I love the ceremony associated with this machine's retirement. As I get older and have more experience replacing gear I installed and worked with for long stretches of my career, I have started to appreciate ceremony more.
In the last 15-ish years I've taken to stashing notes in and around hardware I've installed. I write the date, an anecdote about the weather or the news, and my name. It has been nice to find these notes years later, when I'm lucky enough to be one retiring the gear. I hope, when other people have retired gear I've installed, they get a little kick out of seeing a voice from the past.
VAX power really did make big leaps from from generation to generation. The 11/750 they started with got about 0.65 VUPs and the VAXstation 3200 they ended with topped out at 2.8, over four times faster. The VAXstation 3100 M76 here, about 18 months later, makes another big jump to 7.6.
DEC spent an enormous amount of money on an emitter-coupled logic (ECL) bipolar implementation as the VAX 9000.
They came to discover that the MicroVAX hit ~70% of the performance at a fraction of the cost.
That was an expensive dead end.
"Production problems pushed back its release, by which time these fears had come true and newer microprocessors like DEC's own NVAX offered a significant fraction of the 9000's performance for a tiny fraction of the price.
"Roughly four dozen systems were delivered before production was discontinued, a massive failure. "
We had one of these at CERN in the experimental physics group I was in 1990-1991. I had no idea they were rare or that the line was a failure. It was certainly faster than the other machines the group had access to at the time (except an onsite Cray, access to which was restricted only to members from "Western" countries for political reasons).
Our shop floor system is still written in VAX ACMS, using TDMS for menus, and Oracle Rdb for SQL.
I still have an account. It runs on the Charon emulator.
A little after your time, I wrote code on a DECstation 5000/240 running MIPS/Ultrix. I'm surprised that wasn't faster, and/or available to you.
The vector extensions might have been your niche though.
It wasn’t a discovery: DEC knew the 9000 wouldn’t be competitive against VLSI implementations! DEC’s board really should’ve replaced Ken Olsen by 1985 at the latest.
> There are several quotes by prominent engineers on the NVAX project that describe Olsen's unwillingness to kill the 9000 even after being told point-blank that it would not be competitive by the early 1990s,[14] and his outright rejection that such a thing was even possible.[18]
The line between stubbornness and persistence is very thin. It's easy to rationalize post-factum that A was right to push his idea even when everyone told him it won't work, and B was stupid not to listen when everyone told him it won't work.
DEC tried something and failed, that gives me more joy than seeing Intel failing without trying. Transition from Otellini to Krzanich driven by the risk-averse board was painful to watch from within.
Right; if you claim Olsen should have listened to all the experts in 01989 telling him VAX was doomed because it was CISC instead of RISC, you are also pretty much claiming Andy Grove should have listened to all the experts in 01989 telling him the 80486 was doomed because it was CISC instead of RISC. They were right about the VAX and wrong about the 80486, but that wasn't clear for about five years. See my comment at https://news.ycombinator.com/item?id=43811508 for more details.
ECL vs. CMOS is a slightly different battle, but HPC companies (and Amdahl) kept shipping new commercially viable ECL processors until the mid-90s, and the VAX 9000 was sort of aimed at those markets. Tera shipped their first bipolar MTA in 01998, and if they'd managed to ship it a year or two earlier, it might have been commercially viable too. So it wasn't crazy to think that an ECL VAX 9000 would be the success that the ECL VAX 8000 series had been. It was just wrong.
Supposedly, DEC started to lose its way when Olsen no longer had Doriot around to mentor him.
Olsen was an old-school MIT guy - an engineer making "affordable" tools for other engineers and scientists. It was a personal crusade for him, and his biggest competitor was always the ghost of the IBM System 360.
He literally couldn't imagine the economies of scale that VLSI made possible, or how to pivot from DEC's traditional customer base into commodity computing at one extreme and the more open Sun/SGI workstation market at the other.
There's a parallel universe where DEC invented the commodity PC and made the Internet run on it. It's likely quite an interesting place.
DEC always had top-tier engineering talent and had the Alta Vista search engine well before they understood how valuable it could be.
DEC's management and vision were also top-tier, until suddenly they weren't.
At Sun we were delighted that DEC's board wouldn't replace Ken Olsen!
Only towards the end. By the time the CVAX microprocessor inside the VaxStation 3100 shipped in 1988, it was too late for the VAX line.[1] Other microprocessors already dominated. The VAX 11/780 came out in 1977. It was a long time until a faster model.
There were rumors of a VAX 11/790, and it eventually came out as the VAX 8000 family in 1984. That had about 4x the performance of the 11/780. It was not a microprocessor; the CPU still took up several large boards. The VAX 8000 had to compete with all the M680x0 machines, and soon, the Intel 386 machines. It was worse on price/performance. DEC had to go back and make a fast VAX microprocessor, which they did. But it was too late.
>The 11/750 they started with got about 0.65 VUPs
"1 VUP was equivalent to the performance of a VAX 11/780 completing a given task. One VUPS was roughly equivalent to 1 MIPS, and can be used interchangeably in most cases."
It's worth reading John Mashey's post on why the VAX was displaced by RISC: https://yarchive.net/comp/vax.html
Basically, making VAXen go faster was too much work, so they fell farther and farther behind, and they were already behind MIPS on performance in 01988. Mashey was at MIPS at the time and tells his recollections of how DEC started buying MIPS chips to build DECstations around.
The 11/750 came out in October 01980; one of the last VAXstation 3100s came out in November 01989: https://gunkies.org/wiki/VAXstation_3100_Model_38/48 (not sure about the M76 specifically). The Dhrystone performance results from https://netlib.org/performance/html/dhrystone.data.col0.html puts the 11/750 at 0.50–0.62 MIPS (very close to your 0.65) and the last VAX listed, a VAX 8650, release date December 01985, at 5.0–6.3 MIPS.
A factor of 10 performance gain over 62 months was pretty amazing; that's 56% performance improvement per year, a doubling of performance every 19 months, slightly faster than Dennard scaling. But you'd expect it to be faster than Dennard scaling because the VAX 8000 was ECL, not CMOS; it ran at 18MHz in 01985!
And that's pretty much when the rotting VAX whale splattered at the bottom of the cliff. If you're right that in 01989 they were still stuck at 7.6, four more years later, they'd only gained an additional 35% in performance over those four years, about 8% per year. If we average over the 9 years in your comment, that's 14× performance increase, 34% per year. In isolation that would be staggering—it's faster than the precipitous drop in solar power prices over the last 20 years—but it's far slower than the performance improvement CMOS got from Dennard scaling.
Other architectures (RISC, and the less demanding CISC of the 80386) were speeding up much faster. The 20MHz 80386 released in 01985 had already very nearly caught up with the VAX 8650 in computrons (though not I/O), despite being mere CMOS instead of ECL. The Netlib page documents that an AlphaServer 8400 5/300 was getting 464 MIPS on Dhrystone 2.1 in, say, 01996 https://top500.org/system/166696/. That's 800× the speed of the 11/750 16 years earlier, a 52% speedup per year.
(This is a little bit personal for me because my first internet access was on a VAX, and I think a friend of mine committed suicide due to the failure of the VAX line.)
Speeding up has slowed down since the collapse of Dennard scaling in 02006. Roy Longbottom reports http://www.roylongbottom.org.uk/Raspberry%20Pi%20Benchmarks.... that his Raspberry Pi 3 gets 2469 VAX MIPS and his 3.9GHz "Core i7" gets 16356 VAX MIPS. The i7 is of an unknown vintage (the brand name covers chips from 02009 to 02017) but I'm going to guess it's closer to the end, about 02016, 20 years after the AlphaServer and only 35× faster, only a 19% speedup per year. But this time it's affecting RISC architectures too, and it's GPUs that have left conventional single-thread-focused CPUs in the dust.
It was the mid 80's. That was happening to everything. The days of the VLSI scaling inflationary period are almost forgotten now, but it was a wild ride while it lasted.
It was like, blink and you missed virtual memory. Blink again and now you count in gigahertz.
Out of curiosity I looked up a similar-vintage uni host of similar vintage. Latest online evidence I can find dates to the early aughts.
Is there any reliable way of finding a last-effective date for a given hostname?
In related news, many of the services once offered through a locally-administered host are now provided by ... large commercial platforms. Time has moved on, it seems.
For those curious: the hosts of similar vintage were of similar vintage.
>Path: agate!agateway!CMSA.BERKELEY.EDU!CLIFF
Looks like the email was sent/received via UUCP or hit one UUCP path.
Some email certainly was, a lot of local uucp nodes polled ucbvax - this was at a time when UUCP routing was largely manual and people quoted their email relative to well known nodes (like ucbvax) - my business cards said "..!ucbvax!unisoft!paul"
No, this is a Usenet posting. Bang paths are still common there.
Thanks, forgot :)
I found this earlier this year when researching for a school presentation - it was a joy to read, but sadly did not help me much with my research! I was trying to find some code supposedly hosted on Ucbvax from a research paper. I appreciated the humor in its passing though, and if I ever take an important server down I would love to do a ceremony like this.