Re: Opteron


"Hoff Hoffman" <hoff-remove-this@xxxxxx> wrote in message
news:452faeb5@xxxxxxxxxxxxxxxxxxxxxx
Neil Rieck wrote:

[...snip...]

As for HP's OpenVMS and HP-UX product lines and plans, Intel Itanium is
the processor and HP Integrity is the platform. Itanium is supported for
OpenVMS I64 and for HP-UX, it works, and Integrity servers available.

Comparable Integrity system configuration prices are rather lower than
those of an equivalent Alpha system, as well.

And for completeness, I know of no plans to port OpenVMS to x86-64.

--

The easiest approach to this end is an Alpha or VAX emulator, obviously.

There's a whole sequence needed for a platform port and -- starting with
access to the source code for the operating system and for the core build
tools -- this is involving roughly 30 to 40 gigabytes of source code, give
or take, and various work in a set of compilers. Plus core layered
products.

This if you want to ask HP for all this, and particularly for the pieces
needed to perform the port out in the open-source community.

Then -- assuming HP then decides to parallel Sun Solaris in opening up
the source pool -- you have to figure out what to do to get it all to work
on x86-64, and that's not a small project in itself. That's likely a
full-time job of a number of software engineering folks for probably two
years, and quite possibly rather more. And you're probably going to have
to add on a migration over to a new database and/or to Subversion or CVS
or such.


I'm sure I'm not the only one currently wondering if HP execs have plans to
slowly kill off OpenVMS. (this statement is based upon their latest actions;
dumping OpenVMS support to India will also enable more North American
companies to provide third party support which sounds very similar to all
the companies supporting different flavours of LINUX)

If they intend on killing the OS, then putting OpenVMS source code into the
public domain is better than destroying it. While it may no longer be
commercially feasible to have 100 paid engineers do a platform port, having
1000 people doing it for free might be an alternative. I'm not saying this
port would be painless, but at least OpenVMS would eventually find itself
living on another architecture.

Neil Rieck
Kitchener/Waterloo/Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/links/cool_openvms.html


.

Relevant Pages

  • Re: OpenVMS Itanium port progressing well says Gorham!
    ... do with the actual port itself. ... This was the BS spouted about OpenVMS when it apparently ... >>have been asking for a port to a low cost commodity platform ... >>100,000 for a 4 way server is commodity pricing. ...
    (comp.os.vms)
  • Re: Intel prepares to kill off the Pentium 4
    ... It might well be rather cheaper to port the existing DCL into the target environment, particularly as the discussion expands out from DCL and into the commands and utilities and lexical functions. ... The boxes are EFI based, and the platform easily suffices for the needs of HoffmanLabs. ... It would be easier to port to a box with EFI than with BIOS, but the work involved in this particular area of an operating system port is a rounding error in the context of the effort involved in a porting project for something of the scale of OpenVMS. ... There is still real effort in getting the environment and the compilers for a wholly new platform. ...
    (comp.os.vms)
  • Re: Intel prepares to kill off the Pentium 4
    ... Roughly thirty-five million lines of OpenVMS source code and various and correspondingly giant build and test procedures, multiple languages, object code generation for x86-64, memory management for x86-64, linking for x86-64, signal handling, dealing with ACPI differences, dealing with I/O devices and interfaces, multiple third-party products. ... Be done with the port by lunchtime. ... The next step past this discussion: who would buy this OpenVMS x86-64 port, ... will Intel/Amd at one point start to "enforce" UEFI for the X86 platform? ...
    (comp.os.vms)
  • Re: Stay on Alpha forever?
    ... The expertise in the particular application bits and the platform bits, and the availability of compatible tools, can and does tend to decline over time, too. ... Emulation does not, however, remove risk -- the software stack is even longer, and you end up with the same basic issues involved with hardware support, albeit from the hardware emulator vendor, and from the underlying operating system software platform. ... Porting a non-trivial application off of OpenVMS Alpha -- whether to OpenVMS I64 or to another platform -- will almost certainly cost more. ... You can choose to incrementally move toward more portable code here, too, though there's no payback in that effort until and unless you actually do port. ...
    (comp.os.vms)
  • Re: Intriguing sticker on my Jensen machine
    ... At one time this platform ran Windows NT, OpenVMS ... and Tru64 Unix (they dropped support quite a few versions ago). ...
    (comp.sys.dec)