sea during rough-weather environmental tests for shipboard use. However, I'd be the first to admit that the O/S is getting a bit obsolescent. The Series-300? Well, don't hold your breath waiting for it to be replaced with the s800 just yet. It's got features (and more coming) which will ensure that it'll be around for a few years more! And already it can run circles around my 2-cpu s500. I suspect that Rocky-Mountain BASIC will keep s300 sales cruising along for quite some time yet, even if HP-UX users all convert to s800's. And ... it's a relatively inexpensive box. Personally, I'm caught between two worlds when it comes to the philosophy of continuous system upgrading .... as a former customer, I remember the constant battle for funds to achieve this .... as an employee of HP, I see the need for customers to modify the traditional policies of "use it until it drops". Technology is simply moving too fast! It's tough to please everyone when using finite resources. The Series-800 should finally allow us (HP) to more easily retrofit software upgrades to older HP-PA machines as time goes on. You'll see HP-PA around for a very, very long time. Thanks for listening! Customers : keep sending us your thoughts, please! Chuck Munro hpunsca!chuck
irf@kuling.UUCP (Bo Thide) (03/20/88)
In article <1080002@hpfcmr.HP.COM> chan@hpfcmr.HP.COM (Chan Benson) writes: >> We don't even think an 800 series can compete with a 500 with 2 or >> 3 computers in terms of multi-user capablility. > >I assume you mean 2 or 3 cpus. Yes, that was a typo. Sorry! > I don't know what your applications >are, but having gone from a 500 to an 840 here at work, I have to >say that the 840 blows the doors off a 500 for multi-user central system. Our applications are science-oriented computing (not only number crunching but also text formatting with nroff and TeX and lots of I/O-oriented tasks). A week ago bought a second, used 540 (very cheap ....) and added its FOCUS II CPU and 1 MB of FOCUS RAM to our existing 2 CPU/4 MB RAM 540 to speed it up even further. We had noticed some problems with 10 people logged in. Since we haven't got 'monitor' for HP-UX 5.21 so we have not been able to analyze our new configuration yet. By the way, we plan to use our new 540 cage as an extra I/O cage thus saving some $4500... -Bo -- >>> Bo Thide', Swedish Institute of Space Physics, S-755 90 Uppsala, Sweden <<< Phone (+46) 18-300020. Telex: 76036 (IRFUPP S). UUCP: ..enea!kuling!irfu!bt