Submitted by This latest 2017-2019 product roadmap includes, for the first time, the latest support roadmap. There is also further details about the next OpenVMS V8.x and V9.0 release for Itanium, along with the “early adapter” release of V9.0 of OpenVMS for x86 servers.
Development is continuing at a steady pace.
That is a lot of work. Entirely different CPU, creating compilers that don’t currently exist, refactoring the kernel to operate in the same reliable way as on Vax/Alpha.
I suppose it is what DEC should have done all that time ago instead of putting the effort into Alpha hardware/software and then onto Itanium with HP. A better move would have been to migrate to Intel x86 despite the hardware initially being a generation behind.
VMS on x86 is achievable now with hardware emulators, the x86 port will put the cat among the pigeons in that market. However, there are still some advantages that an emulated Alpha/vax has over a x86 native VMS platform as there are certain aspects that it can never achieve such as byte for byte cross compilation to target systems, the x86 result will always be different from the original vax/alpha and so the emulators will retain the upperhand when it comes to secure legacy system replacement.
A little perspective:
At the time that VMS was ported to Alpha, x64 wasn’t even a dream, and there was nothing enterprise about the x86 hardware offerings. Alpha served DEC well.
In hindsight, the Itanium effort was a complete waste, but even then, Intel wasn’t even thinking about x64 – and it would have been even more ridiculous to port to 32-bit x86.
Hindsight? Ha! Everyone knew it was a waste while it was going on. That’s why many of us are still bitter about the Alphacide.
I’d agree about hindsight – but at the time it was clear to many that itaniaum was the wrong choice but probably the only one available to DEC/Compaq/HP. A better bet would have been to wrest x86 away from Intel and extend it to 64bits just as AMD eventually did. DEC still thought of themselves as big players (as they really were for a while) and they thought that their own proprietary 64bit Alpha hardware was the sure fire way to lock in existing customers. x86 was not considered a grown-up solution at the time. Hindsight is a good thing but AMD got it right, grabbed the x86 specification and took it to 64bit before Intel even knew what was happening, killing Itanium in the process but it still didn’t help AMD that much…
That is true. I mourned Alpha’s passing too. I’ve still got an Alphastation 500 sitting around. Itanic was more about Intel positioning product than it was about advancement of the state of the art.
Absolutely it’s a massive body of work and I’m impressed with their progress thus far. Here.. this is a bit older but you’d probably get a kick out of what is at what stage:
https://vmssoftware.com/pdfs/State_of_Port_20160906.pdf
I’m surprised at using llvm, but if it works…
They really wrote “early adapter”. Argh…
There was a Holy War here at NASA in the last 80s and early 90s between Unix and VMS. Unix won.
I agree. Also yes, i386 of the era was truely a joke compared to these days. Remember, most people still ran MessyDOS when ia64 was being imagined. IA64 was indeed tragic. One minicomputer O/S, One UNIX were also buried due to it and a third UNIX severely harmed.
I still run a few Alpha’s here and a TS10 (converted) and ES40 suit my needs. Should try and do something with these DS20E’s but I just don’t really need them.
I still tend to gravitate to VAX and my 4000m96 and VLC do most of my needs.
Although I have to wonder about the business case. Most possible clients are going to be legacy cases, I don’t think there’s much interest in VMS as a new target.
Last time I touched VMS was when I was bored and the lab I worked at had some mothballed alphas, and there was some hobbyist program where you could get a hold of the OS for very low cost. I think you had to renew the licenses every year, and just about everything had to be licensed, from the TCPIP stack, to the compilers, etc. all requiring their own individual licenses.
The system was interesting, in a software nerd archaeological kind of way, but struck me as a bit hostile development-wise.